OPERATIONS RESEARCH THEORY AND APPLICATIONS Fourth Edition JK SHARMA Formery Professor Faculty of Management Studies University of Delhi, Delhi MACMILLAN(© Macmillan Publishers India Led, 1987, 2003, 2007, 2009 ‘All rights reserved. No part ofthis pubicatin may be reproduced or tmansmited, n any for or by any means, without permission. ABy ‘etson who does any inauthorised act in relation to this publication nay be liable to eiminal prosecution and civil elas for damages. First published, 1997 Reprined, 1998, 2000.2002 (sir mes) ‘Second edition, 2003 ‘Reprinted, 2003-2006 fight times) ‘Third edition, 2007 Reprined, 2008 Fourthedition, 2009 Reprnied, 2009 (nice) MACMILLAN PUBLISHERS INDIA LTD. Delhi Bangalore Chennai Kotkat, Mumbai ‘Anmesabad Bhopal Chandigarh Coimbatore Cuttack CGuwabati Hubli Hyderabad Jaipur Lacknow Madurai [Nagpu: Patsa Pune Thirwvananthapuram Visakhapatzam Companies and representatives throughout the weld ISBN 10; 0280-63885.6 ISBN 13: 978-0230-63885-3 Published by Rajiv Beri for Macmillan Publishers Inia Lid 210 Aasar Road, Daryagan} New Deli 110 002 Lasertypeset by Sara Assignments ‘A4(B-49, Johi Colony, Patarganj Ext, Delhi 110 092 Printed at Baba Barkha Nath Printre 37, MIE, Bahadurgarh, Haryana 124 SOT “Ts bok ant ec aig pues Te aba) fe nk vv ei esol caro es tbe Get Be bak do ve any xg copyright oe nelle pope igs aay pene in my mana woo, eee be mor) sae Ben ule ‘Chek any soa al apy ba een ene nin pe ol he plier ning er ese onContents Preface to Fourth Edition Preface to First Edition Chapter Chapter 1 Operations Research: An Introduction 1.1 Operations Research — A Quantitative Approach to Decision Makin 12 The History of Operations Research 3 13 Definitions of Operations Research 14 Features of Operations Research Approach 6 13 Operations Research Approsch wo Problem Solving 7 Conceptual Questions A 7 16 Models and Modelling in Operations Research 8 = Classification Besed on Structure 9» Classification Based on Function or Parpose_10 * Classification Based on Time Reference 1 + Classification Based on Degree of Certainty * Classification Based on Method of Soluion or Quantification V1 1.7 Advantages of Model Building 12 18 Methods for Solving Operations Research Models 12 ‘+ Anelytical (or Deductive) Method 12 + Numerical (or lerative) Method 12 # Monte Carlo Method 12 1.9 Methodology of Operations Research 12 L109 Advantages of Operations Research Study 15 LIL Opportunities and Shortcomings of the Operations Research Approach 16 L12 Features of Operations Research Solution 16, 1.13 Applications of Operations Research 16 114. Operations Research. Models in Practice 17 115 Computer Software for Operations Research 19 Conceptual Questions B19 Chapter Summary 20 Chapter Concep’s Quiz 20 Case_Study 22 Puzzles in Operations Research 24 2 Linear Programming: Applications and Model Formulation 21 Introduction 28 22. Structure of Linear Programming Model 28 * General Structure of an LP Model 28 + Assumptions of an LP Model 29 23 Advantages of Using Linear Programming 29 24 Limitations of Linear Programming 30 25 Application Areas of Linear Programming 30 26 General Mathematical Model of Linear Programming Problem 31 2.7 Guidelines on Linear Programming Model Formulation 32 28 Examples of LP Model Formulation 33 + Examples on Production 33» Examples on Marketing 44 » Examples on Finance 48 + Examples on Agriculture $4 « Example on Trarsportation $6 + Examples on Personnel Conceptual Questions 59 Self Practice Problems 60 ‘Hints_and_Answers 66 Chapter Summary 69 Chapter Conceps Quiz 69 vi 1-26 27-70 37Contents Chapter Chapter 4 Linear Programming: The Sim Chapter 3 Linear Programming: The Graphical Method 71-102 3.1 Introduction 72 32 Important Definitions 72 33 Graphical Solution Methods of LP Problems 72 + Extreme Point Solution Method 13 + Examples on Maximization LP Problem 73 + Examples on Minimization LP Problem 78 » Examples on Mized Constraints LP Problem 81 « Iso-profit (Cost) Function Line Method 88 + Comparison of Two Graphical Solution Methods 89 34 Special Cases in Linear programming 90 + Alternative (or Multiple) Optimal Solutions 90 + Unbounded Solution 91 + Infeasible Solution 93» Redundancy 95 Conceptual Questions 95 ‘Self Practice Problems 9§ Hints_and_Answers _99 (Chapter Summary 101 Chapter Concepts Quiz_101 41_Iniroduetion 104 42. Standard form of an LP Problem 104 4.3 Simplex Algorithm (Maximization Case) 106 44 Simplex Algorithm (Minimization Case) 115 Two-Phase Method IIT + The Big-M Method 122 Selr Practice problems A 130 ‘Hinis_and_Answers 133 45 Some Complications and Their Resolution _135 ‘= Unrestricted Variables 135 + Tie for Entering Basie Variable (Key Column) 137 = Me for L Basic Variable (Key Row) ~ Degeneracy 138 46 Types of Linear Programming Solutions 139 = Altemative (Multiple) Optimal Solutions 139 © Unbounded Soluion 140 + Infeasible Solution 141 ‘Conceptual Questions 142 ‘Self Practice Problems B 143 ‘Hints_and_Answers 144 Chapter Summary _145 ‘Chapter Concepts Quiz 145 Case Study 147 5 Duality in Linear Programming 149-172 .1__Introduction 150 52 Formulation of Dual Linear Programming Problem 150 « Syrmetrical Form 150» Econonic Interpretation of Duol Voriobles_\51 + Economic Interpretation of Dual Constrains 152 + Rules for Constrcting the Dual from Primal ‘Self Practice Problems A156 Hints_and Answers 156 53 Sundaxd Resuls on Duality 157 + Printiple of Comblementary Slackness 181 54 rial Significance of Dual 55 _Advaniages of Duality 163 Conceptual Questions 168 Self Practice Problems B 164 ‘Mims_and_Answers_166 ‘Chapter Summary 168 Chapter Concepts Quiz 168 ry ‘Theorems of Duality 170Contents Chapter 6 Sensitivity Analysis in Linear Programming 173-204 6 soduction 174 62__ Sensitivity Analysis _174 ‘s Change in Objective Function Coefficient (cj) 174 = Change in the Availability of Resources (6) 181 © Change inthe Inpur-Out Coeficienis (ays) 8 + Addition of a New Variable 217 = Addition of a New Variable (Column) 191 + Addition of a New Consirain (Row) 194 Conegpival_ Questions 200 ‘Self Practice Problems —200 ‘Hints_and_Answers 203 Chapter Summary 203 Chapter Concepts Quiz _203 Chapter 7 Integer Linear Programming 205-238 21_Introduction 206 72 Types of Imeger Programming Problems 206 73 Enumeration ard Cutting Plane Solution Concept 207 74 Gomory’s All Integer Cutting Plane Method 207 ‘= Method for Constructing Additional Constraint (Cut) 208 = Steps_of Gomory’ All_Inieger Programming Algorithm 208 Self Practice Problems A” 216 ‘Hints_and_Answers 219 75 Gomony’s Mixed-Inieger Cutting Plane Method 220 = Method for Constructing Additional Constraint (Cut) 220 = Steps of Gomory’ Mixed Integer Programming Algoritln 222 26 Tanck and Bound Method 225 7.7 Applications of Zero-One Integer Programming _232 + Capital Budgeting Problen 232 + Fixed Cost (or Charge) Problem 232 ‘= Plant Location Problem 233 Conceptual Questions 234 ‘Self Practice Problems B 234 ‘Hiin's_and_Answers 235 Chapter Summary 236 Chapter Concepts Qui Case Study 237 Chapter 8 Goal Programming 239-258 81_Introduction 240 32 _ Difference Beween LP and GP Approach 249 83 Concept of Goal Programming 240 « Distinction Among Ohjectives, Goals and Constraints 241 84 Goal Programming Model Formulation 241 * Single Goal with Multiple Subgoals 241 + Equally Ranked Multiple Goals 242 = Ranting and Weighting of Unequal Multiple Goals 243 = Algorithy to Formulate GP Model 244 + General GP Model 244 ‘Graphical Solution Method for Goal Programming 244 Ke Modified Simplex Method of Goal Programming 74% 37 _Altemative Simplex Method for Goal Programming 250 ‘Conceptual Questions 253 ‘Self Practice Problems 253 Chapter Summary _255 ‘Chapter Concepts Quiz _256 Case Study 257 Chapter 9 Transportation Problem 259-312 21_Inireduction 260 92 Mathematical Model of Transportation Problem _260 236 CopyrightContents Chapter Chapter ‘+ General Mathematical Model of Transportation Problem 261 93 The Transportation Algorithm 262 9.4 Methods for Finding Initial Solution 262 ‘= North-West Corner Method (NWCM) 262. « Least Cost Method (LCM) 263 = Vogel's Approximation Method (VAM) 268 Conceptual Questions A 268 ‘Self Practice Problems A268 ‘Hinis_and_Answers 268 95 Test for Optimality 269 ~ Dual of Transporation Model 269 + Economie Inerpretation of u,¥ and vjs 210 {Sep of MOD) Method (Wrantporiaion Algor) “21 = Close-loop in Transportation Table and lis Properties 272 Conceptual Questions B28} ‘Self Practice Problems B_ 281 ‘Hinls_and_Answers 283 96 Variations in Transportation Problem _283 = Unbalanced Supply and Demand 283» Degeneracy and its Resolution 286 = Alternative Optimal Solutions 290 + Prohibited Transportation Routes 293 9.7_Maximization Transportation Problem 297 98 Trans-Shipment Problem 209 Conceptual Questions C_301 ‘Self Practice Problems C301 ‘Hins and Answers 305 Chapter Summary _307 Chapter Concepts Quiz _307 Case Study _ 308) ‘Appendix: Theorems and Results 310 10 Assignment Problem 10.1 Introduction 314 10.2 Mathematical Models of Assignment Problem 314 10.3 Solution Methods of Assignment Problem 315 ‘= Hungarian Method for Solving Assignment Problem 315 Conceptual Questions A321 Self Practice Problems A 321 ‘Hinis_and_Answers 323 10.4 Variations of the Assignment Problem _323 Multiple Optimal Solutions 323 + Maximization Case in yment Problem 323 ‘© Unbalanced Assignment Problem 326 + Restrictions on Assignmenis 326 Conceptual Questions B 330 ‘Self Practice Problems B33 Hims and Answers 332 10.5 A Typical Assignment Problem 333, 106 Travelling Salesman Problem 337 Self Practice Problems C 339 Hinws and Answers 340 Chapter Summary 340 Chapter Concepts Quiz 341 Appendix: Important Results and Theorems 342 11 Decision Theory and Decision Trees 11 Introduction 344 112. Steps of Decision-Making Process 345 113 Types of Decision-Making Environments 345 114 Decision-Making Under Uncertainty 346 313-342 343-390 ‘= Optimism (Maximax or Minimin) Criterion 346 + Pessimism (Maxinin or Minimax) Criterion 346Chapter Chapter Contents + Equal Probabilities (Laplace) Criterion 347 + Coefficient of Optimism (HHurwics) Criterion 34 + Regret (Savage) Criterion 347 Conceptual Questions A 350 Self Practice Problems A 350 Hints and Answers 351 11S. Decision-Making Under Risk 351 + Expectsd Monetary Value (EMV) 351 + Expected Opportunity Loss (EOL) 354 + Expected Value of Perfect Information (EVP) 356 11.6 Posterior Probabilities and Bayesian Analysis 364 Conceptual Questions B 369 Self Practice Problems B 369 Hints and Answers 371 117 Decision Trees Analysis 372 118. Decision-Making with Utilities 383, + Usility Functions 384» Utility Curve 384 * Construction of Utility Curves 384 Self Practice Problems C 386 Hints and Answers 387 Chapter Summary 388 Chapter Concepts Quiz 389 Case Study 390 12 Theory of Games 391-324 124 Introduction 392 122 Two-Peron Zero-Sum Games 393 12.3 Pure Strategies (Minimax and Maximin Principles): Games With Saddle Point 395 + Rules 10 Determine Saddle Point 385 Conceptual Questions A297 Self Practice Problems A 398 Hints and Answers 399 124 Mixed Strategies: Games Without Saddle Point 399 12.5 The Rules (Principles) of Dominance 400 126 Soluion Methods Games without Saddle Point 401 + Algebraic Method 401 Arithmetic Method 409 » Matrix Method 410 * Graphical Method 412 + Linear Programming Method 416 Conceptual Questions B 420 Self Practice Problems B 420 Hints and Answers 422 Chapter Summary 423 Chapter Concepts Quiz 423 13 Project Management: PERT and CPM 425-482 13.1 Introduction 426 132. Basic Differences Between PERT and CPM 426 * Significance of Using PERTICPM 427 133. Phases of Project Management 427 + Project Planning Phase 427 Scheduling Phase 427 Project Control Phase 428 134 PERT/CPM Network Components and Precedence Relationships 428 + Rules for AOA Network Construction 430.» Errors and Dummies in Network 431 Conceptual Questions A 434 Self Practice Problems A 434 135. Critical Path Analysis 436 + Forward Pass Method (For Earliest Event Time) 436 + Backward Pass Method (For Latest Allowable Event Time) 437 + Float (Slack) of an Activity and Event 437 « Critical Path 438 Conceptual Questions B 442 Self Practice Problems B 442 Hints and Answers 445, 136 Project Scheduling with Uncertain Activity Times 445Contents Chapter + Estimation of Project Completion Time 446 Conceptual Questions C450 Self Practice Problems C 450 Mints and Answers 453 137 Project Time-Cost Trade-off 453 + Project Crashing 453 + Time-Cosi Trade-off Procedure 454 Self Practice Problems D462 Hints and Answers 465 138 Updating of the Project Progress 465 19 Resoures Allocation 467 + Resowce Levelling 467 + Resource Smoothing 467 Self Practice Problems E478 Chapter Summary 479 Chapter Concepts Quiz 479 Case Stuly 481 14 Deterministic Inventory Control Models 483-552 4.1 Introduction 484 142. The Meaning of Inventory Control 484 143. Furetional Role of Inventory 485 144 Reesons for Carying Inventory 586 145 Factors Involved in inventory Problem Analysis 487 ‘+ Inventory Cost Components 488 + Demand for Inventory Items 490 + Replenishment Lead Tine 490 + Order Cycle 490 + Lead Time or (delivery lag) 490 + Siock Replenishment 490» Length of Planning Period 490 146 Inventory Model Building 490 + Steps of inventory Model Building 491 + Replenishment Order Size Decisions and Concept of EOQ 491 + Classification of EOQ Models 491 147 Single Item Inventory Control. Models Without Shortages 492 + Model 1(a): EOQ Model with Constant Rate of Demand 492 + Model 1(b): EOQ Model with Different Rates of Demand 495 + Model Ic: Economic Production Quanity Model when Supply (Replenishment) ls Gradual 500 Conceptual Questions A 503 Self Practice Problems A 503 Hints and Answers 505 148 Single Item Inventory Control Models With Shortages 506 + Model HM(a): EOQ Model with Constant Demand and Variable Order Cycle Time 06 + Model 11(6): EOQ Model with Constant Demand and Fixed Reorder Cycle Time $08 + Model H(c): EOQ Model with Gradual Supply and Shortage Allowed 510 Conceptual Questions B S12 Self Practice Problems B. 512 Hints and Answers 513, 149 Muitiftem Inventory Models with Constraints 514 + Model IIN(a): EOQ Mode! with Warehouse Space Constraint S14 + Model I11(): EOQ Modet with Investment Constraint S15 + Model IN): BOO Model with Average Inventory Level Consiraint 516 * Model Il): EOQ Model with Number of Orders Constraint 517 Self Practice Problems C518 14.10 Single Item Inventory Control Models with Quantity Discounts S18 * Model IV : EOQ Model with All Units Quantity Discounts 519 Self Practice Problems D 522 Hints and Answers 523 14.11 Inventory Control Models with Uncertain Demand 525 + Reorder Level with Constant Demand 525» Service Level 526 « Additional Stocks S28 14.12. Information Systems for Inventory Control 530 + The Q-Sjsiem wiih Uncerwin Demand 530Chapter 15. Probabil Chapter Contents + The Q-sysiem with Uncertain Demard and Lead Time $37 * Application of Q-System: Two-Bin System S37_» The P-system with Uncertain Demand 537 + Comparison between Q-sysiem and P-ssstem 540 Conceptual Questions C541 Self Practice Problems E542 Hints and Answers $43 1413 Selective Inventory Control Techniques $44 Conceptual Questions D549 Self Practice Problems F 549 Chapter Summary $50 Chapter Concepts Quiz $50 Case Study 551 tic Inventory Control Models 553-570 151 Introduction $54 152. Instantancous Demand Inventory Control Models without Setup Cost 554 + Model I: Optimal Order Point — Marginel Analysis Approach $55 + Model il: Single Period EOQ Model for Uncertain Demand (Newsboy Problem) SST + Model Iil(a): Instantaneous Demand with Shortages, Discrete Replenishment $87 + Medel I11(3): Instantaneous Demand with Shortages Continuous Replenishment $60 + Model II(c}: Reorder Lead Time Without Set-up Cost Model 561 Conceptual Questions A 563 Self Practice Problems A 563 Hints and Answers S64 153 Continuous Demand Inventory Control Models without Set-up Cost 564 + Medel 1V(a): Continuous Demand, Discrete Replenishment S64 + Model IVI(}): Continuous Demand, Continuous Replenishment 567 154 Instantaneous Demand Inventory Control Model with Setup Cost 568 4+ Madel V- Instantaneous Demand, Continous Replenishment 568 Conceptual Questions B 569 Self Practice Problems B 509 Hints and Answers $70 Chapter Summary 570 16 Queuing Theory 571-624 161 Introduction 572 162 The Structure of « Queuing System S72 + Calling Population Characteristics $73» Queuing Process 37% + Quewe Discipline S76 + Service Process (or Mechanism) 516 163. Performance Measures of « Queuing System 577 + TransiontState and Steady-State S18 « Relotionthips Among Performance Measures 579 164 Probebility Distributions in Queuing Systems 580 * Distribution of Arrivals (Pure Birth Process) 580 + Distribution of Imerarrival Times (Exponentiel Process). $81 + Distribution of Departures (Pure Death Process) $81 + Distribution of Service Times 381 Conceptual Questions A 582 165 Classification of Queuing Models 582 + Solution of Queuing Models 383 166. Single-Server Quesing Models 583 + Model |: (MI) (o4FCES)} Exponential Service ~ Unlimited Queue 583 + Model If: {QM/M/1) + (#/SIRO)) 390 + Model It: ((MM/) : (NIFCS)} Exponential Service ~ Finite or Limited) Queue 591 Conceptual Questions B 592 Self Practice Problems A 593 Hints and Answers 594Contents, Chapter Chapter 167 Multi-Server Queuing Models 595 # Model IV: {(M/Ms) : (o¥®CFS)} Exponential Service ~ Unlimited Queue 595 + Model V: ((WMs) : (NIFCFS)} Exponential Service ~ Limited (Finite) Queue 600 Conceptual Questions C 603 Self Practice Problems B_ 603 Hims and Answers 604 168 Finite Calling Population Queuing Models 604 * Model Vi: (MIM/L) : (MIGD)} Single Server ~ Finite Population (Source) of Arrivals 604 + Model Vit: {(M/M/s) : (MGD)} Multiserver ~ Finite Population (Source) of Arrivals 605 Self Practice Problems C 608 169° Multi-Phase Service Queuing Model 608 + Model Vil: ((M/>E,/I) : (=/FCFS)) Erlang Service Time Distribution with k-Phases 608 Self Practice Problems D. 611 Hinss and Answers 612 16.10 Special Purpose Queuing Models 612 ‘+ Model IX: Single Server, Non-Exponential Service Times Disribution — Unlimited Queve 612 + Model X: Single Server, Constant Service Times ~ Unlimited Queue 613 + Model Xi: Service Level Cost Model 613 Chapter Summary 615 Chapter Concepts Quiz 615 Case Stidy 616 Appendix 16: Probability Distribution of Arsivals and Departures 618 Appendix 16B: Erlangian Service Time Distribution with K-Phases 622 17 Replacement and Maintensnce Models 625-658 17.1 Introduction 626 172 Types of Failure 626 + Gradual Failure 626 + Sudden Failure 626 173. Replacement of Items Whose Efficiency Deteriorates with Time 627 ‘+ Model 1: Replacement Policy for lems Whose Running Cost Increases with Tine and Value of Money Remains Constant During a Period 627 ‘+ Model II: Replacement Policy for items Whose Running Cost Increases with Time but Value of Money Changes with Constant Rate During a Period 634 Conceptual Questions A 640 Self Practice Problems A 640 Hints and Answers 642 174, Replacement of Items that Completely Fail 642 + Individual Replacement Policy 645» Group Replacement Policy 645 Conceptual Questions B 651 Self Practice Problems B 651 Hints and Answers 652 175 Other Replacement Problems 653 + Staffing Problem 653 + Equipment Renewal Problem 654 Self Practice Problems C 656 Hints and Answers 657 Chapter Summary 657 Chapter Concepts Quiz 658 18 Markov Chains 659-686 181 Introduction 660 182 Characteristes of a Markov Chain 650 183 Applications of Markov Analysis 660 184 State and Transition Probabilities. 661 185 Multi-period Transition Probabilities 663 + Steps of Constructing Marie of Transition Probabilities 663 186 Steady-State (Equilibrium) Conditions 672 ' Procedure Jor Determining Steady-State Condition 673 187 Absorbing States and Accounts Receivable Application 618Chapter Chapter Chapter Contents Conceptual Questions 680 Self Practice Problems 680 Hints and Answers 683 Chapter Summary 683 Chapter Concepts: Quiz. 683 Case Study 684 19 Simulation 687-722 191 Introduction 688 192 Simulation Defined 689 193 Types of Simulation 690 194 Steps of Simulation Process 690 195 Advantages and Disadvantages of Simulation 692 196 Stochastic Simulation and Random Numbers 692 + Monte Carlo Simulation 693 + Random Number Generation 693 197 Simulation of Inventory Problems 695 198 Simulation of Queuing Problem: 701 199 Simulation of Investment Problems 706 190 Simulation of Maintenance Problems 707 19.11 Simulation of PERT Problems 711 19.12 Role of Computers in Simulation 713 * General Purpose Programming Languages 713.» Special Purpose Simulation Languages 713 1913 Applications of Simulation 714 Conceptual Questions 714 Self Practice Problems 715 Chapter Summary 718 Chapter Concepts Quiz. 718 Appendix: The Seven Most Frequent Causes of Simulation Analysis Failure and How to Avoid Them 720 20. Sequencing Problems 723-740 20.1 Introduction 724 202 Notations, Terminology and Assumptions 724 203 Processing m Jobs Through Two Machines 725 + The Algoritim 725 Conceptual Questions A 730 Self Practice Problems A 730 Hints and Answers 731 204 Processing m Jobs Through Three Machines 731 + The Algorithm 731 Self Practice Problems B» 733 Hints and Answers 734 205 Processing Jobs Through m Machines 734 206 Processing Two Jobs Through m Machines 736 Conceptual Questions B 739 Self Practice Problens B 739 Hints and Answers 739 Chapter Summary 739 Chapter Concepts Quiz 740 21 Information Theory 741-760 211 Introduction 742 212 Communication Processes 742 + Memoryless Channel 743» The Channel Matrix 743 + Probability Relation in a Channel 743 + Noiseless Channel 744 213 A Measure of Information 744 « Properties of Entropy Function, H_ 74S 214 Measures of Other Information Quantities 747Contents Chapter Chapter Chapter + Marginal and Joint Entropies 747 © Conditional Entropies 748 + Expected Mutual Information 750. + Axiom of an Entropy Function 751 + Basic Requiremens of Logaritimic Eniropy Functions 751 21.5 Channel Capacity, Efficiency and Redundancy 753 + Channel Capacity 753 « Efficiency 753 « Redundancy 753 21.6 Encoding 754 + Objectives of Encoding 784 21.7 Shannon-Fano Encoding Procedure 755 218 Necessary and Sufficient Condition fer Noiseless Encoding 757 Conceptual Questions 758 Self Practice Problems 759 Hints and Answers 759 Chapter Summary 760 Chapter Concepts Quiz 760 22. Dynamic Programming 761-796 221 Introduction 762 222. Dynamic Programming Terminology 762 23 Developing Optimal Decision Policy 763 + The General Algorithn 764 24 Dynamic Programming under Certainty 764 * Mode! I: Shortest Route Problem 164 ‘+ Model Il: Multiplicative Separable Return Function and Single Additive Constraint 761 ‘© Model Ill: Additive Separable Retwn Function and Single Additive Constraint 774 ‘+ Model IV. Additively Separable Return Function and Multiplicative Constraint 783 225 Dynamic Programming Approach for Solving Linear Programming Problem 788 Conceptual Questions 790 Self Practice Problems 791 Hints and Answers 194 Chapter Summary 795 Chapter Concepts Quiz 796 23° Classical Optimization Methods 797-820 23.1 Introduction 798 232 Unconsirained Optimization 798 + Optimizing Single-Variable Functions 798 ‘+ Conditions for Local Minimum and Maximum Value 799 + Optimizing Multivariable Functions 803 Self Practice Probiems A 806 Hin and Answers 807 233 Constrained Multivariable Optimization with Equality Constraints 808 + Direct Substitution Method 808 + Lagrange Multipliers Methods 809 Self Practice Probiems B_ 814 Hints and Answers 815 234 Constrained Multivariable Optimization with Inequality Constraints 815 + KuhnTucker Necessary Conditions 815 + Kuhn-Tucter Sufficient Conditions 816 Conceptual Questions 819 Self Practice Problems C 819 Hints and Answers £20 Chapter Summary 820 Chapter Concepts Quiz #20 24 Non-Linear Programming Methods 821-862 24.1 Introduction 822 2 The General Non-Linear Programming Problem 824 243. Graphical Solution Method $24Chapter Chapter Contents ‘Self Practice Problems A_ 827 Hints and Answers 528 244 Quadratic Programming 829 + Kuhn-Tucter Conditions 829 + Wolfe's Modified Simplex Method $30 * Beales Method $35 24S Applications of Quadratic Programming 841 Conceptual Questions A 844 Self Practice Probiems B_ 844 Hints and Answers $45 246 Separable Programming 845 + Separable Functions 843 + Definitions 846 + Pieceswise Linear Approximation of Non-linear Functions 846 ‘+ Mixed-Integer Approximation of Separable NLP Problem 847 + Steps of Separable Programming Algorithm 848 Conceptual Questions B 852 Self Practice Problems C853 Hints and Answers 853 247 Geometic Programming 853 + General Mathematical Form of GP 853 + Primal GP Problem with Equality Constrains 857 248 Stochastic Programming 859 + Sequential Stochastic Programming 860 + Now Sequential Stochastie Programming $60 * Chance-Constrained Programming 860 Self Practice problems D861 Hints and Answers $62 Chapter Summary 862 25 Theory of Simplex Method 863-884 25.1 Introduction 864 252 Canonical and Standard Form of LP Problem 864 283 Slack and Surplus Variables 865 * Basic Solution 866 + Degenerate Solution 866 + Cost (or Price) Vector 867 Conceptual Questions A 869 Self Practice Problems A 869 Hints and Answers 869 254 Reduction of Feasible Solution to a Basic Feasible Solution 870 255 Improving a Basic Feasible solution 876 2546 Alternative Optimal Solutions 879 28.7 Undounded Solution 879 258 Optimality Condition 880 259 Some Complications and Their Resolution 880 + Unrestricted Variables 881 * Degeneracy and its Resolution $81 Conceptual Questions B 883 Self Practice Problems B 884 Hints and Answers 884 Chapter Summary 884 26 Revised Simplex Method 885-900 26. Inioduetion 886 262. Standard Forms for Revised Simplex Method 886 ‘ Revised Simplex Method in Standard Form | 886 263. Computational Procedure for Standard Form I 888 = Steps of the Algorithm $89 264 Comparison of Simplex Method and Revised Simplex Method $99 Conceptual Questions 909 Self Practice Problems 900 Hints and Answers 900 Chapter Summary 900Contents Chapter 27 Dual-Simplex Method 271 Inwoduetion 902 27.2 Dual-Simplex Algorithm 902 Conceptual Questions 908 Self Practice Problems 908 Hints and Answers 908 Chapter Summary 908 Appendix: Theory of Dual-Simplex Method 909 Chapter 28 Bounded Variables LP Problem 28.1 Introduction 912 282 The Simplex Algorithm 912 Self Practice Problems 920 jmts and Answers 920 Chapter Summary 920 Chapter 29 Parametric Linear Programming 29.1 Introduction 922 292 Variation in the Objective Function Coefficients 922 293 Variation in the Availability of Resources (RHS Values) 927 Conceptual Questions 951 Self Practice Problems 932 Hinis and Answers 932 Chapter Summary 932 Appendix A: Pre-Study for Operations Research AL Linear Combination of Vectors 934 A2 Linear Dependence and Independence 934 A3- Simultaneous Linear Equations and Nature of Solution 936 AA Convex Amlysis 937 A. Supporting and Separ ‘AG Convex Function 941 AT Quadratic Forms 942 Self Practice Problems 943 Hyperplares 940 Appendix B: Selected Tables Table 1 Values of e* and &* 946 Table 2 Poisson Distribution 947 Table 3 Normal Distribution 949 Table 4 Random Numbers 950 Table $ Present Values 951 Table © Cumulative Poisson Probabilities 952 Selected References 901-910 911-920 921-932 933-944 945-952 953-954Operations Research An Introduction “The first rule of any technology used in a business is that ‘automation applied to an efficient operation will magnify the ‘efficiency. The second is hat automation applied 10 an inefficient ‘operation will magnify the inefficiency.” ~ Bill Gates Preview This chapterpresents a framework ofa possible structural analysis of problems pertaining to an organization in order to arrive at an optimal solution using operations research approach. Learning Objectives After studying this chapter you should be able to ‘+ understand the need of using operations research ~a ‘quantitative appreach for effective decision-making ‘© know the historical perspective of operations research approach, ‘© know the various definitions of operations research, its characteristics and various phases of scientific study. ‘© recognize, classify and use various models for solving a problem under consideration. be familiar with several computer software available forsolving ‘an operations research model chapter Chapter Outline 44 12 13 14 18 18 18 19 4.10 1m 442 113 1.14 416 cooo Operations Research - A Quantitative Approach to Dexsion making The History of Operations Reseerch Deiintions of Operations Research Features of Operations Research Approach ‘Operatis Research Approach to Problem Sohwng + Conceptual Questions A Models and Modeling in Operations Research ‘Advantages f Model Bung Methods for Solving Operations Research Models Metnodclogy of Operations Research Advantages cf Operations Research Susy Opporturities and Shorcomings cf the Opertions Resoarch Asproach Fecturee of Operations Research Solusion ‘Applications of Operations Research Operations Research Models in Practice Computer Sottwaro ler Operations Research + Conceptual Questions B Chapter Summary Chapter Concopte Quiz Cato Study Puzies in Operatons Research opytightOperations Research: Theory and Applications Quantitative analysis is the Sslentiie approach to decision-making Decision maker ‘should consider both qualitative and ‘quantitative factors while solving a problem 1.1 OPERATIONS RESEARCH - A QUANTITATIVE APPROACH TO DECISION-MAKING Decision-making in today’s social and business environment has become a complex task. High costs of technology, materials, labour, competitive pressures and s0 many different economic, social as well as political factors and viewpoints, greatly increase the difficulty of managerial decision-making. Knowledge and technology are changing rapidly and are continuously giving rise to problems with little or no precedents. Well-structured problems are routinely optimized at the operational level of organizations, and increased aitention is naw focussed on broader tactical and strategic issues. To effectively address the arising problems and to provide leadership in the advancing global age, decision-makers cannot afford to make decisions by simply applying their personal experiences, guesswork or intuition, because the consequences of wrong decisions can prove to be serious and cosily. Hence, an understanding of the applicability of quartitative methods to decision-making is of fundamental importance to decision-makers For example, entering the wrong markets, producing the wroag products, providing inappropriate services, ete., may cause disastrous consequences for organizations. In the global age, the practice of the operations research (OR) approach must maintain stride with the abovementioned trends. Some people claim that the OR approach does not adequately meet the needs of business and industry. The reasons for its failure are often behavioural in nature. Lack of implementation ‘of results of findings is one of the major reasons. The implementation process presumes that the definition, analysis, modeling, and solution phases of a project have been adequately performed. Among the reasons for implementation failure is the lack of creative problem solving abilities of the decision-maker. Operations research facilitates the comparison of all possible alternatives (courses of action oF act. This helps to know the potential outcomes and permits examination of the sensitivity of the solution to changes or errors in numerical values. It also encourages rational decision-making based on the best available approzches and/or techniques. However, the fact that timely and competent decisions should be ‘an aid to the decision-maken's judgement, not a substitute for it, should be emphasized. That is not to say that the management decision-making is simply about the application of operations research techniques and/or approaches. ‘While solving a real-life problem, the decision-maker must examine a problem from both the quantizative as well as the qualitative perspective. This should be done so that data so obtained, be analyzed from both perspectives in order to suggest a solution to the problem. For example, consider the problem of an investor seeking advice for investments in three alternatives: Stock Market, Real Estate and Bank Deposit. To suggest an acceptable solution, we need to consider certain quantitative factors that should be examined in the light of the problem. For instance, factors like financial ratios from the balance sheets of several companies whose stocks are under consideration; real estate companies” cash flows and rates of retum for investment in property; and how much the investment will be worth in the future when deposited at a bank at a given interest rate for a certain number of years, are factors that need to be examined. However, before reaching a conclusion, certain other qualitative factors, such as weather conditions, state and central policies, new technology, the political situation, etz., also need to be considered. ‘The evaluation of each alternative can be extremely difficult or time consuming for two reasons: First, due to the amount and complexity of information that must be processed, sevond due to the number of alternative solutions available. The number of solutions can be so large that a decision-maker simply would not be able to evaluate all of them in order to select an appropriate one, For these reasons, when there isa lack of qualitative factors, decision-makers increasingly tur to quantitative factors and use computers to arrive at the optimal solution for problems that involve @ large number of altematives. ‘There is an imperative need for an analytical research for presenting a siructural analysis. This can be done by critically examining the levels of interaction between the application process of operations research, various systems and organizations. Figure 1.1 summarizes the operations research approach by sketching the conceptual framework of the main elements of the analysis. This book introduces a set of operations rescarch techniques that would help decision-makers in making rational and effective decisions. It also gives a basic knowledge of mathematics and statistics, as well as of the use of computer software needed for computational purposes.image not availableimage not availableimage not availableOperations Research: Theory and Applications. Operations research ( coorene (i) hoist, and th) oectio- ofiented. approaches te decsion makng Operations research ‘atompte to resolve the conflicts of interest among various secions of the organization and ‘seeks to find the ‘optima solution that 's in the intrest of the organization as a whole + Operations research is a sciemific approach to problem-solving for executive management. ~ HM Wagner {As the discipline of operations research grew, numerous names such as Operations Analysis, Systems Analysis, Decision Analysis, Management Science, Quantitative Analysis, Decision Science were given to it This is because of the fect that the types of problems encountered are always concemed with “effective decision’, but the solution of these problems do not always involve research into operations or aspects of the science of management. 1.4 FEATURES OF OPERATIONS RESEARCH APPROACH ‘The broad based definition of OR, with the additional features is as follows: OR utilizes a planned approach following a scientific method and an interdisciplinary team, in order to represent complex functional relationship as mathematical models. for the purpose of providing a quantitative basis for decision- ‘making and uncovering new problems for quantiative analysis. The broad features of OR. approach to any decision problem are summarized as follows Interdisciplinary Approach For solving a problem interdisciplinary teamwork is essential. This is ‘because while attempting to solve a complex management problem, one percon may not have the complete knowledge of all its aspects (such as economic, social, political, psychological, engineering, etc.). This means we should not expect one person to find a desirable solution to all managerial problems. Therefore, a team of individuals specializing in mathematics, statistics, economics, engineering. computer science, psychology, etc, should be organized in a way that each aspect of the problem can be analysed by a particular specialist in that fleld. This would help to arrive to an appropriate and desirable solution of the problem. However, there are certain problem situations that can be analysed by even one individual. Scientific Approach Operations research is the application of scientific methods, techniques and tools to problems involving the operetions of systems so as to provide those in control of operations with optimum solutions to the problems (Churchman et al.). The scientific method consists of observing and defining the problem: formulating and testing the hypothesis: and analysing the results of the test. The data so obtained is then used to decide whether the hypothesis should be accepted or not. If the hypothesis is accepted, the resulls should be implemented otherwise an altemative hypothesis has to be formulated Holistic Approach While arriving ata decision, an eperations research team examines the relative importance of all conflicting and multiple objectives. It also examines the validity of cleims of various epartmenss of the organization from the perspective of its implications to the whole organization. Objective-Oriented Approach An operations research approach seeks 10 obtain an optimal solution to the problem under analysis. For this, a measure of desirability (or effectiveness) is defined, based ‘on the objective(s) of the: organization. A measure of desirability so defined is then used to compare alternative courses of action with respect to their possible outcomes, Mlustration The OR eppreach attempts to find global optimum by analysing interrelationships among the system components involved in the problem. One such situation is deseribed below. Consider the case of a large organization that tas 2 number of management specialists but the organization is not exactly very well-coordinated. For example its inability to properly deal with the basic problem of maintaining stocks of finished goods. To the marketing manager, stocks of a large variety of products are purely a means of supplying the company’s customers with what they want and when they want it, Clearly, according to a marketing manager, a fully stocked warehouse is of prime importance to the ‘company. But the production manager argues for long production runs, preferably on a smaller product range, particularly if a significant amount of time is lost when production is switched from one variety to another. The result would again be a tendency to increase the amount of stock carried but itis, of course, Vital that the plant should be kept running. On the other hand, the finance manager sees stocks in terms of capital that is unproductively tied up and argues strongly for its reduction. Finally, there appears the personnel manager for whom a steady level of production is advantageous for having better labour relations. Thus, all these people would claim to uphold the interests of the organization, but they do so only from their own specialized points of view. They may come up with contradictory solutions and obviously, all of them cannot be rightimage not availableimage not availableimage not available‘Operations Research: Theory and Applications Fig. 12 Classification of Models The relationship among velocity, distance and acceleration is an example of a mathematical model. In accounting, the cost-volume-profit model is also an example of a mathematical model. Symbolic models are precise and abstract and can be analysed and manipulated by using laws of iathematics. The models are more explanatory rather than descriptive, Benavicur Characters ‘conaity eee] =o — sein Prcedize 1.6.2. Classification Based on Function or Purpose Models based on the purpose of their utility include the following types: Descriptive models Descriptive models characterize things as they are. The major use of these models is to investigate the outcomes or consequences of various alternative courses of action. Since these models cheek the consequence only for a given condition (or alternative) rather than for all conditions, there is no guarantee that an altemative selected with the aid of descriptive analysis is optimal. These models are usually applied in decision situations where optimizing models are not applicable. They are also used when the final objective is to define the problem or to assess its seriousness rather than to select the best altemative. These models are especially used for predicting the behaviour of a particular system under various conditions. Simulation is an exemple of a descriptive technique for conducting experiments with the systems. Predictive models These models indicate the consequence ~ ‘if this occurs, then that will follow’. ‘They relate dependent and independent variables and permit the trying out, of the ‘what if” questions. In other words, these models are used to predict the outcomes of a given set of alternatives for the problem. These models do not have an objective function as a part of the model of evaluating decision alternativesimage not availableimage not availableimage not availableOperations Research: Theory and Applications Exogenous vvarisblee are those whose numerical value depends upon the external envionment that influences an ‘organization Optimization methods yield the best values for the decision variables both for unconstrained and ‘constrained probiems Model construction consists of hypothesizing relationships between variables subject to and not subject to control by decision-maker. Certain basic components required in every decision problem model are Controllable (Decision) Variebles These arc the issues oF factors in the problem whose valucs arc to bbe determined (in the form of numerical values) by solving the model. The possible values assigned to these variables are called decision aliernatives (strategies or courses of action). For example, in queuing theory, the number of service facilities is the decision variable. Uncontroliable (Exogenous) Variables These are the factors. The values of these variables are not under the control of the decision-maker and are also termed as state of nature, Objective Function (or Performance Measures) This is a representation of (i) the criterion that expresses the decision-maker’s manner of evalusting the desirability of alternative values of the deci variables, an¢ (ii) how that criterion is to be optimized (minimized or maximized). For example, in queuing theory the decision-maker may consider several criteria such as minimizing the average waiting time of customers, of the average number of customers in the system at any given time. Policies and Constraints (or Limitations) These are the restrictions on the values of the decision variebles. These restrictions can arise due to organization policy, legal restraints or limited resources such a5 space, money, manpower, material, tc. The constraints may be in the form of equations or inequalities. Funstional Relationships In a decision problem, the decision variables in the objective function and in the constraints, ae connected by a specific functional relationship. A general decision problem modal might take the following form: Optimize (Max. or Min.) Z= f(x) subject to the constraints 8) [5-21 Bs m and x20 where, x= a vector of decision variables (x), x, -. 4 X,) ‘f(8) = criterion or objective function to be optimized ,(8)= the ith constraint 8, A model is referred to as a linear model if all functional relationships among decision variables Xyp Ay» son, in /(@) and g (x) are of a linear form, But if one or more of the relationships are non-linear, the Model is'said to be a non-hnear model. Parameters These are constants in the functional relationships, Parameters can either be deterministic or probabilistic in nature. A deterministic parameter is one whose value is assumed to occur with certainty. However, if constants are considered directly or explicitly es random variables, they are probabilistic paramete's ‘Step 3: Solving the Mathematical Mode! Once # mathematicel model of the problem hus been formulated, the next step isto solve it, that is, to obtain numerical values of decision variables. Obtaining these values depends on the specific form or type, of mathematical model. Solving the model requires the use of various methematical tools and numerical procedures. In general, the following two categories of ‘methods are used for solving an OR model. @ In constrained problems, these values simultaneously satisfy all of the constraints and provide an ‘optimal or acceptable value for the objective function or measure of effectiveness. (® However. these values provide an acceptable value for the objective function, Heuristic methods are sometimes described as rules of thumb which work, An example of a commonly used heuristic is ‘stand in the shortest line’. Although using this rule may not work if everyone in the shortest line requires extra time, in general, it is not a bad rule to follow. These methods are used when obtaining optimal solution is either very time consuming or the model is too complex. ‘Sometimes difficulties in problem solving arise due to lack of an appropriate methodology for it and psychological perceptions on the part of the problem solver. The major difficulties in problem solving can be grouped into the following categories: ( Failure to recognize the existence of a problem + Some people tend to personalize problems * Information is not received to signal the fact that a problem exists fixed amount of the ith resourceimage not availableimage not availableimage not availableOperations Research: Theory and Applications If the measure of effectiveness such as profit, cost, etc, is represented as a linear function of several ‘variables and if limitations on resources (constraints) can be expressed as a system of linear equalities or inequalities, the allocation problem is classified as a linear programming problem. But if the objective function of eny or all of the constraints cannot be expressed 2s a system of linear equalities or inequalities, the allocation problem is classified s a non-linear programming problem. ‘When the solution values or decision variables of a problem are restricted to being integer values or just zero-one values, the problem is classified as an integer programming problem or a zero-one programming problem, respectively. ‘A problem having multiple, conflicting and incommensurable objective Functions (goals) subject to linear constraints is called a goal programming problem. If the decision variables in the linear programming problem depend on chance the problem is called a stochastic programming problem. If resources such as workers, machines or salesmen have to be assigned to perform a certain number of activities such as jobs or territories on a one-to-one basis so as to minimize total time, cost or distance involved in performing a given activity, such problems are classified as assignment problems. But ifthe activities require more than one resource and conversely ifthe resources can be used for more than one activity, the allocation problem is classified as a trarsporaation problem. + Inventory models Inventory models deal with the problem of determination of how much to order at «point in time and when to place an order. The main objective is to minimize the sum of three contlicting inventory costs: The cost of holding or carrying extra inventory, the cost of shortage or delay in the delivery of items when it is needed and the cost of ordering or Set-up. These are also useful in dealing with quantity discounts and selective inventory control. ‘© Waiting line (or Queuing) models These models have been developed to establish « trade-off between costs of providing service and the waiting time of a customer in the queuing system. Constructing a mode! entails describing the components of the system: Arrival process, queue structure and service process and solving for the messure of performance ~ average length of waiting time, average time spent by the customer in the line, traffic inter ‘+ Competitive (Game Theory) models These models are used to characterize the behaviour of two or ‘more opponents (called players) who compete for the achievement of conflicting goals. These models are classified according to several factors such as number of competitors, sum of loss and gain, and the type of siratezy which would yield the best or the worst outcomes. ‘= Network models These models are applied to the management (planning, controlling and scheduling) of large-scale projects. PERT/CPM techniques help in identifying potential trouble spots in a project through the identification of the critical path. These techniques improve project coordination and enable the efficient use of resouress. Network methods are also used to determine time-cost trade-off, resource allocation and help in updating activity time. + Sequencing models The sequencing problem arises whenever there is a problem in determining the sequence (order) in which a number of tasks can be performed by a number of service facilities such as hospital, plant, etc. in such a way that some measure of performance, for example, total time to process all the jobs on al the machines, is optimized. ‘+ Replacement models These models are used when one must decide the optimal time to replace an ‘equipment for one reason or the other ~ for instance, in the case of the equipment whose efficiency deteriorates with time or fails immediately and completely. For example, in case of an automobile, the user has his own measure of effectiveness. So there will net be one single optimal answer for everyone, even if each automobile gives exactly the same service. ‘* Dynamic programming models Dynamic programming may be considered as an outgrowth of mathematical programming, involving the optimization of multistage (sequence of interrelated decisions) decision processes. The method starts by dividing a given problem into stages or sub-problems and then solves those sub-problems sequentially until the solution to the original problem is obtained. ‘© Markov-chain models These models are used for analysing a system which changes over a period of time among various possible outcomes or states. The model, while dealing with such systems, describes transitions in terms of transition probabilities of various states. These models have been used to test brand-loyalty and brand-switching tendencies of consumers, where each system state is considered ta be a particular brand purchaseimage not availableimage not availableimage not availableimage not availableimage not availableimage not availableimage not availableOperations Research: Theory and Applications, Letter Values A P- E- Total — Group | = ——— R- F. s- Total — Group 2 = ——— He TT o- Scere — Hint The solution of this equel-group-total-assumption formulation bas the leter values A - 6, E ~ 7, F~ 2, H=1,0—8,P-4,R—5,S—3, T—9 with equal group touls of 139, If this formulation had no feasible solution, then it would be necessary to solve a sequence of problems in which the right-hand side ofthe first constrait is varied ‘with increasing positive and negative integer values until a feasible solution is found.image not availableimage not availableimage not availableOperations Research: Theory and Applications 5. Linear programming also helps in the re-evaluation ofa basic plan for changing conditions. Ifeondivions change when the plan is partly carried out, they can be determined so as to adjust the remainder of the plan for best results. 2.4 LIMITATIONS OF LINEAR PROGRAMMING In spite of having many edvantages and wide areas of applications, there are some limitations associated with this technique. These are given below: 1. Linear programming trea's all relationships among decision variables as linear. However, generally, neither the objective functions nor the constraints in real-life situations conceming business and industrial problems are linearly related to the variables, 2 While solving an LP model, there is no guarantee that we will get integer valued solutions. For example, in finding out how many men and machines would be required to perform a particular job. non-integer valued solution will be meaningless. Rounding off the solution to the nearest integer will not yield an optimal solution. In such cases, integer programming is used to ensure integer value to the decision variables 3. The linear programming model does not take into consideration the effec of time and uncertainty. Thus, the LP model should be defined in sucha way that any change due to internal as well as external factors can be incorporated. 4. Sometimes large-scale problems can be solved with linear programming techniques even when the assistance of a computer is available. For this, the main problem ean be divided into several smal problems and each one of them can be solved separately 5. Parameters appearing in the model are assumed to be constant but in real-life situations, they are frequently neither known nor constant. 6 Itdeals with only single objective, whereas in real-life situations we may come across conflicting multi- objective problems. In such cases, instead of the LP model, a goal programming model is used to get satisfactory values of these cbjectives 2.5 APPLICATION AREAS OF LINEAR PROGRAMMING Linear programming is the most widely used technique of decision-making in business and industry and in various other fields. In this section, we will ‘iscuss a few of the broad application areas of linear programming. Agricultural Applications ‘These applications fall into categories of farm economics and farm management. The former deals with the agricultural economy of anation ora region, while the latter is concerned with the problems of the individual farm, ‘The study of farm economics deals with interregional competition and optimum allocation of crop production. Efficient production patterns can be specified by a linear programming model under regional, land resources and rational demand constraints. Linear programming can be applied in agricultural planning, e.g. allocation of limited resources such 1s acreage, labour, water supply, working capital, ete., in a way 0 as to maximize the net revenue. Military Applications Military applications include the problem of selecting an air weapon system against the enemy so keep them pinned down, and at the same time ensuring that the minimum amount of aviation gasoline is ‘used. The applications are also used for varying the transportation in such a way that it maximizes the total tonnage of bombs dropped on a set of targets and takes care of the problem of community defence against disester, the solution of which yields the number of defence units that should be used in a given attack in order to provide the required level of protection at the lowest possible cost. Production Management = Product Mix A company can produce several different products, each of which requires the use of limited production resources. In such cases, it is essential to determine the quantity of each product to be produced, knowing its marginal contribution and amount of available resourc: used by it. The objective is to maximize the total contribution, subject to all constraints, toimage not availableimage not availableimage not availableOperations Research: Theory and Applications to prepare and pack 1,000 gallons of A and Rs 28,000 to prepare and pack a quintal of B, whereas in plant 2 these costs are Rs 18,000 and Ks 26,000, respectively. The company is obliged to produce daily atleast 10 thousand gallons of A and 8 quintals of B. Formulate this problem asan LP modelo find ou as to how the company should organizes production so thatthe required amounts of the two products be obtained ai the minimum cost LP model formulation The data of the problem is summarized as follows: ‘Preparation time (hrs) Plant 1:3 hrvthousand gallons 1 hquintal 6 Plant 2: 2 brsthousand gallons 1.5 hr/uintal 6 ‘Minimam daily production 10 thousand gallons 8 cuintals Cost of production (Rs) Plant I: 15,000/Housand gallons 28,000/quinals Plant 2: 18,00/housand gallors _26,000/qintals Decision variables Let 2X)» Fp = quantity of product A (in *000 gallons) to be produced in plants 1 and 2, respectively. X44 = quantity of product B (in quintals) to be produced in plants 1 and 2, respectively. The LP model Minimize (total cost) 2 15,000x, + 18,000x, + 28,000x, + 26,000x, subject to the constraints @ Preparation time (@) 3x, +516, (0) 2x, + 1.544516 (Minimum éaity production requirement (@ x +x, 10, x, +28 and App Ty X20 Example 23 _An electronic company is engaged in the production of two components C; and C; that are used in radio sets. Each unit of C, costs the company Rs 5 in wages and Rs 5 in material, while each of , costs the company Rs 25 in wages and Rs 15 in material. The company sells both products on one- period credit terms, but the company’s labour and material expenses mast be paid in cash. The selling price of C, is Rs 30 per unit and of C, sis Rs 70 per unit, Because of the company’s strong monopoly in these components, itis assumed that the company can sell, at the prevailing prices, as many units as it produces. ‘The company’s production capacity is, however, limited by two considerations. First, at the beginning of period |, the company has an initial belance of Rs 4,000 (cash plus bank credit plus collections from past credit sales) Second, the company has available in each period 2,000 hours of machine time and 1,400 hours of assembly time. The production of each C, requires 3 hours of machine time and 2 hours of assembly time, whereas the production of each C, requires 2 hours of machine time and 3 hours of assembly time. Formulate this problem as an LP model so as to maximize the total profit to the company. LP model formulation The data of the problem is summarized as follows: Budget (Rs) 1Orait 40iunie Rs 4000 Machine time 3 hvunit 2 hrsumit 2,000 hours ‘Assembly time 2 bsiunit 3 hesfuit 1400 hours Selling pice Re 30 Re 70 Cost (wages + material) price RS 10 Rs 40 Decision variables Let x, and x, = number of units of components C, and C, to be produced, respectively. The LP motel Maximize (total profi) Z = Selling price ~ Cost price (30~ 10) x, + (70-40) x = 20x, + 30%,image not availableimage not availableimage not availableET Operations Resoarch: Theory and Applications to maximize his profit, subject of course, to all the production and marketing restrictions. Formulate this problem as an LP model to maximize toval profit. ‘Machine A time 10 minte 8 « 5 60 ~ 2,400 minutes ‘Machine B time ‘60;mimite 8 «5 * 60 = 2,400 minutes Production 1 1 300,000 units'week Marketing 1 x 25,000 units work = + 7,000 unitsiweek Profitanit (Rs) 025 Decision variables Let x, and x, = units of 8-ounce and 16-ounce bottles, respectively to be produced weekly. The LP model Maximize (total profit) Z=0.15r, + 0.25: subject to the constraints © Machine time : (@) *2 < 2,400 and (b) +2 < 2,400 100" 40 60" 75 (@ Production : x, +x $300,000 i) Marketing (@) x, $25,000, (b) x, $7,000 and p20. Example 2.9 A company, engaged in producing tinned food has 300 trained employees on its rolls, each of whom can produce one can of food in a week. Due to the developing taste of public for this kind of food, the company plans to add to the existing labour force, by employing 150 people, in a phased manner, over the next five weeks. The newcomers would have to undergo a two-week training programme before being, put to work. The training is to be given by employees from among the existing ones and it is a known fact that one employee can train three trainees, Assume that there would be no production from the trainers and the trainees during training period, as the training is offthe-job. However, the trainees would be remunerated atthe rate of Rs 300 per week, the same rate would apply as for the trainers. ‘The company has booked the following orders to supply during the next five weeks: ‘Week 1 2 3 4 5 No. of cans : 280 298 305 360 400 Assume that the production in any week would not be more than the number of cans ordered for so that every delivery of the food would be ‘fresh’ Formulate this problem as an LP model to develop a training schedule that minimizes the labour cost over the five-week period. [Delhi Univ, MBA, Nov. 1998, 2000] LP model formulation The data of the problem is summarized as given below: @ Cans supplied Week : 1 2 3 4 4 Number: 280 298 «= 305360400 i) Each trainee has to undergo a two-week training. Gi) One employee is required to train three trainees. (ix)__Every trained worker produces one can/week but there would be no production from trainers nd. trainees during training. (8) Nuraber of employees to be employed = 150 (Wi) The production in any week is not to exceed the cans required. (vi) Number of weeks for which newcomers would be employed: 5, 4,3, 2, 1. From the given information you may observe following facts: (© Workers employed at the beginning of the first week would get salary forall the five weeks; those employed at the beginning of the second week would get salary for four weeks and so on. (©) The value of the objective function would be obtained by multiplying it ty 300 because esch person would get a salary of Rs 300 per week.image not availableimage not availableimage not availableOperations Research: Theory and Applications LP model formulation Let x; 7 quantity of Venus (kilolitres) produced in the company 25 = quantity of Diana (kilolitres) produced in the company 2X3 = quantity of Diana (Kilolitres) produced by hired facilities 2X4 = quantity of Aurora (kilolitres) produced in the company The LP model Maximize (total profit) Z= subject to the constraints @ Special additive 0.30x, + 0.15x, + 0.1Sx, + 0.75x,< 600 ,000x, + 3,500, + (3,500 — 1,000ix, +2,000x, ( Ownmilling fai (ii) Hired milling facility: (iv) Packing (v) Marketing: (i) x 100 (Venus); (ii)x, +x and 15 Xp Xy yO. Example 2.15 Four products have to be processed through a particular plant, the quantities required for the next production period are: Product 1: 2,000 units Product 2: 3,000 units Product 3: 3,000units Product 4 : 6,000 units ‘There are three production lines on which the products could be processed. The rates of production {in units per day and the total available capacity in days are given in the following table. The corresponding cost of using the lines is Rs 600, Rs $00 and Rs 400 per day, respectively 400 (Diana), (ii) 200 <.x, < 600 (Aurore) Formulate this problem as an LP model to minimize the cost of operation. [Delhi Univ., MBA, 1994) LP model formulation Let x, = number of units of product i(/ = 1, 2, 3, 4) produced on production line j (j= 1, 2,3) The LP model 4 4 4 Minimize (total cost) Z = 600 x,, +500 ¥ x,, +400 ¥ x, ft ia i= subject to the constraints (@ Production: (a) Bonu ~ 2,000, ) Ex. 73,000 © 3m @ 3x4 =6,000 @ Line capacity Su 412, *3 Au 22, 3. ® 750 * 100 * 500 Si Sp a 7, 53, tem 200" 100 * 760" 400 ) Bu An, 33. © [60°30 * 0 and x, 20 for all and j Example 2.16 A manufécturer of biscuits is considering making four types of gift packs containing three types of biscuits: Orange cream (OC), Chocolate cream (CC) and Wafers (W). A market research study, recently conducted to assess the preference of the consumers, shows the following types of assortments, to be in good demand:image not availableimage not availableimage not available‘Operations Research: Theory and Applications (iv) Maximum newspaper and magazine advertisement xy ty 50.60 a < ipo or ~ 06x, ~0.6r, +0.4x, + 04x, < 0 (v) Exposure to families with income over Rs 50,000 2,00,000x, + 5,00,000x, +3,C0,000x, + 1,00,000x, 2 45,00,000 and xp xypxy Hy 20. Example 2.20 An advertising agency is preparing an advertising campaign for a group of agencies. These agencies have decided that different characteristics of their target customers should be given different importance (weightage). The following table gives the characteristics with their corresponding importance (weightage). Characteristics Weighage (A) Age 40 years 20 Annual income Above Rs 60,000 30 50 ‘The agency has carefully analyzed of three media and has compiled the following data: ns Data trem Women’s Magazine (25) Reader characterises (Age: 25-40 years 80 70 © (Gi) Anmual income: Above Rs 60,000 co 50 45 (Gil) FemalesMarred 40 35 es Cost per advertisement (Rs) 9,500 25,000 100,000 Minimum number of advertisement allowed 10 3 3 Maximum number of advertisement allowed 20 10 10 ‘Auticuce size (1000s) 750 1,000 1,300 ‘The budge: for launching the advertising campaign is Rs 5,00,000. Formulate this problem as an LP model for the sgency to maximize the total expected effective exposure. LP model formulation Let x,, x, and x, = number of advertisements made using advertising media’ ‘women’s magazines, radio and television, respectively ‘The effectiveness coefficient comesponding to cach of the advertising media is calculated as follows: Maia “Eiectneness Coeiicion: Women’s magazine 0.30 (0.20) + 0.60 (030) + 0.40 (0.50) = 054 Radio 0.70 (020) + 0.50 (0.30) + 035 (00) ~ 0.16 Television 0.60 (020) + 0.45 (030) + 0.25 (0.50) = 038 ‘The coefficient of the objective function, ie. effective exposure for all the three media employed, can be computed as follows: Effective exposure — Effectiveness coefficient ~ Audience s ‘where effectiveness coefficient is a weighted average of audience characteristics. Thus, the effective exposure of each media is as follows: ‘Women’s magazine ~ 0.54 » 7,50,000~4,05,000 Radio = 0.46 x 10,00,000 = 4,60,000 Television =0.38 x 15,00,000 = 5,70,000 The LP modet Maximize (effective exposure) Z = 4,05,000+, +4,60,000x, +5.70.000:, subject to the constraints Budget: 9,500r, +25,000x, + 1,00,000x, $ 5,00,000 Gi. Minimum number of advertisements allowed (@) x2 1G (0) %25;and (0,25image not availableimage not availableimage not availableOparations Research: Theory and Applications Portfolio Data ‘Current price per share (Rs) #0.00 100.00 160.00 120.09 15000 200.00 Projected annual growth rate 008 = 0070.10 0129090. Projected annual dividend per share (Rs) 4.00 4.50» -7.50-5.50-5.75. 0.00 Projected risk return O05 0030.10 0.200.068 —_0.08 The total amount available for investment is Rs 25 lakh and the following conditions are required 10 be satisfied () The maximum rupee amount to be invested in alternative F is Rs 2,50,000. (i) No more than Rs 5,00,000 should be invested in altematives A and B combined. (ii) Total weighted risk should not be greater than 0.10, where (Amount invested in_alternative j) (Risk of alternative j) Total amount invested in all the alternatives, (is) For the sake of diversity, at least 100 shares of each stock should be purchased. (8) At least 10 per cent of the total investment should be in alternatives A and B combined (v) Dividends for the year should be at least 10,000. Rupee retum per share of stock is defined as the price per share one year hence, less current price per share plus dividend per share. Ifthe objective is to maximize total rupee return, formulate this problem as an LP model for determining the optimal number of shares to be purchased in each of the shares under ‘consideration. You may assume that the time horizon for the investment is one year LP model formulation Let x,,.%,x ty Xs and x ~ number of shares to be purchased in each of the six investment proposals A,B, C, D, E and F, respectively Rupee retum per share ~ Price per share one year hence ~ Current price per share + Dividend per share = Current price per share x Projected annual growth rate (i.e. Projected ‘growth each year + Dividend per share) Thus, we compute the following data: Total weighted risk No. of shares purchased Bi =e x Projected growth for each share (RS) 640,700 16.00 14.40 13.50 Projected annual dividend per share (Rs) : 400 450 750 5805.78 Return per share (Rs) 10401150 23.50 19.90 19.25 The LP model Maximize (total return) R= 10.40, + 11.50x, + 23.50x, + 19,90r, +19.25x, + 30.00x, subject to the constraints © 8lx, +100r, + 160x, + 120, + 150x, + 200%, < 25,00,000 (total fund available) @ 200, < 2,50,000 [from condition (i)] Gi) 80x, +100x, $ 5,00,000 [from condition (i)} 80x, (0.05) +100, (0.03) +1003 (0.10) + 1205, (0.02) +1505 (0.06) +200x, (0.08) (i) to, +1005, +1605, +120%, +1505,420%, SSCS or dh, 43g + 1615 + 2a + 91g + 16Kg S Bey 4 HO + Ibe, + 12ey + 154, +208 or = dx, Tx, + Ory + 12k, — 65 — ang SO (v) x2 100, 12100, x2 100, x,2 100, x2 100, x, = 100 {irom condition (iv)] (Wi) 80x, + 100r, 2 0.10 (80x, + 100x,+ 160x, + 120x, + 150x,+200x,) {from condition (v)] or BOr, + 100r, > Rx, + l0x, + 16x, + 12xy + 15x, + 20%, or Tax, + 90%, ~ I6x, ~ 12x, ~15x,~20r, 2 0 (ii) 4x, +4.5x, + 7.5xy + 5.54 +5.75x5 2 10,000 [from condition (vi)] and 20, =1,2,3, 4,5 andimage not availableimage not availableimage not available‘Operations Research: Theory and Applications 4) + 05x, + 1088, = 2, +5, Also, since the company will not commit an investment exceeding Rs 75,000 in any project, therefore the constraint becomes: x, $75,000 for i= 1,2, 3, 4, 5. and Fy Fp Ty Kp tos dy 5y 552 0. 2.8.4 Examples on Agriculture Example 2.29 A cooperative farm owns 100 acres of land and has Rs 25,000 in funds available for investment. The farm members can producea total of3,500 man-hours worth of labour during September — ‘May and 4,000 man-hours during Junc— August. Ifany of these man-hours are not needed, some members of the firm woulduse them to work on a neighbouring farm for Rs 2 per hour during September — May and Rs 3 ner hour during June — August. Cash income can be obtained from the three main crops and two types of livestock: airy cows and laying hens. No investment funds are needed for the crops. However, each cow will require an investment outlay of Rs 3,200 and each hen will require Rs 15. In additon each cow will also require 15 acres of land, 100 man-hours curing the summer Each cow will produce a net annval cash income of Rs 3,500 for the farm. The corresponding figures for each hen are: no acteage, 0.6 man-hours during September ~ May: 0.4 man-hours during June ~ August, and an ‘annual net cash income of Rs 200. The chicken house can accommodate a maximum of 4,000 hens and the size of the catile-shed limits the members to @ maximum of 32 cows, Estimated man-hours and income per acre planted in each of the three crops are: war ‘Man-hours September-May 40 20 25 June-August 50 35 40 Net aanual cash income (Rs) 1,200 800850 ‘The cooperative farm wishes to determine how much acreage should be planted in each of the erops and how many cows and hens should be kept in order to maximize its net cash income. Formulate this problem as an LP model to maximize net annual cash income. “Man-hours Sept-May 10006 «40 © «20s 1 = 3,500 June-Aug She Oh 50 3540 = 1 4,000 Land eee 1 1 - = 140 Cow 1 aS - - 2 Hens = iene 2 = = = 4,000 ‘Net annual cath income (Rs) 3,500 200 1,200 800850 Decision variables Let xx, and x, = number of dairy cows and laying hens, respectively xy %4 and x, = average of paddy erop, bajra crop and jowar crop, respectively. 1X = extra man-hours utilized in Sept-May. 1X; = extra man-hours utilized in June-Aug. The LP model Maximize (net cash income) Z =3,500x, + 200x, + 1,200x, + 800x, + 850%, + 2x, + 3x, subject to the constraints @ Man-hours: 10x, + 0.6x; + 40x, + 20:4 + 25x, + xg = 3,500 (Sept-May duration) 50x, + 0.44, + 50x, + 35, +40, +x, = 4,000 Giune-Aug duration) (Land availabilty: 5x, +x, +, + 155 100 Gi) Livestock: (a) x, $ 32 (dairy cows), (0) x, 5 4,000 (laying hens) and By Ep Xp My Xp Hp %7 2 0.image not availableimage not availableimage not availableOperations Research: Theory and Applications. No more than 40 doctors can start their five working days on the same day. Formulate this problem as an LP model to minimize the number of doctors employed by the hospital. [Delhi Univ, MBA (HCA), 1999) LP model formulation Let x)= number of doctors who start their duty on day j (j= 1, 2,.. ..7) of the week. The LP modet Minimize (total number of doctors) Z= x, +x, +3 + ty +45 +46 + 3 subject t0 the constraints @ tay tagtiyety 235 G Atxta tay ty, 255 Gi) Ht Atay ty tH 260, Wy) ytx tH tH tH 2 50 O) Asta tay ty tay 2 6, (WD) Het ay ty tay ts Z 50 (ii) xy tx, taytzy tay 245, (vil) x) $40 and x, 20 forall Example 2.35 A machine tool company conducts on-the-job training programme for machinists. Trained ‘machinists are used as teachers for the programme, in the ratio of one for every ten trainees. The training programme lasts for one month, From past experience it has been found that out of the ten trainees hired, only seven complete the programme successfully and the rest are released. ‘Trained machinists are also needed for machining. The company's requirement for machining for the next three months is as follows: January 100, February 150 and March 200. in addition, the company requires 250 machinists by April. There are 130 trained machinists available at the beginning of the year. Pays per ‘month are: Each trainee 2 Rs400 Each trained machinist (machining and teaching): --Rs4,900 Each trained machinist idle Rs4,700 Formulate this problem as an LP model to minimize the cost of hiring and training sckedule and the company’s requirements. LP model formulation Let Xp» = trained machinist teaching and idle in January, respectively ‘xy %q = ttained machinist teaching and idle in February, respectively X51 ¥¢ = trained machinist teaching and idle in March, respectively The LP model Minimize (total cost) Z~ Cost of training programme (teachers and trainees) + Cost of idle machinists + Cost of machinists doing machine work (constant) 400 (10, + 10xy+ 10x.) + 4,900 (x, +5 +x) + 4,700 (x, +x, +39) . subject to the constraints (Total trained machinists available at the beginning of January = Number of machinists doing machining + Teaching + Idle 130 = 100 +x, +x oF xj +x) = 30 (i) Total trained machinists available at the beginning of February = Number of machinists in January + Joining after training programme 1304 7x, = 150+ 25424 oF Try —¥y— y= 20 In January there are |Orx, trainees in the programme and out of those only 7x, will become trained machinists. Gi) Total trained machinists available at the beginaing of March = Number of machinists in Janusry + Joining after training programme in January and February 130+ 7x, + 7x; = 200+ x5 + 35 Te, + Try —¥,— 4 = 70 iv) Company requires 250 trained machinists by April 130 + 7x, + 7x4 + Try =250 ‘Ixy + Tay + Tag =120 and Ji Xp ty ty Xo Je 20.image not availableimage not availableimage not available15, 16, ‘Operations Research: Theory and Applications 17, ‘A manufacturer preduces three models (| and il) of a cartain Pe) Rae ea EE face) Beason, Nay vet Sine || “hea ean areCar n ase NP ia ‘oo and8.00 te pas reaped Th mura ‘souinons aro te tees rote or on ie Snum na, 90g 48 240.000 “ile Wael Roars ‘par Uk of ha Wd im eunean nae a aaa 7 7 3 : The bicens ae 3 i B ; Pains Price Minimum Maximum Maximum —Quentty Coffee par kg Strongth Acidly Per Demanded (fis) Cafioine (kg) __ Cont xX 4 65 38 22 35,000 x 55 6D 025,000 ‘Assume that 35.000 kg of plains X ard 25.000 kg of plains XX. are to be sold. Formulate this problem as an LP model to maximize sales ‘A manufacturer of metal office equipments makes desks, chairs, cabinets and book cases. The work is caitied out in the thres ‘major manufacturing deparimenis: Metal stamping, Assembly and Finishing. Exhibits, A, B the and © gve the requisite data of the problem, Exhibit A Department Time Required per Unt ‘Available ‘of Product (hrs) Products __ Time per Desk Chair Cabinet Bookcase Week (hrs) Sampng 4 2 3 3 #00 Assembly 10 6 8 7 1,200 Fishing 108 8 '800 Exner ‘Department Gost (As) of Operation per Unit of Product Deck Char Cabnet Bookcase ‘Sampng 15 a 2 12 Assembly 30 18 24 2 Finishing 36 28 26 24 Exhibit C: Selling price (Rs) per urit of product Desk 175 Chair 95 Cabinet; 145 Bookcase 490 In orcer to maximize weekly profits, what should be all production programme? Assuming thatthe items preduced can be sold, which department needs to be expanded for increasing proits? Formulate this problem as an LP model ‘The POR stone company sells stone procured from any cf three adjacent quarries. The stone sold by the company conforms to the following specication Material X equal to 30% atotial ¥ equal to or lose than 403% Matetial Z between 30% and 40% Stone fram quarry A costs Ais 10 per tenn properties: Material X : 20%, Materal ¥ : 60%, Material Z 20% ‘Stone from quarry B costs Rs 12 per tenne and has the folowing propentes: Material X : 40%, Material Y : 20%, Material Z : 20% ‘Stone fram quarry C costs Rs 15 per tonne and has the following properties: Material X : 10%, Nateral ¥ : 40%, Material Z : 50% From what quairies should the PQR store company procure rocks in order to minimize cost per tonne of rock? Formulate this problem as an LP model. [Delhi Univ. MBA, 1997 ind has the following 18. 19, 20. ‘The labour time for each unit of models twice that of model I and three times that of mode! Il The entire labour force of the factory can produce the equivalent of 2,500 units of model ‘A market survey indicates that the minimum domand for the three models is 500, 500 ard 375 unis, respectively. However the ratios of the number of units produced must be equal to 32:5, Assume that the profit per unit of models | I and I is Ais 60, Fs 40 and As 100, respectively. Formulate this problem as an LP mocel to determine the number of urits of fact product that will the maximize amount 0! prot A company manufactures two models of garden roles: X ané Y. When preparing the 2008 budget, it was found that the limitations on capacity were represented by the folowng woek!y ‘Production maxima: ‘Model Foundry Machine-shop Contribution per ‘Vode! (RS) x 10 200 720 ¥ 240 150 80 {Tw adgiion, the material required for medel X was in ston supply land sufficient only for 140 units per week, guaranteod for the year. Formulate this problem as en LP model to determing the ‘Optimal combination of output. ‘A company manufacturing television and radio sets has four ‘major departments: chassis, cabinet, assembly and fra testing. ‘Tho morthly capacties of these are as follows: Television chassis 1,300 or Cabinet 4,000 or ‘Assembly 2000 or Testing 3,000 or “The contibution of a television se is Ps 500 and that of a radio ‘Set As 250. Assume that the company can sell any quantiy of either product. Formulate this problem as an LP model to termine the optimal combination of television and radio sets, ‘A company wants to plan production for the ensuing year 80 4 to minimize the combined cost of production and inventory storage. in each quarter of the year, demand is anticipated 10 be 65, 80, 135 and 75 respectively. The product can be manufactured during regular time at 2 cos of Rs 16 per unit produced, or during overtime at @ cost of RS 20 per unit. The table given boiow gives data perthent to production capacitios, ‘The cost of carrying one unit in inveniory per quarter Is Rs 2 ‘The inventory evel at the beginning of the rat quarter is zero ‘Quarter Capacities (units) Guarterly Fegular Time Overtime __ Demand 1 © 10 65 2 2 10 20 3 8 20 135 4 7 10 75 Formulate the problem as an LP madel 60 a0 to minimize the production plus storage costs for the entire year [Dem Univ, Ba, 1998)image not availableimage not availableimage not availableOperations Research: Theory and Applications, HINTS AND ANSWERS 1. Letx,,x, = number of units of products A and B to be produced, respectively Max Z = 40x, + 30x, subject tox, +x $ 3,000 (Manchours) 4 $8,000 ; x, $ 1,200 (Marketing) and 3,20 2 Let xy, x; =mumber of productive mins of process 1 and 2, respectively, Max Z = 300, + 400:, subject to 53, +4r, $200) (Max amount of crude A 3x, #5r; £150] and B) ‘Sx, + 4x, 2100) (Market requirement of 83, +4x, 280 | gasoline X and Y) and ay 80. 3. Let x, = number of units purchased per month (@= 1, 2, 3~ April, May, June) ‘number of units sold per month (= 1,2, 3 — May, June, July) Max Z = (90y, + 60y, + 75ys) - (75x, + 75e, + 60x) subject toy, Sx, < 150 Sm +H —y, $150 yy £4, +4 tx 9, —92 £150 nto tment ty and 45420 forall 4. taj andy “mm of wf ai ad pend) oe mame ies sl tg ana sre) subject to 1 hey +i $2.50 radon ian x, 2500; x, 2500; x, 2375 (Market demand) 1 giai g%2 = 5% (Ratios of production) 5. Let x, and x, = parts of A and B per hour manufactured, respectively. x Z = (55 +65)-{ 224441722200) nara ones 1] 20,4 17505), BTiS as = 1.205; + 140K sbjectto OA+Z si A+B st G+ <1 (Manufacraring capacity) wd yr 20, 6. Let xjq = mumber of units manufactured in month i (1, 2,3 = Oct, Nov., Dec.) during shift ‘J (= 1, 2~ regular, overtime) and shipped in month (k= 1, 2, 3 - Oct, Now., Dee.) Bent Set pat 629 + Se + Frag Qn ttt Ay + Oh + Sy + Shy subject © ayy tay taps $1500 Xoi2tXy3 $1500 (Regular time) 1 $1,506 Sittin thas € 750) Xi tXz) $750} (Overtime) X53 < 750) An * in = 1000 fai ttm +a + tm Aus i * 2213 * Aan tia tg 7 2,000 and Xin 2 Ofor all, jk 7. Let. 2, = number of gallons of wine B and C in the blend, respectively. Max Z = 20 +x, + x, (20427}433x, #323, +x, + (Resultant degies proof of tend) (20%0.32)+02", 2043, +4, subject 10 30 a +0882 50.25 (Acidity) (20107) + 1.08%, +1.08 GORMOT NB *1O48 5 1.05, (Specific erevity) 20% 5304 (Quilty) and xymp 20, 8. Let. tp and x5 = quantity of foods 1, 2 and 3 w be used, respectively, Min Z = 1.50%, + 2.00, + 1.20%) x+y S12, subject 10 350% +250x, + 200.5300 250%, +300; +150x,2 200 100s; +150x, + 75x52100 75x, +1251, + 150%, 2100 and agg 20. 9. Let x and x, = number of soccer balls of types X and Y, respectively. Min Z = (2 hrs) (Rs 5.50/he) x, + (4 hrs) (Rs 8.50/h) x, + has) (Rs 5.50/m) x + (6 hrs) (Rs 8.50Vhr) x, (Se, + 6750x, subject © 2x, + 3x, S80 (Semi-skilled hour) 4x, + 6% 4 150 (Skilled hours) x S 15 all X) mS 10 (Ball Y) and spe 0.image not availableimage not availableimage not availableOperations Ressarch: Theory and Applications Eeakeraic tata Ree ee dara se hea on a ten of on? (0) value of a decision variable (b) highest value is chosen among allowable decisions 55 RN wares 5 ee ee eas a because (a) resources must be limited pete Fc ce a eagnpeen world problems 33, Nori-negativty condition in an LP model implies (a) none of the above (bd) a positive coefficient of variables in any constraint () express the objective function in words ‘34. Which of the following is an assumption of an LP modal (0) verbally identity decision variables (a) divisibity (b) proportionality (a) of the above (c) additivity: (d) all of the above ov mmcnucpaut un — | on Sadana tae etn a & et oe Oe eee fat ae a pa BS gmaeadingine am ee fey Eee BSimage not availableimage not availableimage not availableOperations Research: Theory and Applications ‘These two points are then connected by a straight line as shown in Fig, 3.1(a). But the question is: Where are ‘these points satisfying 4x, + 6, 360, Any point above the constraint line violates the inequality condition. But any point below the lin does not violatethe constraint. Thus, the inequality and non-negativty condition can only be satisfied by the shaded area (feasible region) as shown in Fig. 3.1). * ¥4 4 +6,-260 % Ok secss 2 Se Gir O.x = 0) ® Osn-20 ®KO oe PadXcors) o Fig. 34 am é * Graphical Solution & of LP Problem as, ¥ o 20 0 60 80-90 + o ™ @ (b) Similarly, the constraints 3x, < 180 and 5:, < 200 are also plotted on the graph and are i the shaded area as shown in Fig. 3.1(b). Since all constraints have been graphed, the area which is bounded by all the constraints lines including all the boundary points is called the feasible region (or solution space) The feasible region is shown in Fig. 3.1(b) by the shaded area OABCD. 3.(0 Since the optimal value of the objective function occurs a one of he extreme points of the feasible region, it is necessary to determine their coordinates. The coordinates of extreme points of the feasible region are: O = (0, 0), A = (60, 0), B = (60, 20), C = (30, 40), D = (0, 40). (i Evaluate objective function value at each extreme point of the feasible region as shown in the table: 15(0) + 1060 0 0,0 4 (60, 0) 15(60) + 10(0) = 900 g (60, 20) 18(60) + 1020) = 1.100 c G0, 40) 15(30) + 10(40) = 850 D (0, 40) 15(0) + 10(40) = 400 (i) Since we desire Z to be maximum, from 3(ii), we conclude that maximum value of Z = 1,100 is achieved atthe point extreme B (60,20) Hence the optimal solution tothe given LP probiem is: 1) x9 20 and Max 21,100 FSS TENET TE” Remark. To-determine which side of constant equition isin the feasible region, examine whether the Fee selon IS origin (0, 0) satisfies the constraints I it does, then all points on and below the constraint equation towards Of constrans that" the origin are feasible points. If it does not, then all points on and above the constraint equation away satisfies all of the from the origin are Feasible points constants on ” a resources, Example 3.2 Use the graphicsl method to solve the following LP problem. Maximize Z= 2x, + subject to the constraints Oty <1, Dy +mS6, @ x) -%) 22, Gv) x, 251 and 20, | Solution Plot on a graph each constraint by first treating it as a linear equation in the same way as discussed ‘earlier. Use the inequality condition of each constraint to mark the feasible region as shown in Fig. 3.2. The feasible region is shown by the shaded area. Here it may be noted that we have not considered the area belowimage not availableimage not availableimage not available‘Operations Research: Theory and Applications 3.33 Examples on Minimization LP Problem Example 3.6 Use the graphical method to solve the following LP problem. br + ey subject to the constraints @ Sx 4x,210, Gy +x26, (iil), +4r,212 and p20. Minimize Z Solution Plot on a graph each constraint by first treating them as @ linear equation in the same way as discussed earlier, Use the inequality condition of each constraint to mark the feasible region as shown in Fig 3.6. sy Os4n-10 @ as inné of PO. @ ay e4y=12 @ ° Feasible 9 Region 4 Fig. 9.6 2 Graphical Solution SELE Probe, Oa ee ‘The coordinates of the extreme points of the feasible region (bounded from below) are: A = (12, 0), B = (4,2), C=(1, 5) and D = (0, 10). The value of objective function at each of these extreme points is as follows: 4 (12, 0) 312) = 2(0)= 36 a 34) + 20) ~16 as) 1) + 215)=13 10) (0) + 2(10) = 20 ‘The minimum value of the objective function 2~ 13 occurs at the exireme point C (I, 5). Hence, the ‘optimal solution to the given LP problem is: x, = 1, x,= 5, and Min Z = 13. Example 3.7 Use the graphical method to solve the following LP problem. Minimize Z=—x, + 2x, subject to the constraints O-m+3y, 510, Gi) ty s6 iii) x)-x, 62 and x 20. 4 QO-m+iqg=10 Ox tqes Fig. 3.7 Graphical. Soluton of LP Probiemimage not availableimage not availableimage not availableEE¥E Oporations Research: Theory and Applications Example 3.11 Use the graphical method to solve the following LP problem. Maximize Z= 71, + 3x subject to the constraints @ x, +2423 @ x+n<4 (i) 0 Wr, + 6, ° 0.0) 1040) + 6(0) = 0 4 (0, 9) 1040) + 6(9) = 54 5 67, 607) 1067) + 6(6077)= 60 € 6,9) 1046) + 6(0) = 60 Since on two different extreme points B and C the value of objective function is same, i.e. Max Z = 60, wo alternative solutions: x; = 6/7, xy = 60/7 and x,= 6, x) = 0 exist. Remark Ifa constraint to which the objective function is parallel does not form the boundary of the feasible region, the multiple solutions will not exist. This type of a constraint is called redundant constraint, i.e. @ redundant constraint is one whose removal does not change the feasible region 3.4.2 Unbounded Solution Sometimes an LP problem will not have a finite solution. This means that when one or more decision variable values and the value of the objective function (maximization case) are permitted to increase Ynpaunded. infinitely, without violating the feasibility condition, then the solution will be infinite. This is known as en te soutton an unbounded solution. It is important here to note that there is a difference between a feasible region vatable and the ing unbounded and an LP problem being unbounded. It is possible that in a particular problem the Prfit can be made Teale region may be urbounded but LP problem may not be unbounded, 12. an unbounded feasible lua region may yield some definite value of the objective function. The general cause for au unbounded LP St me mavimizaton problem is an improper formulation of the real-life problem. LP problem Example 3.21 Use the graphical method to solve the following LP problem: peers Maximize Z.~ 3x, + 4x, subject to the constraints @ 4-4 @ -x, +4450 and mp 20. Solution Plot oa graph each constraint by frst treating it as a linear equation in the same way as discussed earlier. Then use the inequality condition of each constraint to mark the feasible region as shown in Fig. 321 4 QOarn--1 @ -nen-0 ; pac as Fig. 3.21 Urbounded oT ™ Solution 1 may be noted from Fig. 3.21 that there exist an infinite number of points in the convex region for which the value ofthe objective function increases as we move from the extreme point (origin), to the right. That is, both the variables x, and x, can be made arbitrarily large and according the value of objective function Z will also increase. Thus, the problem has an unbounded solution Example 3.22 Use graphical method to solve the following LP problem: Maximize Z= 3x, +2, subject to the constraints, Oa-x x20 1 @ ures and aOperations Research: Theory and Applications Solution The constraints are ploted on graph as usual as shown in Fig. 3.22. The solution space is shaded and is bounded jrom below. It is noted here that the shaded convex region (solution space) is unbounded from above. The two comet of the region are, A = (0, 3) and B= (2, 1). The value of the objective function at these comers is: Z(A) and 2(B) = 8 Fig. 3.22 Graphical Solution of LP Problem Since the given LP problem is of maximization, there exist « number of points in the shaded region for which the value of the objective function is more than 8. For example, the point (2, 3) lies in the region and the function value at this point is 12 which is more than 8. Thus, both the variables x, and x, can be made arbitrarily large and accordingly the value of Z will also increase. Hence, the problem has an unbounded solution. Example 3.23 Use graphical method to solve the following LP problem Maximize Z=Se,+4x, subject to the constraints @x,-2yS 1, (i) x, +24y23 and Xy%20 [PT Univ, BE, 2001) Solution Constraints are plotted on 2 graph as usual as shown in Fig. 3.23. The solution space is shown shaded and is bounded from below. This shaded convex region (solution spsce) is unbounded from above. ‘The twoextreme points ofthe solution space are, A(0, 3/2) and BQ, 1/2). The value of objective function at these points is Z(A) =6 and Z(B) = 12. Since the given LP problem isof maximization, these exists a number of points in the solution space where the value of objective function is much more than 12. Hence, the unique value of Z cannot be found as it occurs at infinity only. The problem, therefere, has an unbounded solution ” Solution space (0,372) Fig. 3.23 Unboundea SolutionLinear Programming: The Graphical Method Example3.24 Solvethe followingLP problem graphically MaximizeZ~~ 4x, +35 subject to the constraints O y-x) 50, @us4 and sky 2 0. [Punjab Unix, BCom, 2005) Solution The solution space satisfying the constraints and the non-negativity restrictions is shown shaded Fig 3.24. The line x, ~ x, =0 isdrawn by joining origin point (0,0) and has a slope of 45°. Also asx, —x,< 1 $Xp the solution space due to this line is in the upward direction, O w-* Feasible Region 3 2 1 Fig. 2.24 Unbounded x Solution 0 123 4 5 ‘Since objective functions of maximization, therefore the value of can be made arbitrarily large. Hence, this LP problem has an unbounded solution. Value of variable x, is limited to4, while value of variable x,can be increased indefinitely. 3.4.3 Infeasible Solution Infeasibiliy is condition that arises when there is no solution to an LP problcm that satisfies all the constraints simultaneously. This means, there would be no unique (single) feasible region. Such a problem 45 infosaibie arises when a wrong model is formulated that has conflicting constraints. Any point lying out side the solution les feasible region. It violates one or more of the given constrains. outside tne feasible Infeasbilty solely depends on the constraints and has nothing to do with the objective function. This f29in. it voces type of a situation requires the ability of the decision-maker to resolve the conflicting requirements of hex corotrants resources so that « decision that is acceptable to all sections of the organization can be made. Example 3.25. (Prodlem with inconsistence system of constrains) Use the graphical method to solve the following LP problem: Maximize Z = 6x, 4x, subject to the constraints @ 2 44x Sd Gi) Ax, + Bey BIE and x20 ° Dey 16 @ 2 H4q=4 Fig. 3.25 Aan Infeasible Se Solution Solution The constraints are plotted on graph as usual as shown in Fig. 3.25. As you can see there is ‘no unique feasible solution space that allows us to obtain a unique set of values of variables x, and x, ‘that satisfy all the constraints. Hence, there is no feasible solution to this problem because of the conilicting constraintsOperations Research: Theory and Applications Fig. 3.26 ‘An Infoasible Solution Fig. 3.27 No Feasible Solution Example 3.26 Use the graphical method to solve the following LP problem: Maximize Z= x, + subject to the const @ 3x, +2e512 GH Sxy=10 x, +28 v)-x +424 and x20 Solution The constraints are plotted on graph as usual as shown in Fig, 3.26. The feasible region is shaded. As shown in Fig. 326, the three shaded areas indicate non-overlapping regions. All of these can be considered feasible solution areas because they all satisfy some subsets of the constraints Ox,+25 On-n-8 o 2 4 6 8 0% However, we cannot find any particular point (x,,x,) in these shaded regions that can satisfy all the constraints simultaneously. Thus, the LP problem has an infeasible solution Example 3.27. Solvethe following LP problem graphically Maximize Z=3r+2y subject to the constrainis @ -2x+3v<9, Gil) 3-2 < -20 and xy20 [Punjab Univ, BE Elect.) 1996] Solution There are two sclution spaces (shaded areas) shown in the Fig, 3.27. One of these solution space is satisfying the corstraint ~ 2x +-3y <9 while the other is satisfying the constraint 3x ~ 2y <~20. These two shaded regions in the first quadrant do not overlap and hence there is no point (x, ») common to both the shaded regions. This implies thatthe feasible solution to the problem does not exist. Consequently, this LP problem esnno! solved graphically.Linear Programming: The Graphical Method 3.4.4 Redundancy ‘A redundant constraint is one that does not affect the feasible solution region (or space) and thus redundancy of any constraint docs not cause any difficulty in solving an LP problem graphically. As previously shown in Fig, 3.12 constraint &x, + 4x, 2 80 is redundant. In other words, a constraint is said to be redundant when it may be more binding (restrictive) than the another Redundancy is a situation in which fone or more constraints donot affect the feasible solufon region CONCEPTUAL QUESTIONS Srpiain te graph! math of wohing an LP prob. | Give a triof desorption of en LP problem with iluatratione How can it be solved graptically? What is meant by he term Teesible region"? Why must his be a Welldetinad boundary forthe maxiization problem? Wht is a feasiiliy region? Is necessary thet it should always, bbe a convex eet? Define iso-profit and iso-cost lines. How do these hep us to obtain a soluton to an LP problem? | Define the corcept of convexity. Why must fhe feasible region | exhibit the property of convexity in. an LP problem? Explain the procecure of generating extreme polnt solutions to an LP probiem, pointing out the assumptions made, if any. 18. It hes boon sald that each LP problom that has a feasible ragien has an infinte number of solutions. Explain 9. You have just formulated a maximization LP probiem and are preparing to solve it graphically. What erieria shoud you ‘consider in deciding whether it would be easier to soive the ‘problem by the extrsme point enumeraton method oF he iso- profit ine method, 10. Under what condition is t possible for an LP protlem to have ‘mote than one optimal soluton? What do these alternative ‘optimal solutions represent? 14, What is an tnfeasitie solution, and how does it occur? How | ths condition recognized in the graphical method? 12, What is an unbounded solution, and how is this concition recognized in the graphical method? SELF PRACTICE PROBLEMS Comment on the solution of the folowing LP problems (i) Max Z = 4x + 4, (W) Max Z = 8x, + 3x, () Max Z = 4%, + 2x) sublet to | x, + 24 5 10 wbpee 4's 248 subject tox, + 2u <6 ox, + 6x xy 88 “H+ Ge? <6 and m2 0 meT Me and Hy 20 and My m2 0 (iv) Min 2 = x, ~ 2) (v) Max 2 = 6%, + 12% (vl) Max 2 = 3x, 4 5%) eubjct to Be + = 8 subject to x, + 2xy 5 10 subject 0a + a = 100 wh + 2% 5-24 2x; — 5% 5 20 5x, +10% 5 400 na Hy BBO Ke ys 6x, + Bx < 440 and wy EO ond my He 0 (il) Max 2 = 1.78%, + 15% (vi) Max Z = Sx, + 2x, (0) Max 2 = 5x, + 3% subject to Be, + 5x < 220 subject to 2x, + Sx < 9 subjost to Ox + Say = 15 4x, + 5x 5 20 x, - 5y 2 20 5x, + 2xp < 10 215; x 2 10 and Jy 20 and 20 and Hy PO Ierale, 83c Engg), 1985) (x) Max 2 =) 4% subject oy + St 8,4 23 and My mE 0 Soke the folowing LP problems graphically and stalo what your solution Indicates, () Min Z = 4x, = 2x (i) Min 2 = 3x; + 5x (i) Min 2 = 20x, + 10%, subject oy + S14 subject to — 3x; + 4x, 5 12 subject 10x; + 2x) 5 40 Bx, + 2x 298 ay pe? a+ E90 y+ 4 S28 2x + 3x2 12 4x, + 3x 2 60 and % m2 0 S42 2 and Hy 20 and n> 0image not available12, 14 18. 16. ‘A company produces two types of presentation goods A and B that require gold and siver. Each unt of type A requires 39 Of siver and 1g of gold while B requites 19 of slver and 2g Of gold. The company can produce 9g of siver and 8g of (oc. each unit of type A brings a profit of Rs 40 and fat of ype B Ae 50, dotarmine the number of unite of each typo that should be produced in order to maximize the profit. ‘A turn makes two types of luniure: chairs anc tables. The Cconirbuton to profit by each product as calculated by the ‘accounting department is— Rs 20 per chair and Ris 30 por table Both products are to be processed on three machines M,, M, | ‘and My. The time required in hours by each procuct and total time avalable in hours per week on each machine ere as follows: ‘Machine Chair ‘Table Avalabio Timo (te) My a a 26 ™ 5 2 50 Me 2 6 80 How should the manuactrer scheaule is production in forder to maximize proft? The ABC company has been a producer of picture tubes for television sets and of certain printed cicuis for radios. The Company has just expanded Into ful-scale preducton and ‘marteting of AM and AV-FW radi. I! has built a new plant that can operate fer 48 hours per week. Preducton ol an AM radio in the now plant wil require 2 hours and production of fan AM-FM radio will require 3 hours. Each AM radio will Contibute Ris 40 towards the profis while an AM-FM radio will contibute Rs 80 towards the profits. The marketing department, ‘afer extensive research, has determined that a maximum of 115 AM radios and 10 AV-FM radios can be sold each weak. Formulate this prodlem as an LP model so as to determine the optimal producton mix of AM-FM radios that will maximize profts. [Deh Univ., MBA. 1998, 2004) ‘The productien of a certain manufacturing firm involves a ‘machining process that acquires raw materials and then ‘converte thom info (unassembled) parte. Thoso parte aro then sent to one ofthe two divisions for beng assembly into the final product. Dvision 1 ie used for product A, and Divison 2 for product B. Product A requires 40 unis of raw material land 10 hours of machine processing line. Product B requies ‘80 units of ram material and 4 hours of machine precessing lime. During the period, 09 unis of raw material and 80 nous ‘of machine processing ime are avalable. The capabilites of the two assembly dvisons dung the petiod are 6 and 9 Units, respectively. The profit contibution per unit to profit and ‘overhead (fied costs) is of As 200 fer each unit of product ‘A and of Fs 120 for each unt of product 8. With this information, formulate this problem as a linear programming model and’ determine the optimal level of cutput for the two ‘Products using the graphic: method. ‘A local travel agent Is pinning a charter trip to @ famous sea, resot. The eight-day seven-right package includes the round- tip fare, surface transporaion, boarding and lodging end Selected tour options. The charter tro is restrcted to 200 [persons and past experience Indates that there will not be any problem in arranging 200 passangers. The problem faced by he travel agent is to deterine the number of Deluxe Standard ané Economy tour packages to offer for this Charter. Al three cf these plans each differ in tarms of their ‘seating and sarvice for the fh, qualty of accommodation, ‘meal plans ard tour options. The following table summarizes the estimated prices of the three packages and the Corresponding expenses of the travel agent. The travel agent hhag hired an srera for a Mat foe of is 2,00,000 for tha entra trip. The per person price and costs (or the three packages: fare ae fellows: Linear Programming: The Graphical Method (EEE Tow Plan Price Rs) Hotel ‘Meals and Costs (FsjOther Expenses (Rs) Deluxe 10,000 3,000 4,750 Standard 7,000 2,200 2,500 Economy 6,500 1,900 2,200 In planning the ti, the folowing considerations must be taken into. account: () Atleast 10 percent of the packages must be of the deluxe type li) At least 35 percent but not more than 70 per cent must be of the standard type, (i) Atleast 80 percent must be of the economy type. (Ww) The maximum numter of deluxe packages aralatie in any aircraft is restricted to 60. (¥) The Notel desires that at least 120 of the tourists should be on the deluxe standard packages together ‘The travel agert wishes to determine tne number of packager to ofer in each typo of trip co ae to maximize tho {otal prof. (@) Formulate ths problemas a inear programming problen, (b) Restate the above linear programming protiem in terms of two decision variables, takng advaniage of the fact that 200 pactages will be sold (©) Find the optimum solution using graphical methods for the restated linear progranming problem and interpret your results [Delhi Univ., MBA 2004) [A publisher of textbooks isin the process of bringing a new book in the market. The book may be bourd by either clath (oF hard paper. Each ciot-bound book sold contibutes Fs 24 towerds the profit and each paperbound book contributes AS 23. k takes 10 minutes fo bind @ cloth cover, and 9 minutes to bind a paperback. The total avaiable time for binding is 80 hours. Aller considering a number of market surveys, it Is predcted that the cbth-cover sales will be arything more than 10,000 copies, but the paperback sales will not be more than 86.000 copies. Formulate this problem as a LP problem and solve it gaphically PQA Feed Compary markets two feed mixes for catte. The feed mix, Fetiex, requis at least twive as much wheat as barley. The second mix, Multiplex, requires at least Iwice as ‘much barey as wheat. Wheat costs Rs 1.50 per kg, and only 1,000 kg are availabe ths month. Barley costs Fs 1.25 per «9 and ',200 kg is avaiable: Ferlex sals for Rs 1.80 per kg up to '99 kg and.each addtional kg over 99 sells for Rs 1.65. Multiplex sels at Rs 1.70 per kg up to 99 kg and each additonal kg over 90 sels for is 1.55 per kg. Bharat Farms is sure to buy any and all amounts of both mixes thet POR Feed Company will mix. Formulatethis problem asa LP problem to determine the product mix so as to maximize the profs. (On October 1st, a company received a contract to supply 8,000 unis of a specialzed product. The terms of contract requires that 1000 units be shipped in the north of October; 3,000 units in November and 2,000 units in December. The ‘company can manvtacture 1,500 units per month on regular ‘ume and 750 units per month on overtime. The manutecturing ‘cost per fem produced during regular time is As 3 end the ‘cost per ilem produced during overtime is Fs 5. The monthly storage cost is Re 1. Formulate this protiem as a linear programming problem so as to minimize the total cos A emall-scale manufacturer has preducton facilites. for producing wo different products. Eech othe products requites thvoe dliferont operators: ging, ascombly and tasting Product {requires 15, 20 and 10 minutes to grind, assemtle land teat, respectively, on the other hand product I requires 75, 40 and 45 minutes for ginding, assembling and testing, respective. The production run calls tor at least 7.5 hous21 22. 23. Operations Research: Theory and Applications. of grinding time, at least 20 hours of assembly and at least 15 hours of testing time, f manufacturing product | costs AS 60 and roduct Il costs Rs 60, determine the number of units fof each product tho firm should produce in order to minimize the cost of operaion. ‘A frm ie engaged in breeding piga. The pige are fed on various products grown in the farm. Because of the need 10 fensure certain nutrient constivents, it is necessary to adatonaly buy one o two produets (call them A and B) ‘The conten! of the various products (per unt) in the nutront constituents (e.g. vitamins, protains, etc) is given in the following table ‘Nutrient ‘Nutriont Cortent in Minimum Amount Product ‘of Nutaent A a M, 96 6 108 M, 3 2 36 Ms 20 10 100 “The last column ofthe above table gives the minimum quantity ‘of nutrient constituents M,, M, and M, that must be given to the pigs. If products A and B cost Rs 20 and Fs 40 per unit, respectively, how much af each of these ‘wo products should be bought so that the iotal cost is minimized? Solve this LP problem graphical ‘A company manufacturng television sets and radios has four major ‘departments: chassis, cebinet, assembly and fnal tesing. The monthly capaciies of these are as folbws: Department __‘Telovieion ‘Radio Capacity, Capaciy Chassis 1,500 or 4,500 Cabinet 1,000 or 8,000 Assomly 2,000 or 4,000 Testing 3,000 or 9,000 The contribution of television is As 160 each and the contribution of radio is A 250 each, Assume that the company can sell any quantity of eitier product. Formuate this problem as an LP problem and solve I to determine the ‘optimal combination of output The ABC Clcthing Stoves is planning their annual shin and pant sale. The owner, Mr Jan ie planring to use two diffrent forms of advertising, viz, radio and newspaper ads, in order to promcte the sale. Based on past experience, Mr Jain feels Confident that each newspaper ad wil reach 40 shit custorrers| and 80 fant customers. Each rado ad, he believes, wil reach 30 shit customers and 20 pant customers. The cost of each ‘newspaper ad is As 300 ard the cost of each radio spo: is Re 450. The advertizing agency wil propare the advertising {and it will require 5 man-hours of preparation for each newspaper ad and 15 mamhours ol preparation for each radio spot. Mr Jai's sales manager says that a minimum of 779 mar-nours should be spent on preparation of advertising in order to fully uilize the servces of aivertsing agercy. [Mr Jain feels that in order to have a suncesstul sale, the advarising must reach at least 360 shit customers and at least 400 pant customers. (). Formufate this problem 2s an LP model oo a6 to determine how much should the sale be advertised Using 2ach meda in ord to minimize costs and stil attain the objectve of Mr Jain, (Save tne LP problem usng the grapnical metnoc 24, The commander of a small tank has beon ordered to win and ‘occupy a valley located in a rver delta area. He has 4 heavy tanks and 10 fight lanks, Each heavy tank requires 4 men to operate, whereas the ight tank requires 2 men. The total ‘number of men avaiable is 29. The fire power of the hasvy tank io tree times that of the light tank. Yet the commander feols that he shoud use more light tanks than heavy ones since the light ones are more effecve against guerias, () Formulate tis problem as an LP model so as to determine the number of tanks in each type fo be sent into the combat, keeping in mind thet the commander Wwahes to maxinize the total fre power. (If you could persuads the commander that ho rule ‘more light tanks than heavy tanks’ may not be applicable due to the fact that quella wavlare i absent in the area, how many tarks of each type could be usec? 25. A timber company cuts raw timber ~ oak and pine logs — into wooden boards. Two steps are required to produce boards from logs. The first step involves ‘emoving the bark from the logs. Two hours are required to remove the bark from 1,000 fect of oak logs ard thrae hours per 1,000 feet of pine logs ‘After the logs have been debarked, fey must be cut info boarde. It takes 2.4 houre per 1,000 fect of oak loge to bo cut Into boards and 1.2 hours per 1,000 feet of pne logs. The bark moving machines can operate fer up to 60 hous per week, While for cutting machines this number is limited to 48 hours er week The company can buy @ maximum of 18,000 feet (of raw sk logs and 12,000 feet of raw pine logs each week ‘The prof per 1,000 feet of processed logs i FS 1,200 and fs 1,200 for oak and fine logs, respectivay. Formulate this problem as an LP mode! and solve it to determine how many feet of each type of log should be processed each week in order to maximize the prof. 26. Upon complotng tho construction of fis house, Mr Sharma discovers that 100 square leet of pywocd scrap and 80 aquere feet of white pine scrap are in unusable form, which can be used for the construction of tables and bookcases. It ‘takes 16 square fest of plywood and 16 square fest of write pine to constrict a book case, By seling the finished products 1 a local tumture store, Mr Sharma can realize 2 prot of AS 25 en each table and Fs 20 on each bookcase. How can he ‘mos! proftably use the leftover wood? Use the graphical method to solve this LP problem, 27. A manufacturer produces electric hand saws and clectic ‘rile, for which the demand exceeds he capaci. The production cost of a saw is is 6 and the production cost of | dell Re 4. The shipping coct fe 20 paice lor a caw and 30, pase fora dil A saw solls for Fs 9 and a drill slls for As 5.50, The budget alows a masimum of Ris 2,400 for preduction ‘costs and Rs 120 for shipping costs. Formulate this problem 188 ah LP model and soe it determine the number of SANS and drills that should be produced in order to maximize te ‘excess of sales over production and shipping costs. 28. A company produces two types of pens. say A and B. Pen A is of superior qualty and pen B is ofan infeior quality. Profs fon pan A and 8 are Rs 5 and Rs 3 per pen, respectively. Tha ‘aw mateial required for producing single pen A is twice as that ff pen B. The supply of faw mataral is sufiiont only fer 1,000 pens of B per day. Pen A requires a special cip and only 400 ‘uch clipe are availble per day. For pen B, only 700 cipe ere avaiable per day. Solve this LP problem graphicaly to find the product mix so that he company can meke the maximum prt. [Deni Unv., MBA, Nov. 1996) 29, Two proaucts A and B are 1 be manutacured. One single Unt of product A requires 24 minutos of punch pross tine land 5 minutes of éssembly timo. The prott for product A is Re 0.60 per unit. One single unt of product B requires &‘minutes of punch press time and 2.5 minutes of welding tine. | The proft for product & is Re 0.70 per uni The capacity of the punch press deparment available for these products is 41,200. minutesiweek. The welding department has an idle ‘capacity of 600 minutesiweek and the assambly department hhas the capacity of 1,500 minutesweek Formulate this Linear Programming: The Graphical Method ‘checks his bins and finds he has 700 kg of protein source land 500 kg of carbohydrate in the house. How mary bags ‘of each food should he mix in order to maximize his profits? Formulate this provlem as an LP model and solve it to determine the optimum fooé-mix that woud maximize the prof. Brotlem as_an LP model to ceternine the quarities of | 32, An adversing agony wishes to roach two types of auconces: products A and B that would yield the maximum. ‘customers with monthly income greater than Rs 15,000 20. A nibber company is engaged produchg three diferent {larget ausionce A) and cusiomers wi anual incomes of kinds of tyres A, B and C. These three different tyres are Jess than Rs 15,000 (target audience 8). The total advertising procuces at two deren plants of te company, which have ‘budget Is Rs 300000, One progtamme ot TV advertising Gtteont production capacities, In anormal S-hour working costs As 50,000; ene programme of radio advertising costs day, Plant 1 produces 50, 100 and 100 tyres of ypos AB. Re 20,000. For contract reasons, at leas! 2 programmos and C, respesivelr. Plant 2, produces 60, 60 and 200 tres ‘ug fo be on TY, and the number of radio programmes must of types A, B and C, respectively. The menthy demand for be limited to 5. Surveys Indoate that a sing TV progamne types A.B and CIs 2/500, 3,000 and 7000 nits, respectively roaches 450,000 customers in target auciorce A and 60,000 ‘The daly cost of operaton of Pants 1 and 2 is Rs 2500 and in the tagel audence B. One radio programme reaches Re 9,500, respectvaly. Formule the problem ae an LP 20,000 in target suctonce A and 80,000 in ta target auchonce rmocel and ‘soWve it to detarmine how the company can B, Formulate fis problem as an LP model and sole i t0 tminiize the numbor of deys cn which % operations, per determine the medians thal would maximize the total reach month, at the two plans, so fal the tal cost is aso {Dati Univ, MBA. 2005 lupczeyaobbin settee cable 33. A company has two grades of inspectors - 1 and 2 — who 31. Kishore Joshi mixas pet food in his basement on a smal are fo be assigned a quality contol Inspecton. I is required ‘scale. He advertises two types of pet food: Diet-Sup and at least 2,000 pieces be inspected per 8-hour day. A grade GroMore. Contrbuton trom Diet Sup is Fs 1.50 a bag and 4 inepector can check place atthe rato of 0 per how, wah {fom GroMare Rs 1.10 a bag. ath are mixed fom to basic fan accuracy of 97 percent. A grade 2 inspector checks at ingredients — a protain source and a carbohytate source the rate of 00 pieces per hour with an accuracy of 96 Diet Sup and Gro-iore requre ingredents in these amounts Beant eee “The wage feo of grade 1 Rspecior Is As 5 per hour wile Protein Carbohyarate that of a grade 2 iespecor is Rs ¢ per hour. An error made = a eee by an inspectr cost Fs 9b the Company. There se only Diel-Sup (7 kg bag) 4kg 3g 9 grade 1 inspectors and 11 grade 2 inspectors available in Gromore (8 Kg bag) 20 1g the company. The company wistes 10 assign work t0 te a avaiable inspectors £0 as to minmize the total cost of the Kishore has the whole weekend aheed of him, but wil ot inopection. Fomulie ths problem as « LP mode and solve be able 0 procure more ingredents over the weekend, He it by using the graohical ahd. [Dab Un, MEA. 2008 HINTS AND ANSWERS 1. (@)Altemative solutions ()_Infeasbe soli ant 5.2 0 (i) Unhounded soltion drs, #y =O; 4) 32 and Mak 2'= Wa 32 (ix) Unbounded solution (0) Redurdant 4S. Let x ands, = number of bel of types A and B produced, ‘onstrints are second and third ()_Infeasble solution Terie. (i) No feusiblesoltion (il) Usbounded Man Z = 040s, + 0.303 solution subject to (i) x, + 4) < 800: (ii) 2x, +x < 1,000 (2) x, = 2019, x, = 4519 and Max 2 = 535/19 Gi) 540; Gy) ry S700 2) No feasible solution and xp. 20 = 2 6 3," 8x," 6and Min Z © 20 Uns. x; = 200, % = 600 and Max = Rs 260 Gi) x)= 3.x, = 2 and Min Z = 19 6 Let x; and x, = numberof days plant at G and J workin July Gi) 5, ~ 63 ~ 12 and Min Z ~ 240 respectively. 3. Let x, and x, =number of pieces of P, and P, to be Min Z = 600, + 400, Suamutcne, special subject 10 () 1.500%, + 1.500% 2 20,000 Max 2~ SOx, + 100%, Gi) 0005, + (000%, > 40,000 subject to (i) 10x + Ie, < 2500 (ii) 2.000%, + $0001, 2 44,000 (ii) 4x, + 10x, < 2000 and Ky ap 20 (i) 3, + ax, s 430 drs. j= 12x, = and Min Z = 8800, and ey Lat, and 15 = numberof pars manufactured of automobile A Ans. x = 1875, x) = 125 and Max Z = Rs 21,87. Let andx, = number of chairs and tables produced, respectively. Max Z= 2x, + Hx, subject t0 i) 2c, + Sigs I Gi) 6x, < 30 and 3, respectively Mas Z = {S—2420/25-414/284175/38} + (6-3+20/40+14/35-+17.5/25} a2 = 4.80x, + 4.00%,‘Operations Research: Theory and Applications subject to Aye Asc @pAate o3+3<, WS+351, GW S+2<1 and aem20 Let x and xy = number of units of casting X and Y to be ‘manufactured, respectively. Max Z = (300r, + 360%) ~ (120%, + 120%,) ~ 1,000 = 180c, + 240x, ~ 1,000 subject to) 4x, + 2s, $240 Gi) 2 4 Sy 23» 40 am 20 20 and Max Z = 5,600. Let x; and x = number of production hours for plants A and B needed 10 complete the orders, respectively. Min Z = 9x, + 10x, subject to () 2x, +44, 290; (i) 4x, + 3x, 224 Gi) We, +28, > 60 and 20 Ans. x, = 3872 and x; = 15/4 and Min Z = Rs 195, Let x; and ay ~ number of units of products A and B, respectively. Max Z = 30x, + 40x, subject to) x; S20;.x;2 10, (i) 4x + 4x 5 100, j) 4r, + 6x, $180, (iv) x) +x, $40 and 20 Let x, and x, = number of units of A and B to be produces, respectively. Max Z = 40x, + 50x, subject to () 3%, +4) $9; (iH +2yS8 and x20 Ans. x, = 2,4, = 3 and Max Z = Rs 230. Let x, and x = number of chairs ard tables to be made, respectively Max Z = 20x, + 30x, subject to () 3x, +3x536; (ii) Sey +2x, $50 2e, + 6p < 60 eo oi fy = numb owl of AM tid ANT doe wie ont epee Max Z = 40r, + 80x, subject 10) 24, +3548; Gi) x, $15; Gi) x, $10 and xem? 0 Ans. x; = 9, x," 10 and Max Z = Rs 1,160. ‘number of units to be preduced of products A and B, respectively. Max Z = 2006, + 120%, Let x and x; a1 2, 25, | subject 10 (i) 4x, + 80x; < 800, (i) 10x, + Oe, < 80, Gi) pS 6 sd and xn 20 Ans. x, = 5, x, = 7.5 and Max Z = Rs 1,900. Let x,, 15 and x, = number of Deluxe, Standard and Economy packages restricted to 200 persons, respectively. Max (net profit) = [Price ~ (Hotel costs + Meals)] ~ Flat fee for the chartered aircraft = (10,000 ~ 3,000 ~ 4,750), + (7,000 ~ 2200 ~ 2,500}c5 + (6,500 = 1,900 ~ 2200), - 2,00,900 = 2,250, + 2300s, + 2400s, ~ 2,00,000 subject 10 (@) 20-< 3, $60 [fom () and (ii) conéition) © 260 [fiom (i) condition) © TSx 5140; @ x+y ts = 00; @ x +e,2 1 and x2.) 2 0 Since x, +, += 200, x)= 200—x, ~ x, Substituting value ‘of x, both in the objective fmetion and set of constraints, we set Max Z =~ 180x, ~ 100%) + 2,80,000 subject © () 20x, £60; Gi) 10< x < 110; Gi) 120 1 31.000 + £8.00 2 1 (i) 5j2,000 + y2,000 > 1 (6s) 13,000 + 539,000 2 1 and xp R20 Let x, aud xg = umber of feet of raw oak logs and raw pine logs tobe processed each week. respectively Max Z = 1,800e, + 1,200, subject 10 () 2x + Bx, 5 60, Gi) 4518, 20 (iy 244, + 12, 5 48, liv <2 andLinear Programming: The Graphical Method CHAPTER SUMMARY The graphical solution approaches (or methods) provide a conceptual basis for solving large and complex LP problems. A graphical method is used to reach an optimal solution to an LP problem that has a number of constrains binding the objective function. that has only two decision variables. Both extreme point methods and the iso-profit (or cost) function line method are used for graphically solving any LP problem CHAPTER CONCEPTS QUIZ True or False 1 2 ‘The graphical method of soving linear programming problem {is useful because ofits applicabily to many real fe stuations. “The problem of infeasibity fn tnear programming can only be solved ty making additonal resources avalable which in turn changes the constrains of the problem, ‘A linear programming problem is unbounded because constrains are incorrectly formulates |f.@ LP problem hae mero than one ection yielding the eame ‘objective functional value, then such optmal solutons are known as allernatve eptimal solutions. 5. Since the constraints to a linear programming are aways near, we can graph them by locating only two erent points 6. An optimal solution does not nacessarly use up al he limited resources avalable 7. The intersecton of any two constraits is an extreme point which ie-a comer of the feasible region. 8. When there are more than one optimal solution tothe problem ‘then the decision-maker will be uneble to judge me best ‘optmal solution among then. 9. If ll the constrains are > inequalties in a inear programming problem whose objective function is to be maximized, then the Solution of the problem is unboundee. 10. The problem caused by redundant constraints is that two Isoprofit lines may not be farallel to each other. Fillin the Blanks 11. Constrats in a tnear programming requiing al varables to be zero or positive are known as ‘constraints {and rest from limited resources are relrred as constraints 12, Instead of maximzing profi in linear programming problem, we ensure linearly of the objectve function by maximizing 13. The nuns PORIS of the convex set give the basic feasible Soluton tb the linear programming. 14. A basic feasble solution is said to be Ht the values fall variates are nenzere and poslive 418. fan opimal soluton to a linear programming problem exéts, He BE se (Of the feasble sokuton, 16. A constaint in linear programming must be expressed as Iinear a inear 17. When more than one solution best meets ihe objective of the linear programming problem ther it is said 10 have sat salutio, 18. re OOGUFS When NO value of the variable is able {0 satiay ‘all the constraints in linear programming problem simultaneousy, 19. AN existence of oojectve of the problem for ary tin Is one fof he major: of te near programming problem. 20. Any two isoproft or isocost ins fora Given inear programming problem are 10 each other 2 23, 24 26. 28. ‘Multiple Choice ‘The graphical metiod of LP_problem uses (a) objective funcion equation (©) constrain equations (c) linear equations (@) all of the above A feasible solution to an LP problem (@) must satsty all of the problem's constraints simultaneously (©) need not satisty all of the constaints. only ome of them (c) must be a comer point of the feasible region (@)_ must optimize the value of tho cbjoctve function Which of the folowing statements is tue with respect to the ‘optimal soluten of an LP problem (@) every LP problem has an optimal solation (©) optimal solution ot an LP problem always occurs at an extreme point (c) at optimal solution all resources are compitely used (@) if an optimal solution exists, there will aways be at least one at a comer An iso-proit Ine represents (a) an infinite number of sotions all of which yield the same proft (©) an infirite number of solution all cf which yisld the same (c) an hnfinite number of optimal solstions (@) @ boundary of the feasbie region It an iso-profit tne yieking the optimal solution coincides with constraint line, then (a) the soliton is unbounced (©) the solution is infeasible {c) the constraint whieh coincides Is redundant (@) none of the above ‘While plating constraints on a graph paper, terminal points on| both the axes are connected by a staigh! line because (a) the resources are tmited in supply (0) the objective function is a linear function (c) the constraints are linear equations or inequalites (@) all of the above ‘A constraint in an LP model becomes redundant because (2) two txo-profit ine may be pavallal to each ot (6) the soluton is untounced (@) thie conetraint ie rot edited by the aoluton value (@) none of the above It wo constraints do net intersect in the positive quarant of| ‘the grapa, then (a) the problem is infeasiblo (0) the soluton is untounced (c) one of the constraints is redundant (4) none of the above Constraints in LP problem ere called active if they (a) represent optinal solution (6) at optimality co not consume all the available resources2. cs 1 1 16. 19. 31 ‘Operations Research: Theory and Applications e) both of (2) and (b) (@) none of te above ‘The solution space (region) of an LP problem is unbounded due to (@) an incorrect fermultion of the LP model (©) objective function is unbounded {e) neither (a) nor (6) (a) both (a) and (b) ‘The while solvng @ LP model graphical, the area bounded by the constraints 's called (@) feasile region (c) unbounded solution constraints (c) two constraints are parallel eee 1 2F 38F 4 ron-nogative, structural 2. ‘an extreme pont 16 requirement 20. (22 (a) a ah @ 2) 321) {() infeasible region (d)_ none of the above ‘Alternative solitons exit of an LP model when (a) one of the constrants is redundant (©) objective function equation is parallel to one of the | (@) all of the above | 33. While solvng a LP problem, inleastilty may be removed by | (@)_ adding another constrain. | (©) adding another variable | {c) removing a constant | {@) removing @ variable 98. a non-redundant constant is removed from an LP problem, then | {@). feastte cexion wil become larger |B) teesibe region wil become smelir | {6) solution wil Bocome infeasible {@) none of the above 35. Tone ofthe constraint of an equation in an LP problem has | an urwourded soliton, ten |G) ‘solution to'such LP problem must be degererate | ©) asi rgion shouts have aire segment | (©) aerate soiions exist {@) none of to above T 87 67 727 B&F 8F wT Contribution 13. oxtremo 14. nondegenaratve, basis, equation, lnequalty 17. atematve 18, feasibly ravalel 25 @) 28) 7 2H (a) 2.) 3D. LC) @ 3Linear Programming: The Simplex Method “Checking the results ofa decision against its expectations shows executives what thelr sirengihs are, where they need 0 improve, ‘and where they lack knowledge or information’ ~ Peter Drucker Preview The purpose of this chapter is to help you gain an understanding of how the simplex method works. Understanding the underlying principles help to interpret and analyze solution of any LP problem Learning Objectives After studying this chapter, you should be able to ‘© understand the meaning of the word ‘simplex’ and logic of using simplex method, convert an LP problem into its standard form by adding slack. surpius and artificial variables. © set-up simplex tables and solve LP problems using the simplex algorithm, interpret the optimal solution of LP problems. revoggize special cases such as degeneracy, multiple optimal solution, unbounded and infeasible solutions. chapter Se Chapter Outline 44 Introduction 42 Standera Form of an LP Problem 43. Smplex Algorithm (Naximization Case) 44 Smplex Algorithm (Minimzaticn Cese) + Sof Practice Problems A eerie er pemees 45. Some Complicatons and heir Resolution 446. Types of Linear Programing Soluions + Concepul Quesions «Sel Pracice Problems B + Hints and Answers Ch Chesle Scam Chapter Concepte Quiz Caso study .BIZ Operations Research: Theory and Applications ‘Simplox method ‘examines ‘comer pponts of the feasible region, usng matrix row ‘operations, until an ‘optimal soluion Is, found 41 INTRODUCTION Most real-life problems when formulated as an LP model have more than two variables and are too large to be interpret with the help of the graphical solution method. We therefore need a more efficient method to suggest an optimal solution for such problems. In this chapter, we shall discuss a procedure called the simplex method, which is used for solving an LP model of such problems. Thus method was developed by G B Dantzig in 1947 ‘The simplex is an important term in mathematics, one that represents an object in an »-dimensional space, conrecting n + 1 points. In one dimension, a simples is a line segment connecting two points; in two dimensions, it is a triangle formed by joining three points; in three dimensions, it is a four-sided pyramid, having four comers The concept of simplex method is similar to the graphical method. In the graphical method, extreme points of the feasible solution space are examined in order to search for the optimal solution that lies at ‘one of these points. For LP problems with several variables, we may not be able to graph the feasible region, but the optimal solution will still ie at an extreme point of the many-sided, multidimensional figure (called an n-dimensional polyhedron) that represents the feesible solution space. The simplex method examines the ‘extreme points in a systematic manner, repeating the same set of steps of the algorithm until an optimal solution is found. It is for this reason that it is also called the iterative method. Since the number of extreme points (comers or vertices) of the feasible solution space are finite, the ‘method assures an improvement ia the value of objective function as we move from one iteration (extreme point) to another and achieve the optimal solution in a finite number of steps. The method also indicates ‘when an unbounded solution is reached. 4.2 STANDARD FORM OF AN LP PROBLEM ‘The use of the simplex method to solve an LP problem requires that the problem be converted into its standard form. The standard form of the LP problem should have the following characteristics: (All the constraints should be expressed as equations by adding slack or surplus and/or artificial variables. (ii) The right-hand side of each constraint should be made non-negative ifit is not already, this should be done by multiplying both sides of the resulting constraint by — 1. (iil) The objective function should be of the maximization type. ‘The standard form of the LP problem is expressed as: Optimize (Max orMin) Zc) x, +¢% +--+ y+ 08, +0594... + 05 subject to the linear constraints yy Xt Oi ayt oss Hyg ky +5, = by Gy) X, + ayy) + 12 Hyg ky +5) = By Opt Xt Ogg + + Aye XS and Ipkpecetphesyerndg 2 ‘The standand form of the LP problem can also be expressed in the compact form as follows: Optimize (Max orMin)Z= ¥ ox; + 05, (Objective function) in subject to the constraints Bays, +5, =, i= 12,..am (Constrainis) i and 48,20, forall ‘andy (Non-negativity conditions)Linear Programming: The Simplex Method In matrix noiations the standard form is expressed as: ‘Optimize (Max orMin)Z = ex +08 subject to the constraints Axts=b, and x,s20 where, €= (¢), C3... 0, )is therow vector, x = (X),Xp,.. 4.%y)!sD = (b,, Byy- «4 By)! aNd $= (5455839 +45) FE column vectors, and A fs the m x 1 matrix of coefficients of variables,x,.. 2 in the constrains, Remarks Theconstrained optimization (maximization or minimization) problem, may have 1. (@)_n0 feasible solution i.e, there may not exist values x, (= 1,2...) that satisfy every constraint, (6) aunique optimum feasible solution. (©). morethan one optimum feasible solution, i. alternative optimum feasible solution. (@ 2 feasible solution for which the objective function is unbounded, i.e. the value of the objective function can be made as large as possible in a maximization problem or as small as possible in a ‘minimization problem by selecting an appropriate feasible solution. 2. Any minimization LP problem can be converted into an equivalent maximization problem by changing the sign of 6's in the objective function. That is, Maine $95 Manin $5) 3. Any constraint expressed by equality (~) sign may be replaced by two weak inequalities. For example, i equivalent to following two simultaneous constraints, Oye Hay ay bag t= ay ey + aygty toot aghy Sb, and ayyxy + 449%) + 4, Three types of additional variables, namely (i) slack variables (s) (i) surplus variables (~ s) and (i) artificial ‘variables (4) are added inthe given LP problemto convert it into the standard form for the following reasons: (@) These variables allow us to convert inequalities into equalities, thereby converting the given LP problem into form that is amenable to algebraic solution. (b) These variables permit us to make amore comprehensive economic interpretation ofa final solution, (©) Help usto get an intial feesible solution represented by the columns of the identity matrix. ‘The summary of the extra variables to be added in the given LP problem in order to convert it into ‘a standard form is given in Teble 4.1. + apt, 2 b, + Less than or A slack variable 0 o Yes equal to (s) is added + Greater than or A surplus variable ° ° No equal to (2) is subtracted, and anartificial variable =~ M +M Yes is added + Equal to (=) Only an artificial +M Yes variable is added Remark 4 slack variable represents an unused resource, either inthe form of time on a machine, labour hours, money, warehouse space or any number of such resources in various business problems. Since these variables don’t yield any profit, therefore such variables are added to the original objective function with zet0 coefficients, A surplus variable represents the emount by which solution values exceed a resource. These variebles are also called negative slack variables. Surplus variables, like slack varisbles carry a 2er0 coefficient in the objective function. Definitions Basicsolution Given system of m simultaneous linear equations in n(> m) unknowns, Ax =B, where A is an m x n matrix and rank (A)=m. Let B be any m x m non-singular submatrix of A obtained by reordering m ‘Stack variable represents a (quanti of unused resource; it le added to less-than (oF equal Constraints in order te get an equally constraint Table 4.1 ‘Summary of Additional Variables Added in an LP ProblemOperations Research: Theory and Applications linearly independent columns of A. Then, a solution obiained by setting nm variables not associated with the columns of B, equal to zere, and solving the resulting system is called asic xoluion to the given system of equations. ‘Them variables, which may be all different from zero, are called basic variables. The m x mnon-singular sub-mattix Bis called a basis matris and the columns of Bas basis vectors. If Bis the basis sub-matrix, then the basic solution to the system of equations will be x, = B'b. Basic feasible solution A basic solutionto the system Ax= bis called basic feasible ifxy> 0. Degenerate solution A basic solution tothe system Ax=b is called degenerate if one or more ofthe basic variables vanish. Associatedcost vector Let xgbe abasic feasiblesolution to the LP problem Maximize Z=ex subject to the constraints Ax=b, and x20. ‘Then the vector €p = (Gp, y3+- + Cp)» called the cust vector associated with the basi feasible solution x, and ¢p. is coefficient of basic variable x 4.3. SIMPLEX ALGORITHM (MAXIMIZATION CASE) ‘The steps of the simplex algorithm for obtaining an optimal solution (if it exists) to @ linear programming, problem are as follows: ‘Step 1: Formulation of the mathematical model i) Formulate the mathematical model of the given linear programming problem. = (i) Ifthe objective function is of minimization, then convert it into one of maximization, by using the Bimount ot resource following relationship Jane esd Minimize Z = — Maximize Z*, where Z* =—Z. ands aed (it) Check whether all the 5 (/= 1,2, ...m) values ae positive. IF any one of them is negative, then Sturt consraints muliply the corresponding constraint by — | in order to make b, > 0. In doing so, remember to in oer gel change 8 < type constraint to a 2 type constraint, and vice verse. feasly constant (iv) Express the mathematical model ofthe given LP problem in the standard form by adding additional variables to the left side of each constraint and assign a zero-cost coefficient to these in the objective function. (v) Replace each unrestricted variable with the difference of the two non-negative variables; replace each non-positive variable with a new non-negative variable, whose value is the negative of the original variable, Step 2: Set-up the initial solution Write dowa the coefficients of all the variables in the LP model in a tabular form, as shown in Table 4.2, in order to get an initial basic feasible solution (x, = Bb. Com 5 ene elaisD: 1 Z= dept, et 0 0 0 0 0 ° Table 42 BLY. coelficients) = (BLV. coefficients) {nial Simpiet (Values of B.V.) —* (th column oF data matrix) Table o- ah G2 00 0Linear Programming: The Simplex Method BU ‘After having set up the initial simplex table, locate the identity matrix and column variables involved in it. This matrix contains all zeros except positive elements 1's on the diagonal. This identity matrix is always a square matrix and its size is determined by the number of constraints. The identity matrix so obiained is also called a basis matrix (because basic feasible solution is represented by B = 1} Assign the values of the constants (5's) to the column variables in the identity matrix [because xg- B"'b=1b=b]. ‘The variables corresponding to the columns of the identity matrix are called basic variables and the remaining ones are non-hasic variables. In general, ifan LP model has » variables and m ( 0 andall other elements are negative (i. ay < 0), then there exists an unbounded solution to the given problem. (Gil) If at least one c,—z,> 0, and each of these columns have at least one positive element (i.e. ay) for some row, then this indicates that an improvement in the value of objective fumetion Z is possible. Step 4: Select the variable to enter the basis If Case (ii) of Step 3 holds, then select a variable that has the largest cj ~ z, value to enter into the new solution. That is, 4724 = Max (6-2); G2)? 0} ‘The column to be entered is called the key or pivor column. Obviously, such a variable indicates the largest per unit improvement in the current solution. Such a variable therefore indicates the largest per unit, improvement in the current solution. Step 5: Test for feasibility (variable to leave the basis) After identifying the variable to become the basic variable, the variable to be removed from the existing set of basic variables is determined. For this, each number in xp-column (ie. 6, values) is divided by the corresponding (but positive) number in Degeneracy when there faa te in the minimum ratio. varlabe 10 enter Into the next solution the set of variables present in the sclution, have positive non-zero value and are Isted in the D(=X5) colurm of a ‘simplex table; these vatiables are abo called basic variables Non-basic variables aro tose hich are not in the "basis’ and have zero-valueOperations Research: Theory and Applications sttuation in which no ‘solution satisfies al constrains of an LP protlom @,~ row values ‘present net prot or bss resulting from invodueng fone unit of any arable into the “basis or solution the key column and a row is selected for which this ratio, [(constant column)/(key column)) is non- negative and minimum. This ratio is called the replacement (exchange) ratio. That is ay ay Xo = sin 80 2 = Min [is 5 >>} ‘This ratio limits the number of units of the incoming variable that can be obtained from the exchange. It may be noted here that division by a negative or by a zero element in key column is not permitted. ‘The row selected in this manner 1s called the key or pivot row and it represents the variable which will leave the solution. ‘The element that lies at the intersection of the key row and key column of the simplex table is called key of pivor element. Step 6: Finding the new solution (If the key element is 1, then the row remains the same in the new simplex table. (Gi) Ifthe key element is other than 1, then divide each element in the key row (including the elements {in Xp-column) by the key element, to find the new values for that row. (iii) The new values of the elements in the remaining rows of the new simplex table can be obtained by performing elementary row operations on all rows so that all elements except the key element in the key column are zero. In other words, for each row other than the key row, we use the formula: aes (ae a) lames xv) (ieee | the new row, that is row \eldieer fey Hemaees replaced in Step 6 (i) ‘The new entries in cy (coefMicient of basic variables) and xp (value of basic variables) columns are updated in the new simplex table of the current solution, ‘Step 7: Repeat the procedure Go to Step 3 and repeat the procedure until sll entries in the ,~ 2; Tow are either negative or zero. Remark The flow chart of the simplex algorithm for both the maximization and the minimization LP problem is shown in Fig. 4.1 Example 4.1 Use the simplex method to solve the following LP problem. Maximize Z= 3x, + Sxp + 4x5 subject to the constraints @ 2x, +3, $8, (i) 2xj+5ryS10, i) 3x + 2x, + 4x, < 15 and XpIyH20 (Rewa Msc (Maths). 1995; Meerut MS¢ (Stox.), 1996; MSc (Maths), 1998: Kurukshetra, MS¢ (Stat), 1998) Solution Step 1: By introducing non-negative slack variables 5}, s) and s, in order to convert the inequality constraints to equality the LP problem becomes: Maximize Z = 3x, + Sx, + 4x, + 0s, + 05, + 0s, subject to the constraints @ 2 +3q +H =8 GD 2m_tSxyts=10, (il) 3x, +2y +4xy +9515 and Hyp ye Se Spy 20 Step 2: Since all b,(RHS values) > 0, (i = |. 2, 3) we can choose the initial basic feasible solution as: 8,s,=10,s,=15 and Max Z=0 i This solution can also be read from the inital simplex Teble 4.3 by equating row-wise values in the basis (B) column and solution values (x) column. Stop 3: To sce whether the current solution given in Table 4.3 is optimal or not, calculate G-4=G~ & B= 5 ey, for non-basic variables x,, x, and x, as followsnear Programming: Te Simplex Method. 42; = (Basic variable coefficients, ¢,) * (jth column of data matrix) 2)+0(0)+0(3) = 0 forx,-column (3)+0(2)+0@) = 0. forx,-column 4 7 0(0)+0(5)+0(4) = 0 foray-cokumn ‘These values are now subtracted from c, valves in order to calculate the net profit. This is dane by introducing ‘one unit of each variable x,, x and x, into the new solution mix. G74 =3-053, G-%=5-0=5, c-2=4-0=4 The z, and cj ~z; rows are added into the Table 43. ‘The vahies of basic variables, 5, 4) and 23 are given in the solution values (xq) column of Table 4.3 ‘The remaining variables that are non-basic atthe current solution have zero value. The value of objective function at the current solution is given by Basic variable coefficients, ¢g) * (Basic variable values, Xp ) =0 (8) +0(10)4 0(15)~0 That is, SR ee Bon eae, ecloo- eclo-e G73 Table 4.3 ® 2 a 6 3 i Initial Solution Since all ¢,—z, 2 0 (/= 1, 2,3), the current solution is not optimal. Variable x, is chosen to enter into the basis because c, ~z, = 5 isthe largest positive number in the x,-column, where all elements are positive, ‘This means that for every unit of variable x, the objective function will increase in value by 5. The x column is the key column, ‘Step 4: The variable that isto leave the basis is determined by dividing the values in the xg-column by the comesponding elements in the key column as shown in Table 4.3. Sines the exchange ratio, 8/3 is minimum in row 1, the basic variable s, is chosen to leave the solution (basis). Step 5 (Iteration 1): Since the key element enclosed in the circle in Table 4.3 is not 1, divide all elements of the key row by 3 in order to obtain new values of the elements in tis row. The new values of the elements in the remaining rows for the new Table 44 can be obtained by performing the following elementary row operations on all rows so that all elements except the key element 1 in the key column are zer0. R, (new) — R, (old) + 3 (key element) ~ (8/3, 23, 3/3, 03, 1/3, 03, 0/3) = (8/3, 23, 1, 0, 1/3,0, 0) 1-2" 83-143 15-2 83293 0-2%23=-4/3 3-2%28-58 aaah 2-21 1-0 S-2% 08s 4-2 0-4 0-2*13=—25 0-25 13=-23 1-2* 0-1 0-2% 0-0 0-2 0 12% 0-1Operations Research: Theory and Applications. Eee woes 0 ® 4 0 4 eole-e Table 4.4 Improved. Solution =18 = 38 t ‘The improved basic feasible solution can be read from Table 4.4 as: x, = 8/3, s, = 14/3, 95 29/3 and x) = = 5, =0. The improved value of the objective function is Z = (Basie variable coefficients, ey ) x (Basie variable values, xy (8/3) +0 (1423) +0 (29/3) = 4023 Once again, calculate values of cj — in the same manner as discussed earlier to check whether the solution shown in Table 4.4 is optimal or not. Since c, — z, > 0, the current solution is not optimal. Heration 2: Repeat Steps 3 to 5, Table 4.5 is obtained by performing following row operations to enter variable x, into the basis and to drive out s, from the basis. R, (new) = R, (old) +'5 (key element) = (14/15, — 4/15, 0, 1, ~2/15, 1/5, 0) 293 -4% 14/is = sons 33-4 — 4/15 = 41/15 ona 0 4-4x 1-0 = 23-4 215 =-218 0-4 IS=-4 1-4» o=0 ‘Table 4.5 is completed by calculating the new z, and c)—z, values and the new value of objective function: 2=5(23) + 44/15) + ONS) =34/15 gy 34/15 = 1IMS for x,-column 2,2 5(13) +4215) + 01S) =1713 cy = ITS = =1713 for sy-column 5@) +40) + 0C45) = 4/5 6-732 0-45 = — 45 fors,-column The new objective function valve is given by Z = (Basic variable coefficients, ¢y) » (Basic variable values, xp) =5 (8/3) +4 (14/15) +0 (89/15)= 256/15 ‘The improved basic feasible solution is showa in Table 4.5. = 83 2 1 0 1 0 0] @ayas)=4 4 wis |-ais 0 1-75 os 0 - ° was |G 0 0 ais —4s_1| aonsyeainsy~ 217 > 2-25615 ous 54 mAs aso Table 4.5 MMs 0 0 -17S -45 0 Improved Solution tLineer Programming: The Simplex Method eration 3: In Table 45, since e, ~ zis still a positive value, the current solution is not optimal. Thus, the variable x, enters the basis and s, leaves the basis. To get another improved solution as shown in Table 4.5 perform the following row operations in the same manner as discussed earlier. By (new) +R, (old) « 15/41 (key element) > (BOIS x 15/41, 41/15 * 15/41, 0 « 15/41, 0% 15/41, 2/15 x 15/41, —4/5 » 15/41, 1 15/41) > (89/41, 1, 0, 0, ~ 2/41,— 12/41, 15/41) 83-28 x 803= S0/al 14S + 4/15 = 89/41 = 62/41 2329p x ae 0 =a +ansx t= 0 1-23 1 o+Misx c= 0 O-2* O- 0 VANS = on 1B - 23 x -24i= 15/41 DIS +415 *— 2/41 = — 6/41 O-2nx-241= gia) 1S + 4/15 «1241 = 5/41 0-28 x Iii = —10/a1 o+ans< 11 = aan 3 4 5041 ° 1 0 isa BL — 1041 4 e 241 0 0 1-641 Sai ait 3 a 8941 1 ° o -2H 11 asia Z= 76541 % 3 s 4 4s41 Hata 5 -§ ° 0 -asi1 = 261 = 14t In Table 4.6, all c, ~ zj < 0 for non-besic variables, Therefore, the optimal solution is reached with, 2, = 89/41, x) = 50/41, x, = 62/41 and the optimal value of Z = 765/41. Example 4.2 A company makes two kinds of leather belts N belt A and belt B. Belt A isa high quality belt and belt B is of lower quality. The respective profits are Rs 4 and Rs 3 per belt. The production of each of type 4 requires twice as much time as a belt of type B, and if all belts were of type B, the company could make 1,000 belts per day. The supply of leather is sufficient for only 800 belts per day (both A and B combined). Belt A requires a fancy buckle and only 400 of these are available per day. There are only 700 buckles a day available for belt B. ‘What should be the daily production of each type of belt? Formulate this problem as an LP model and solve it using the simplex method. Solution Let x, and x, be the number of belts of type A and B, respectively manufactured each day. ‘Then the mathematical LP model would be as follows: Maximize (total profit) 2 subject to the constraints ( 2, +x,$1,000 (Time availability), i) x, + x5 $ 800 (Supply of leather) (ii) x, <400 x, $700 and 1,220 13 | oauerrnonne ‘Standard form Introducing slack variables s,, s,s, and s, to convert given LP model into its standard form as follows, Table 4.6 Optimal Solution‘Operations Research: Theory and Applications Maximize Z = 4x, + 3x, +05) + Os, + 00, + Ory subject to the constraints © 2x, +x,4+5,= 1,000, Gi) x, +3 +5, = 800 Gi) x, +5, = 400, (iv) x +54 = 700 and Hp Xpp Sp Sy Sp 5420 Solution by simplex method An initial feasible solution is obtained by setting x, initial solution is: s, = 1,000, s = 800, s, = 400, s, = 700 and Max Z = 0. This solution from the intial simplex Table 4.7 ‘Thus, the an also be read 0 5; 1,000 Bet 1 76 0 | 1,000/2 = 500 0 4 y 4 0 1.0 0| 8001= 390 ° 5 400 @ o 0 0 1 0} 4001-400 ° 4700 Ce ne eS z-0 s Deo eOn TS be wu Oo TO. -% yeaa sop fe SC yee a) t ‘An Improved Solution to the basis and removing variable sy from the besis. Ry (new) > Ry (Old) + 1 (Key element) Ry (new) ~+ R, (Old) ~ 2Ry R, (new) — R, (old) ~ 2; (new) ‘The new solution is shown in Table 4.8 rin order to get an improved basic feasible solution by entering (new) o a 200 ° 18-20 | 2m = 20-5 o 5 400 Or 1 to | ator = aon: 4 x 400 fad Oot Te ho niet = o P; 700 Oo bo 0 ot | tant = tee! 7600 3 Age aso ea nami uccdy aU: 6-5 Dieseae wey ei femeah an t ‘The solution shown in Table 4.8 is not optimal because cy ~z» > 0 in.x,-column. Thus, again applying. the following row operations to get a new solution by entering variable x, into the basi variable s, from the basis, we get R, (new) ~ Ry (old) + | (key element) Ry (new) + R, (old) ~ R, (new) Rj (new) > R, (old) ~ f (new) ‘The improved solution is shown in Table 4.9, is and removingLinear Programming: The Simplex Method IRI 93 Heaton 6 «tence l Ceram 3 zy ORES. cer peearaD aa ° a 0 0 <1 1 @ 5} s0n=20- 4 x 20 0 6 to} anor = 400 ° % oo -1 0 tt | S00 250 Z= 220 ee 97 eS geet ah nile ee my Table 4.9 , Improved Solution The solution shown in Table 4.9 is not optimal because c; ~ z5 > 0 in s,-column, Thus, again applying the following row operations to get ¢ new solution by entering variable, s, into the basis and removing variable 5, from the basis, we get Ry (new) ~ R, (old) + 1 (Key element) Ry (new) — R, (old) + 2R, (new) Ry (new) + R, (old) = R, (new); Rg new) — R, (old) — 2R, (new) ‘The new improved solution is shown in Table 4.10. 6 ‘ 3 0 0 o o 3 % 00 ° ope ° ° 0 % 200 ° o 1 1 ° 4 x 200 1 0 1 0 0 ° 4 100 ° © 1 ° 1 Z= 2600 7 3 7 0 0 Table 4.10 ° Osea ° 0 Optimal Solution Since all cj 2; < 0 correspond to aon-basic variables columns, the current solution cannot be improved further. This means that the current basic feasible solution is also the optimal solution. Thus, the company ‘must manufacture, x, = 200 belts of type A and x, = 600 beits of type B in order to obtain the maximum profit of Rs 2,600. Example 4.3 A pharmaceutical company has 100 kg of A, 180 kg of B and 120 kg of C ingredients available per month. The company can use these materials to make thre: basic pharmaceutical products namely 5-10-5, 5-5-10 and 20-5-10, where the numbers in each case represent the percentage of weight of A,B and C, respectively, in each of the products. The cost of these raw materials is as follows: 4 80 6 20 ics 50 Inert ingredients 20 ‘The selling prices of these products are Rs 40.5, Rs 43 and 45 per kg , respectively. There isa capacity restriction of the company for product 5-10-5, because of which the company cannot produce more than 30 kz pet month, Determine how much of each of the products the company should produce in order to maximize its ‘monthly profit. [Delhi Univ, MBA, 2004, AME, 2005] Solution Let the P,P, and P, be the three products to be manufactured, The data of the problem can then be summarized as follows:‘Operations Research: Theory and Applications Table 4.11 Initial Solution zB a er Pe % SK 1m y 2% MHS Cost per'kg (Rs) 802052085 Cost of P, =5% <80+10%%20+5% *50-+80% x 20-4 +2 42.504 16= Rs 2450 per ke Cost of P,=9% *80+5% 204 10% *50-+80% «20=4 +145+16 =Rs 26 perkg Cost of P, = 2% x 804+ 5% 204+ 10% <50-+65% «20=16+1+5+413 =Rs35 perkg Let x xy and xy be the quantity (in kg) of P,P and Py, respectively to be manufactured. The LP problem can then be formulated as Maximize (netprofit)Z = (Selling price - Cost price) « (Quantity of product) = (40,50 ~ 24.50) + (43 ~26) xy + (45 ~35)x,= Ir, + 1729+ 10, subject tthe consiaits wi gh ort esting sasmransdean ga thn +s s 180 or 2x, +444, $3,600 4 20"! 530 aia aya Standard form Introducing slick vatnble r,s, andy 10 Sonvert the given LP mode into ite standard form as flows: Maximize Z = 16, + 17x, + 10%, +0s, + 55+ 055 +05, subject to the constraints sx tart = 2,000, Gi) 2s, Hay +45 +4 3,600 Gi) 1,424,424, += 2.400, Gv) x 5,= 30 and Hy Sp Sp Sy S420 1 1 Hyp tet 7gts S120 oF x14 24 + 2x, 2,400 Thus, Solution by simplex method \n initial basic feasible solution is obtained by setting x, =x, ‘the initial solution shown in Table 4.11 is: = 2,000, s, = 3,600, x5 = 2,400, s, = 30 and Max Z= 0. G7 MBIT FO 0. uO Oh 0 2,000 1 1 4 1 0 0 | 20001 = 2,000 0 3,600 2 1 1 0 t 0 0 | 3.6001 = 3,600 o 2.400 1 @ 2 0 0 14 8 | 24002 = 1,200 ° 20 Sue eele. Sono aS 3 ot ERO a oI OTe Bee |e Med hs Onn DS Since ¢y in onder to get a new improved solution by entering variable from the basis. R, (new) ~ R, (old) + 2 (Key element) Ry (new) > R, (old) — R; (new) 2, = 17 im x,-column is the largest positive value, we apply the following row operations into the basis and removing variable s, (new) — Ry (old) ~ Ry (new)Linear Programming: The Simplex Method FEE ‘The new solution is shown in Table 4.12. oF 6 17 0 0 9 0 9 0 4% so [12 0 3 1 0-12 0] son)= 1600 o Sy 2,400 32.0 «20 «80 «61 -12 0] 2,400%3/2)= 1.600 ” 4 120 | v2 1 1 0 © 12 9} t200;72)= 2400 6 a 0 @le 0. oe he 30 = 30> Z= 20400 3 mo 7 0 0 Wao ig twa ay tO 8s cae ieee ale Improve Solution ‘The solution shown in Table 4.12 is not optimal because ¢, ~z, > 0 in.x,-column. Thus, again applying the following row operations to get a new improved solution by entering variable x, into the basis and removing the varisble s, from the basis, we get Rg (new) —> Ry (old) + | (Key element), Ry (new) > R, (old) ~ (1/2) R, (new) Ry (new) R, (old) ~ (3/2)R, (new); __R, (new) - R, (old) - (1/2) R, (new) ‘The new solution is shown in Table 4.13. g7 6 7 wo o 0 0 0 5 785 ae schagel 0-12 -1R 0 5 2355 ° ° phi 1-12-30 v7 i 1185 o 1 12 one Paes 16 a 30 1 6 one 0 0 1 Z= M25 3 [Se i OA OIA) © Table 419 g-y | 0 0-70 0 192 ~ is Optimal Soliton Since all cj ~ zj < 0 corresponding to non-basic variables columns, the current solution cannot be improved further. This means that the current solution is also an optimal solution. Thus, the company must ‘manufacture, x, = 30 kg of P,, x, = 1,185 kg of P, and x, = 0 kg of P, in order to obtain the maximum ret profit of Rs 20,625. 4.4 SIMPLEX ALGORITHM (MINIMIZATION CASE) In certain cases as will be discussed below, it is difficult to obtain an initial basic feasible solution. Such cases @ when the constraints are of the < type Eayx, 0). In this case, we add attficial variables, 4,(/~ 1,2, ...m) to getan initial basic feasible solution. Theresulting system of equations then becomes: Bent a- %pSp4j20, 751,20. ‘and has m equations and (n+ m + m) variables (.c.n decision variables, m artificial variables and m surplus variables). An initial basic feasible solution of the new sysiem can be obtained by equating (n+2m—m) = (n + m) variables equal to zero. Thus the new solution to the given LP problem is: 4, 6,(7= 1, 2, ..., m), which does not constitute a solution to the original system of equations because the two systems of equations are not equivalent. Thus to get back to the original problem, artifical variables must be dropped out of the optimal solution. There are two methods for eliminating these variables from the solution, © Two-Phase Method * Big-M Method or Method of Penalties ‘The simplex method both for the minimization and the maximization LP problem may be summarized in the form of a flow chart. (Fig. 4.1) > Convert LP mndel ato Standard orm by adding eter slack arabes, surplus variables andlor arificial variables 1 Dee coefficient ofthese variables inthe objective function T ‘Setup inital simplex tablet obtain inal solution T ‘Compate sjand g 5 aloes ) Maximizton of Max or Min pe? I Do No 5-4 This ston opal ils os? Yen ‘Select key Solari with Wve oz value largest positive ej value = Sekt ey row with Min (xj ay a> 0) + Ill ze, then curt sltion _unkounded and stop the procedare 1 Ticaity kay clement the tersectionof key ow ard key column, T ‘pine te cnn i he Simplex be by (@) fist obtaining key row values, and (b) apply elemestary row operationsLinear Programming: The Simplex Mothod BERA Remark Arificial variables have no meaning in a physical sense and are only used as a tool for ‘generating an initial sotution to an LP problem. Before the final simplex solution is reached, all artificial variables must be dropped out from the solution mix. This is done by assigning appropriate coefficients to these variables in the objective function. These variables are added to those constraints with equality © and greater than or equal to (2) sign. 4.4.1. Two-Phase Method In the first phase of this method the sum of the artificial variables is minimized subject to the given constraints in order to get a basic feasible solution of the LP problem. The second phase minimizes the criginal objective function starting with the basic feasible solution obrained at the end of the first phase. Since the solution of the LP problem is completed in two phases. this is called the two-phase method. Advantages of the method 1. No assumptions on the original system of constraints are made, i. the system may be redundant, inconsistent or not solvable in non-negative numbers. 2. Itis easy to obtain an initial basic feasible solution for Phase I 3. The basic feasible solution (if it exists) obtained at the end of phase I is used to start Phase 1 ‘Steps of the Algorithm: Phase! ‘Step 1 (a): Ifall the constraints in the given LP problem are of the (=) type, then Phase Tl can be directly used to solve the problem. Otherwise, the necessary number of surplus and artificial variables are added to convert constraints into equality constraints (©) Ifthe given LP problemis of minimization, then convertit to the maximization type by the usual method. ‘Step 2: Assign zero coefficient to each of the decision variebles (x,) and to the surplus variables; and assign — 1 coefficient to each of the artificial variables. This yields the following auxiliary LP problem. Maximize Z*= $ (-1) 4, subject to the constraints Say x, +4, ia and x, 4,20 ‘Step3: Apply the simplex algorithm to solve this auxiliary LP problem. The following three cases may arise at optimality. (@ Max Z*=0 and atleast one artificial varisble is present in the basis with postive value. This means that no feasible solution exists for the original LP problem. (©) Max Z” = 0 and no artificial variable is present in the basis. This means that the besis consists of only decision variables (s/s) and hence we may move to Phase II to obtain an optimal basic feasible solution on the original LP problem. (© Max Z* = 0 and at least one artificial variable is present in the basis at zero value. This means that 0 feasible solution to the above LP problem is also a feasible solution to the original LP problem. Now in order to arrive at the basic feasible solution we may proceed directly to Phase I or else eliminate the artificial basic variable and then proceed to Phase Ul Once an artificial variable has left the basis, it has served its purpose and can, therefore, be removed from the simplex table. An artificial variable is never considered for re-entry into the basis. Remark The LP problem defined above is also called an auciliary problem. The value of the objective function in this problem is bounded from below, by zero, because the objective function represents the sum of artifical variables with negative unit coefficients. Thus, the solution to this problem can be obtained in a finite number of steps. Phase Il: Assign actual coefficients to the variables in the objective function and zero coefficient to the axtificial variables which appear at zero value in the basis at the end of Phase I. The last simplex table of ‘An artifeial vvarlable Is added 10 the LP constraints to get an inital soliton to an LP problema ‘Operations Research: Theory and Applications Phase I can be used as the initial simplex table for Phase II. Thea apply the usual simplex algorithm to the modified simplex table in order to get the optimal solution to the original problem. Artificial variables that do not appear in the basis may be removed. Example 4.4 Use two-phase simplex method to solve the following LP problem. Minimize Z= x, + x, ‘Subject to the constraints @ Ax +yz4 and xpay20 Solution Converting the given LP problem objective function into the maximization form and then adding surplus variables s, and s, and artificial variables 4, and A,, in the constraints. The problem becomes: ‘Maximize Z* =x, ~x) ‘subject to the constraints @ y+ H-5, 44-4 and Bs Sp spp Sp dy Ay 20 where Z"=—Z Phase I: This phase starts by considering the following auxiliary LP problem: Maximize Z* =~ A, ~ Ay ‘subject to the constraints Ot HH tA WD xt IHW 44: and apap sp sy Ap Ay 20 ‘The initial solution is presented in Table 4.14. @ x +7%27 @ x +--+, 5 6 0 o Ost 1 =i 4 4 2 eS 0 aL OY Si oe 7 iC? Olt ese Hai 3 ae 1 1 SE Seeey Table 4.14 Ga 3 ogee: wy Initial Solution t Atificial variables 4, and 4, are now removed, one afer the other, maintaining the feasibility ofthe solution Iteration 1: Applying the following row operations to get an improved solution by entering variable x, in the basis and first removing variable 4, from the basis gives us the results shown in Table 4.15. Note that the variable x; cannot be entered into the basis ss this would lead to an infeasible solution, Ry (new) > R, (old) + 7 (key element); R, (new) > R, (old) ~ R, (new) q~ ° ° 0 Gy nal emt “1 A 3 Broo =1 1 ° # 1 a Que ons Table 4.15 Fas = aan pr aie Improved 5-4 to woo Solution 4 i‘ * This column can permanently be removed at this stage.Linear Progamming: The Simplex Method BRE Iteration 2: To remove A, from the solution shown in Table 4.15, we enter variable s in the basis by spplying the row operations shown in Table 4.16. Here it may be noted that if veriable x, is chosen as the value to be entered into the basis, it will lead to an infeasible solution By (new) > R, (old) «7; Ry (new) > R, (old) + (U7) R, (new) o> ° 0 0 0 -1 o a a 8B Ba 1 ea o a 4 2 x =I 0 : 0 fae 5 0 0 oo 0 0 Table 4.16 6-4 ° ° 0 0-1 1 improved “Solution * Remove columns 4, and 4, from Table 4.16 Since all ¢ - = $ D correspond to non-basic variables, the optimal solution: x, =0, x, =4, 4) =0,s,=21, 4, =0, 4, =0 with 2° = 0 is arrived at. However, this solution may or may not be the basic feasible solution to the original LP problem, Thus, we have to move to phase II 10 get an optimal solution to our original LP problem. Phase I: ‘The modified simplex table obtained from Table 4.16 is represented in Teble 4.17. 97 at <1 ° 0 a a @ o > 1 2/3 > oe 4 2 1 -1 0 40 Ba-4 2 a2 = 1 0 Ae gr4 1 ° “1 ° zs Table 4.17 eration 1: Introducing, variable x into the basis and removing variable s from the basis by applying the following row operations gives as R, (new) — R, (old) + 13 (key element); —__R, (new) -> R, (old) ~ 2R, (new) ‘The improved basic feasible solution so obtained is given in Table 4.18. Since in Table 4.18, cj ~ <0 for all nor-basic variables, the current solution is optimal. Thus, the optimal basic feasible solution to the given LP problem is: x, = 21/13, x, = 10/13 and Max Z* =~ 31/13 or Min Z = 31/13, 97 af 1 o o ay * 8 1 0 =713 13 = 1013 0 1 3-203 Zn 33 a lee 6/13 3 Table 4.18 9-3 ° 0-613 13 Opiimal Solution Example 4.5 Solve the following LP problem by using the two-phase simplex method. Minimize Z =x, - 2x, ~ 3%, subject t© the constraints @ ~2e, + + 3x, = 2, (2%, + 3x, +41 and ayy HzO. [Meerut, MSc Math), 1996] ‘opyrighted materOperations Research: Theory and Appications Table 4.19 Initia! Solution Solution After converting the objective function into the maximization form and by adding artificial variables 4, and A, in the constraints of the given LP problem, the problem becomes: Maximize Z* =~ x, + 2) + 3x5 subject to the constrains © -2n, + yt 3544, and ye Fp Xy Aye dy 20 where Z* =~ Z Phase I: This phase starts by considering the following auxiliary LP problem: Maximize Z” =— 4, — Ay subject to the constrains @ -2x, +4435 +4, and Kp Xp ky Ay dy 20 ‘The initial solution is presented in Table 4.19. go 0 0 0 Sy <1 , @ Wy +3 tay t Ay @) 2x, + 3x, + 4x, + A, =D o-% 0 4 . 0 01 To first remove the artificial variable 4, from the solution shown in Table 4.19, introduce variable 2x, into the basis by applying the following Tow operations: R, (tiew) + Ry (Old) + 4 (Key element) ; Ry (new) +r R, (old) ~ 3Ry (new) ‘The improved solution so obtained is given in Table 4.20. Since in Table 4.20, ¢,—z) <0 corresponds tonon-basic variables, the optimal solution is: x, = 0, x, = 0,5 = 1/4, A, = 5/4 and 4, = 0 with Max = —5/4, Butat the seme time, the value of Z” <0 and the artificial variable 4, appears in the basis with positive value 5/4. Hence, the given original LP problem does net possess any feasible solution. Example 4.6 Usctworphase simplex methodto solve following LP problem Maximize 2=2x, +2, +21, subject to the constraints 5x, 47x: 44%, $7, (il) ~4x,+7H+Se, 2-2, (il) 34, +44, 8x, 2 297 and xa 0. [Punjab Univ, BE (E & C) 2006 : BE (IT) 2004) Solution Since RHS of constraint 2 is negative, multiplying it by ~ 1 on both sides and expressed it as: Ax = Tr —5x5 82. Phase I: Introducing stack, surplus and artificial variables in the constraints, the standard form of LP problem becomes: Copyrighteimage not availableOperatons Research: Theory and Applications = i u Table 4.24 Z=25n 3 =F In Table 4.24, c, —2,=92 i the laxgest positive value, replacing basic variable s, with non-basie variablex, into the basis. For this spply necessary row operations as usual. The new solution is shown in Table 4.25. 8 oF 302002 a 0 0 2 a 2 0 1 0 kat asa —a101 Z e 6 © 1 37821 sa 632 3 - 1 1 0 0 ezsttSesz1 TSR Table 4.25 223 3 2 «2 «Tw (aSISRL Optimal Soluton 0 0 0 _~27wis1 — 9421 — 65/521 44 = Oand Max 7325/7, 4.4.2. The Big-M Method ‘The Big-M method is another method of removing artificial veriables fiom the basis, In this method, we assign large undesirable (unacceptable penalty) coefficients to artifical variables from the point of view of, the objective function, Ifthe objective function Z is to be minimized, then a very large positive price (called penalty) is assigned to each artificial variable, Similarly, if Z is to be maximized, then a very large negative price (also called penaity) is assigned to cach of these variables. The penalty is supposed to be designated by —M for a maximization problem and + M for a minimization problem, where M> 0. The Big-M method for solving an LP problem can be summarized in the following steps. Step 1: Express the LP problem in the standard form by adding slack variables, surplus variables and artificial variables. Assign a ze10 coefficient to both slack and surplus variables. Then add e very large positive coefficient + M (minimization case) and— M (maximization case) to artificial variable in the objective function. ‘Step 2: The initiel basic feasible solution is obtained by assigning zero value to original variables. ‘Step 3: Calculate the values of ¢, ~ 2; in last row of the simplex table and examine these values. ( Wall c,~z) 20, then the current basic feasible solution is optimal Gi) IF for a column, k, cy ~ 2, is most negative and all entries in this column are negative, then the problem tas an unbounded optimal solution, (Gi) If one or more ¢, ~z, <0 (minimization case), then select the variable to enter into the basis with the largest negative c, ~ 5, value (largest per unit reduction in the objective funetion value). This ‘value also represents the opportunity cost of not having one unit of the variable in the solution. That is, ep #4 Min (6-51 6) -5)< 0} ‘Step 4: Determine the key row and key element in the same manner as discussed in the simplex algorithm of the maximization case. ‘Step 5: Continue with the procedure to update solution at each iteration tll optimal solution is obtained.image not availableOperations Research: Theory and Applications Heration 2: Since the value of c; ~ 2, in Table 4.28 is the largest negative value, variable x, is chosen to enter into the basis. For introducing variable x, into the basis and to remove s, from the basis we apply the following row operations. Ry (new) ~> R, (old) « 5/16 (key element); Ry (new) ~+ Ry (old) ~ (6/5) Ry (new). R, (new) ~ Ry (old) 2/5 R, (new). ‘The new solution is shown in Table 428. 3 S sa [a 1 sis © | Gevcim) = 40 M 4 3 6 0-38 1 | 3/04) = 12> 3 ie 1 loo 18 “Ho z= 2a aM y | 5 eames sie Wem wt f Table 428 gy | o GMR-si6 M4478 0 | Improved Solution t | Heration 3: Asc, —z, <0 in sy-column, the curreat solution is not optimal. Thus, introduce s into the basis and remove 4, ffom the basis by applying the following row operations: Ry (new) —> Ry (old) * 4 (key element, Ry (new) —» R, (old) ~(1/8) Ry (new) Ry (new) — Ry (old) + (1/4) R, (new. ‘The new solution is shown in Table 4.29. o> 5 3 0 0 3 ei 1 ° 1 0 0 2 ° ° 1 8 a 4 1 ° ° Table 429 2-3 x 5 3 Ss ° Optimal Solution eg 0 0 1 0 In Table 4.24, all cj ~ z) = 0. Thus an optimal solution is arrived at with the value of variables xa 4.y= ls, and Min Z = 23. Example 48 Use penalty (Big-M) method to solve the following LP problem. Maximize Z=.x, + 2s, +3x,— x4 subject to the constraints @ xjt2y +3415, fi) 2p +x,75x,=20, Gil) x, + 2y tj +x4= 10 and Xp ty ty X20 [Caticut, Bech. (Engg), 1995; Bangalore BE (Mech. 1995; AMIE, 2004] Solution Since all constraints of the given LP problem are equations, therefore we should add only artifical variables 4, and 4, in the constraints, The standard form of the problem is then stated as follows: Maximize 2 xj + 2r + 3x, ~-xy- MA, ~ Md, subject to the constraints OxtrRmt Mt AA=IS Wty + Sy + A= 70 ii) x, + 2x +x, +4 =10 and Xp Xp ty ty Ay 4,20 An initial basic feasible solution is given in Table 4.30.image not availableimage not availableSSE eee aie Sinciee Nettod) As c, ~ Z value in x,-column of Table 4.34 is the largest negative, therefore variable x, should be entered 18 replace basic variable 4, into the basis. For this, apply following row operations Ry (aew) ~ Ry (old) +2 (key element); Ry (new) > R, (old) ~ Ry (new) to get the new solution as shown in Table 4.35, 600 son 0 0 u Mw ar @® GE ade 1 | 1003 500 a 0 12 i 0 = o| Z= 15,000 + SM 3 | 3Mz+250 500 -M = «M2-250 i 9% ~ 3M2 250-M2 0 - le 4. 94 ne ae Improved Soluton Asc —z) valuc in xj-column of Table 4.35 is the largest negative, enter variable x, to replace basic variable 4, into the basis. For this, apply row operations R, (new) —> Ry (old) * (2/3) (Key element) ; to get the new solution as shown in Table 4.36. 2, (new) — R, (old) ~ (1/2) R, (new) o> 500 0 0 ‘600 x 1 = 8 500 e 0 1 3-29 ‘2 80,0008 0-500 — 7003 — 4003 0 70034003 In Table 4.36, all c,— z; = 0 and also both artificial variables have been reduced to zero. An optimum solution has been arrived at with x, — 100/3 batches of hardcover books, x, = 40/3 batches of paperback books, ata toxal minimum cost, Z = Rs. 80,0003, Example 4.10 An advertising agency wishes to reach two types of audiences: Customers with annual income greater then Rs 15,000 (target avdience A) and customers with annual income less than Rs 15,000 (target audience B). The total advertising budget is Rs 2,00,000. One programme of TV advertising costs Rs 50,000; one programme on racio advertising costs Rs 20,000. For contract reasons, at least three programmes ought to be on TV, and the number of radio programmes must be limited to five. Surveys indicate that a single TV programme reaches 4,50,000 customers in target audience A and 50,000 in target audience B. One radio programme reaches 20,000 in target audience A and 80,000 in target audience B. Determine the media mix to maximize the total reach. [Delhi Univ., MBA, 1995] Solution Let x, and.x, be the number of insertions in TV and radio, respectively. The LP problem can the be formulated as follows: Maximize (total reach) Z (4,50,000 + 50,000) x, + (20,000 + £0,000) x, = 5,00,000 x, + 1,00,000.x, = 5k) +, subject to the constraints © 50,000 x, + 20,000 x, <2,00,000 or Sx, +220 (Advt. budget) (@ x)23_ (Adv. on TV) ii) x $5 (Adut. on Radio) and xp% 20. Standard form After introducing slack/surplus and/or artiicial variables in the inequalities of the constrains, the LP problem in standard form becomes: Maximize Z = Sx, + x, +05, + Os, + Os, — MA subject to the constraints © Sx, + 2x, +5, = 20, Rp Xp Sp Spe 8y Ay Gi) tye and Table 4.36 Optimal SolutionBE) Operations Research: Theory and Applications Solution by simplex method An initial basic feasible solution is obiained by setting xy =x, =.= ‘Thus, s, = 20, A, =3, 53 = 5 and Max Z =~ 3M. This initial solution is shown in simplex Table 437. 93 5 1 0 0 0 -M 3 SN Ser Gor pees et o nee rate rane as =i 5 SMO, Ae Inia! Solution Bee Be eee | ‘As ¢;~ =; value inx;-columnm of Table 4.37 isthe lamest positive value, enter variable 3, to replace ‘asic variable 4 into the basis. For this aply following row operations Ra (new) > R, (Old) + I (key element); R, (new) -> R, (old) — SR, (new) to get the new solution as shown in Table 4:38, gion of Baie ot aie 5 ° Dee ° 3 1 0 0 a 5 6 a 1 Z=18 % 5 oe aes Table 4.38 4 ° fen Oss ae | Improved Sokition + { ‘The solution shown in Table 4.38 is not optimal as cy ~ 2, value in sy-coluran is the largest positive. ‘Thus, enter variable 5; to replace basic variable s, into the basis. For this epply the following row operations: Ry (new) -> R, (old) + 5 (key element); Ry (new) > R, (ald) * R, (new) to get the new solution, shown in Table 4.39. sy 0 as nse 5 1 2 ws o ae 0 i 0 0 Table 4.39 3 z 1 0 Optimal Solution sy a = ie Eas Since all cj~z, 2 0in Table 4.38, the total reach of target audience cannot be increased further. Hence, the optimal 4 insertions in TV and x,= 0 in radio with Max (otal audience) Z = 20.00,000. Example 4.11 An Air Force is experimenting with three types of bombs P, Q and in which three kinds of explosives, viz., A,B and C will be used. Taking the various factors into accoun, it has been decided touse the maximum 600 kg ofexplosive A, atleast 480 kg of explosive B and exactly 540 kg of explosive C. Bomb P requires 3,2, 2kg, bomb Q requires 1,4, 3kg and bomb R requires 4, 2,3 kg of explosives A, Band respectively. Bomb Pi estimated to give the equivalent of a 2 ton explosion, bomb Q, a 3 ton explosion and bomb R, a 4 ton explosion respectively. Under what production schedule can the Air Force make the biggest bang? pytighted rimage not available‘Operations Research: Theory and Appications ° 5 3 % 4 % Table 4.42 Z= 660 Optimal Solution 43 ° 0 1 3 wh 1 0 0 oun 1B 0 1 0 i ry 3 4 0 2 =A ° ° ° an In Table 442, all ¢,~z) $0 and artificial variables A, and A, have been reduced to zero. Thus, an optimal solution has been arrived it with x, = Obombs of type P, x; = 60 bombs of type Q.x, largestbenefit of Z = 660. 20 bombs oftype R, at SELF PRACTICE PROBLEMS A 1. television company has three major departments for maniac turing two of its models — A and B. The monthly capacties of the departments are given as folows: Per Unit Time Requirement (hours) Hours Available Moaer a —_-Mocer 5 th Month Depanment 4.0 20 1,800 Deparment i! 2.5 10 1,200 Deparment ill 4.5 15 1,600 ‘The company makes a proft of Rs 1,000, 4,000 and 5,000 on ‘each tonne of the products A, B and C, respecively. HOW many tonnes of each product should the company produce in order to maximize its profs ‘A manufacturing firm has discontitued the production of @ ‘contain unprofitable product line. This has crested considerable ‘excess production capaci. Management is considering 10 dovoting this excess capacty to one or more of three products; ‘cal them product 1, 2 and 3. The avaiable capscity on the ‘machines that might limit output is summarized in the folowing table: ‘The marginal profit per unit from model A is Fs 400 and from ‘model B is FS 100. Assuming that the company can sel any ‘quantity of sither product due 10 favoureble market conditions, determine the optimum ouput for beth the models, the highest [essible profit for this month end the siack time in the three epartments, 2. A manulacturer of leather belts mates tree types of belts A, B and C which are processed on three mactines M,, Mz and. ‘My, Belt A requres 2 outs on machine M, and 3 nours on ‘machine M, and 2 hours on machine M,, Belt B requires 3 hours. ‘on machine M,, 2 hours on machine M, and 2 hours on machine M, and Belt C requires 5 hours on machine M. and 4 hours on ‘machine M, There are 8 hours of time per day available on ‘machina My, 10 houre of tine por day avallablo en machine M, {ard 1Shours of tme per dey available on machine My. The profit {gained from belt A is Fs 3.00 per unit, rom Belt 8 is As 6.00 er uni, from bel C is Rs 4.00 per unit. What should be the daly production of each type of bel so that the products yeld fe ‘maximum profit? 3. A company produces three products A, B and C. These products require three or O,, Q, and” Oy The maximum ‘quantities of the ores 0,, Q, and Q, avaliable are 22 tonnes, 14 tonnes and 16 tones, respoctvely. For one tome of each of these products, the ore requirements are: A 8 c 3, 3 = 3 o, 1 2 3 3 a 2 a Proft per tonne 1 4 5 (Rs in tmousana) ‘Machine Type ‘Avalladie Time (ia. Mactine-hours por Week) Ming Machine 250 lathe 150 Grinder 50 ‘The number of machine-hours required for each unit of the respective product is as follows: Machine Tape Procuatty (in Machine-heurs or Uni) Product —‘Product—~C*Product 2 2 ling Machine 8 2 3 Lathe 4 a ° Grinder 2 = 1 ‘The pratt per unit would be Rs 20, Rs 6 and Rs 8, respectively {or product 1, 2 and 3. Find how much of each product the firm should produce in order to maximize its profit [Dalhi Univ, MBA 2003) ‘A larmer has 1,000 acres of land on which he can grow com, ‘wheat or soyabean. Each acre of com costs Rs 100 for reparation, requires 7 mer-days of work and yields a profit of Fis0, An acre of wheat costs Rs 120 to prepare, requires 10 rmen-days of wort and yields a proft of Re 40. An acro of ‘soyabean costs Fis 70 to prepare, requires 8 men-days of work ‘and yieids @ profit of As 20. Ifthe fermer has Fs 1,00,000 for, ‘reparation and can count on 8.000 men-days of work, deter- ‘mine now many acres shoud be allocated to each crop in order to maximize profis?8. The annual handmade furniture show and sale Is supposed 10 Linear Programming: The Simplex Method BERT take place next month and the school of vocational studies is | Mixture ‘Specifications Selling Price (Rs/ig) i Having blo Uutin br sus Te es Seodorn spe! ya I yeu ana hecho See Get 5 and they have decided to make styles of chairs - A, 6 and C. 2 _ Each chair must receive work in each class, The time in hours: and quantity of type 3 ‘required for each chair in each class is: 8 2 35% type 1 4 is a5 ye 2 Gia eur Yor‘ Yor snd ay of 4p 9 A 2 3 3 For tam se pnt eps a manag cn = & 3 2 ‘are as follows: é 2 ' i During the next month there will be 120 hours available to the | year ciass, 160 hours to the Il year class, and 100 hours to the Ill year class for producing the chairs. The teacher of the wood.working classes feels thet a maximum of 49 chairs can be soldat the show. The teacher has determined that the profit ‘om each type of chair wil be: A, Fs 40; B, Rs 85 and C, Rs. 39. How many chairs of each type should be made in order to maximize profits at the show and sale? 7. Mr dain, the marketing manager of ABC Typewntter Company is trying to decide how he should allocate his salesmen to the ‘company's three primary markets. Marke | ig in the urban area ‘and the salesman can sell, on the average, 40 ypewriters a ‘waek, Salesmen in the other two markets, II and Ill can sel, on the avorage, 36 and 25 typewrtore por wook, rocpectvoly. For the coming week, three of the salesmen wil be on vacation, leaving only 12 men available tor duty. Also because of lack of ‘company care, a maximum of 5 salesmen can be allocated to ‘market area |. The soling expenses per week for salesmen in ‘each area are Rs 800 per week for area I, Rs_700 per week for area ll, and Fs £00 por week for area ‘Il. The budget for the next week is Rs 7.500. The proft margin per typewriter is Fs 150. Determine how many salesmen should be assigned to ‘each area in order to maximize profte? 8, Throe products - A, B and C - are produced in thrae machina centres X, Y and Z. All free products requie a part of their ‘ranufacturing operation at each of the machine centres. The time required for each operation on various preducts is indicated Jn the following ‘able. Only 100, 77 and 80 hours are availaba ‘ai machine cenites X, ¥ end Z respectively. The profit per unit from A, B and © is Re 12, Re 2 and Re 1, respectively 10. ‘Screw Typo Plant Capacity Manufacturing Cost (g/day x 100) (Asikg) 1 10 450 2 10 3.50 a © 270 What produstion shall this manufacturer schedule for greatest Profi, assuming that he can sell all that he manulactures? A blencer of whisky imports three grades A, B and C. He mixes them according 10 the recipes that specily the maximum or minimum percentages of grades A and C in each blend. These are shown ithe table below: Blond ‘Spoclicaton Price per Unit cs) Blue Dot Not ss ‘han 60% of A 680 Not more than 20% of G Highland Not more than 60% of © 870 Fring Not loss than 15% of A. Old Frenzy Not more than 50% of C 450 Fellowing are the supplioe of the three whiskies along with their cost, Whisky ‘Maximum Quantiy Cost por Avaliable por Dey Unit (Fe) A 2,006 7.00) 8 2,500 5.00 e 200 $00 Products Machine Centres Prot per oe Mit 8) a %7~C 2 8 2 3 4 3 ce 1 2 4 1 ‘Avaliable hous 1007780 (Determine suitable product mx so as to mativize te Poi. Comment on the queries (b) and (c) trom the solution table chained. Satisfy that fll avaiable hours of X and Y have beon uslzed and there is surplus hous of Z. Find out the surpus hours of Z Your aim is to utlize surgus capadty of Z. Can you say trom the table that the Inoducion of more units of Y Is required”? 9. Accortain manufacturer of screw fastenings found that there is a matket for packages of mixed screw sizes. His mar research deta indicated that two miures of three screw types (1, 2 and 3), propery priced, could be most acceptable to the public. The relevant data is: Co) © 12, 13 ‘Show how tabla the firet matrix in @ emplex computation of ‘8 production policy that will meximize profits. ‘An animal feed company must produce on # cily basis 200 kg (of @ mixture that consists ingredients x, and x, ingredient. x, costs Fs 3 por kg and x, costs AS 8 per ko. Not more than ‘Okg of x can be used ard at least 60 kg of x, must be used. Find out how much of each ingredient should be used if the company wants to minimize costs. ‘A det is to contain atleast 20 cunces of protel and 1 ounces of carpenyarate, Tere are twee toods A. B and C avaladie in to market, cosing Rs 2, Re 1 and Rs 3 per unt, respectively. Each Lunt of A contains 2 ounces of rotein and 4 ounces of carboty- date; each unit cf B contains 3 ounces of protein and 2 ounces of carbohydrate; and each unit oC contains 4 ounces of protein amt 2 ounces of cartohyerate. How mary units of each food should the det contain so thatthe cost per unt diet minimum? ‘A pergon requitee 10, 12 and 12 units of chomeale A, B and C, respectively for his garden. A typical tqud procuct contains 5, 2 and 1 unit of A, € and G, respectively per jar. On the other nand 2 Ypice dry product contains 1, 2 and 4 unis of A, 8 and C per Unt Ie liquid produet sels for Ris 3 per ar ard the dry product for Rs 2 per carton, how many of each should be purchased in ‘order to minimize the cost and meet the requirement?‘Operations Research: Theory and Applications 14. A scrap metal dealer has received an oder from a customer for a minimum of 2.000 kg of scrap metal. The customer requires that at least 1,000 kg of the shipment of metal be of high cualty copper that can be melted down and ueod to produce copper tubing. Furthermore, the customer will not ‘accept delivery cf the order if contains more than 175 kg metal tat he deems unfit for commercial use, Le. metal that contains ‘an excessive amount of impuriies and cannot be melted down {and relined proftably. ‘The dealer can purchase scrap metal trom two dierent suppl 1 In unlimted quanttios with following percentages (by weight) ‘of high quality copper and unft scrap. ‘Supplier A Supplier B Copper 25% 75% Unit scrap 5% 10% ‘The cost per kg of metal purchased from supplier A and B are Re 1 ‘and Fis 4, respecively. Determine the optimal quantities of metal to 'be purchased for the ceale from each ofthe two suppirs. 15, A marketing manager wishes (0 allocate his annual advertisng budget of Fs 2,00,000 to two media vehicles - A and 8. The unit cost of a message in mecia A is RS 1,000 and that of B js Rs 1,500. Media Ais a monthly magazine end not more than fone insertion Is desred in ore istue, whereas at least fve ‘massages should appear in media B. The expécted audionce for unit messages in media A Is 40,000 and that of media B is 155,000. Devolop an LP rrodol and colve it for maximizing the taal ellective audience. 18, An acvertisng agency wishes fo resch two types of audioncs: Customers with annual income of greater than Rs 15,000 (target audience A) and customers with anrual income of less than Rs 15,000 (target audience B). The total acvertsing budget is AS 200,000. Gne programme of TV advertsing costs RS 50,000; fone programme of radio adhertiong costs Re 20,000. For contract ressons, at least three programmes cught to be on TV and the number of radio programmes must be limited te five. A ‘survey indicates that a single TV programme reactes 450,000 customers in taget audience A and $0,000 in audence B; and that one radio programme reacnes 20,000 in target audence A ‘and 80,000 in target audience B. Determine the media mix to maximize the folal reach, 17. A transistor radio company manufactures four models A, B, © land D. Models A, B end C, have prof contnbutions of Fs 8, FS 15 and Rs 25 respectively and has model D a loss of Re 1. Each ‘ype of radio requires a certain amount of time for he ‘manufacturing ol components, for assembling and for packing, ‘A dozen urits of model A require one hour for manufacturing, {wo hours for assembling ard one hour tor packing. The corresponding figures for @ dozen urits of model 8 are 2, 1 and 2, and for @ dozen units ef © are 3, § and 1. A dozen unite of ‘model D however, only require 1 hour of packing. During the fonncoming week, the company wil be able 10 make evalleble 415 hours of manufacturing, 29 hours of assembling and 10 hours of packing time. Determine the optimal production sched- te for the company, 18. A transport compary is considering the purchase of new vehicles for providing trareportation between the Delhi Airport ‘and hotels in the city, There are three vehicles under consid- feration: Station wagons, minibuses and large buses. The pur- chase price would be Rs 1,45,000 for each station wagon, Rs 2.50,000 for each minibus and Rs 4,00,000 for each large bus. The beard of dirctors has authorized a maximum amount of RS {50,00,000 for these purchases. Because of the heavy a travel, the new vehicles would be utilized at maximum capacty, regardless of the type of vehicles purchased. The expected net ‘annual profi would be Rs 15.000 ‘or the station wagon, Rs ‘35,000 for the minibus and Fis 45,000 fr the large bus. The ‘company has hired 30 naw driers for tho now vohiclos. They ‘are qualified drivers for all three types of vehicles. The mainte- nance depatmert has the capecty 'o handle an actional 00 ‘tation wagons, A miribus is equivalent to 1.67 station wagons ‘and each large bus 1s equvalent to 2 station wagons in terms Of thelr use of the mainionarce department, Determine the number of each type of vehicle thet should be purchesed in order to maximize prof. 18, Omega Data Processing Company perorms thre types of actin ‘les: Payroll, accounts recetabies, and inventors. Tho proft and time requirement for keypunch, computation and oftice pining, for 8 standard jb, are shown inthe folowing table: ob Profi Time Requrement (Min) Standard —§ Jeb (Fs) — Keypunch Computation Printing Payrol 215 1,200 20 © 100 Ne Receivable 125 1,400 8 60 Inventory 225 800 8 20 ‘Omega guarantees overnight completion of the eb. Any jb scheduied during the day can be completed during the day oF Fight. Any job scheduled during the night, however, must be Completed during the right. The capacities for both day and night are shown in the following table Capacity (Min) __Keypunch Computation Print Day 4200 150 400 Nght 9,200 250 650 Determine the miture of siandard jobs that should be accepted during the cay and night 20. A furniure company can produce four types of chairs. Each chair is fist mage in the carpentry shop and then fumished, waxed and polished in the fishing shop. The man-hours required in each at Chair Type 1 2 8 4 ‘Carpentry shop 4 8 7 10 Fhiching shop 1 01003 40 Contribution per cha (Fs) 12 201840 ‘Tre total number of man-hours available per month in carpentry and finshing shops are 6000 and 4,000, respectiely ‘Acouming an abundant supply of raw material and an atundont ‘demand for finished products, determine the number of each 'ype of chais that should be produced for profit maximization, 21. A farmar hes a 100 acre farm. He can sell all the tomatoes, letuce or radishes that he produces. The price he can obtain ie Re 1 por kilogram for toratcos, Re 075 a hoad for lettuce ‘and Fis 2 per kilogram for radishes. The average yield per ace 's 2,000 kilograms of tomatoes, :,000 neads of etuce and 1,000 ‘klograms of radishes. The feriizer is avaliable at Re 0.50 por ‘klogram and the required amount per acre is 100 kilograms ‘each fr tomatoes and letuce and 50 klograme for radishes, ‘The labour required for sowing cultvating and hawesting per ‘ace is 5 man-days for tomatoes anc radishes and 8 man-days{or lettuce. total of 400 man-days of labour aro avaliable at Rs 20 per manday. Determine the crop mix so as to maximize the farmer's total prot. [Delhi Univ, MBA, 2008) 22. A metal products company produces waste cans, fing cabinets, fle boxes for conespondence, and lunch boxes. Its inputs aro sheet metal of two differant thickness, called A and 8, and manual labour. The inpu-output relationship forthe company are shown in the table given below: Linear Programming: The Simplex Method RRS 24, A company produces electte transformers for the eectecalin- dustry. The company has orders for transformer for the next sic ‘months. The cost of manufacturing a iransiormer is expected to somewhat vary ever the next low monthe due to expected changes in materials costs and in labor rates. The company can produce up to 50 units per monty in regular time ard up to an ‘ditional 20 units per month during overtime. The costs for boty regular and overtime production are stown in the table below: Waste Fling Correspondence Lunch Cans Cabinets Boxes Boxes Sheet meta A 8 0 2 3 Sheet meta B 0 10 ° 0 Manual labour 4 8 2 3 ‘The saies revenue fer unit ct weste cans, fling cabinets, ccortespondence coxes and lunch boes are Rs 20, Rs 400, Rs ‘90 and Rs 20, respecively. There are 225 units of sheet metal AA available nthe company’s inventory. 200 of sheet mata 8, ‘and a total of 190 units of manuel labour. What isthe company’s ‘optimal sales revenue? [Dothi Univ, MBA, 2000, 2002, 23. A bank is planning its operations for the next year. The bank makes fve types of loans. These are listed below, together with the anrual return (in per cent) ‘Type of Lean Annual Retum (per cert) Education 15 Furniture 2 ‘Autorrabile ° Home Corstructon 10 Home repsir 7 Legal requirements and bank policy place the following mis ‘on the amounts of the various types of loans Education loans cannot exceed 10 per cent of the total ‘arount of loans. The amount cf education and furniture toans together cannot exceed 20 per cent othe total amount of loans. Home constriction loan must be atleast 40 per cent of the to amount of tome loane and at least 20 per cant of the tol ‘amount of leans. Homa construction loan must not exceed 25 ‘per cent of the total amount of loans. ‘The bank wishes to maximize its revenue from loan interest, ‘subject to the abeve restrictons. The bank can lend a maximum fof Fis 75 crore, Formulate and solve this problem as @ linear programming problem, Month Van. Feo. Mar. Apnl way June Orders (units) 58 36 34 69 72 43 Cost per unit at 180 170 170 185 190 190 regular time (in ‘000 Fs) Cost per unit at 200 190 190 210 220 220 ‘overtime (in 000 As) “The cost of caring an unsold transformer in stock is Rs 25 000 er month. As on January 1 the company has 15 trantormors in Stock on and it wishes to have no less than 5 in stock on June 30. Formulate a near programing problem to determine the cptimal production schedule. ‘A company rrined diamonds in the locations in the country. Tho three mines difered in torms of their capactios, numer, weight of stones mined, and costs, These are shown in the table below: Due to marketing considerations, a monthy production of exactly 148,000 stones was required. A similar requirement called for at least 1,30,000 carats (The average stone size was a least 130/148 88 carats). The capacity of each mine Is measured in cuble ‘eter. The mning costs are not ircluded from the treatment costs and assume to be same at each mine. The problem forthe company was to meet the markating requirements atthe leest cost, 28. Capacity Treatment ‘Sone Court Crude (of Costs (Carats (Number of Mine eath (Fs. perM®) per M3) stone per processed) Mw) Plant 1 83,000 «0.60 «0360.88, Plant2 310000 0.36 = 0220028, Plant 3 1,90.000 0263021 0.50 Fornulale a near programming model to daternine how much shculd be mined at each location HINTS AND ANSWERS 1. Let x, and x, = units of models A and B to be manufactured, respectively Max Z = 400, + 400, subject to dx + ay $1,600 Say + ay < 1,200 95/2 + 3x52 < 1,600 and ayy 20 Ams.) = 355.5, 4) = 0 and Max Z = 1,42,222.2. 2 Let xj,% and x, = units of gypes A, B and C bolt to be manufactured, respectively. Max Z = 34, + Se, + 4x, subject 0 2x, + 3s) SB Qt Sy S10 Bayt Bey 4a SS and nkyay2 0 ‘Ans. x)= 89/41, = 50/41 x, = 68/41 and Max Z = 775/41, 3. Let xy, x, and xy = quantity of products A, B and C to be produced, respectively Max Z =x, + 4x; + x5 subject 10 ay tan $22 x, +e +45 < 14 By t2q 0 SM and Spt a2 0 Ans. xj ~ 0.x) 7 7.33 > 0 and Max Z ~ Rs 28,000,4 Operations Research: Theory and Applications Let x,,x) and x, = number of units of products 1, 2 and 3 10 be produced per week, respectively. Max 2 = 207, + 6r, + 8x, subject 10 Bx, + 2s) + 3x, < 250 de tdy 2150 2x, + 4550 and fy ty 20 Ams.) 0,5 > Let xj.) and = acreage of com, wheat and soyabean, respectively. Max Z = 305, + 40x, + 20r, 30, x) = 90 and Mex Z= 700. subject to 10x, + 12e, + Tey < 10,000 75, + W0ry + Bry < 8,000 x) ty + $1,000 and pips 20 Ams. xj © 250, x, = 625, xy = O and Max Z = Rs 12,500. number of units of chair of styles A, B and , respectively Max Z = 402, + 38x, + 30x, Let xy, ¥, and x, subject to 2e + 3x + 2ey 120 45, + 3p + xy S160 Bey t Bey + Any = 100 a tats 40 ant sp ky ay 20 Ams. x)= 20,2) © Let x, x and x 20, x, = 0 and Max Z = Rs 1,500, salesman assigned to area, 1, 2 and 3, respectively, 40 150x, + 36% 150r, + 25% 150%, (800s, + 7003, + 500x,) subject © @ x+y ty S12 Gi)x $5 i) 800, + 700%, + S008, = 7,500 and Xp tps 20 Lot x and y = number of kp of ingredients x, x3, respectively Min (tou cost) Z = 3x + 8y 200; (ii) < €0; Max. Z subject to (x + and Gi) » = 60 nye Ans, x= 80, y = 120 and Min Z = Rs 1,200. Let xy, and x, = number of units of food A, B and C, respectively whieh a diet must contain, Min total cost ) Z = 2x, +3 + xy subject t0 2x, + 2s) + 4x, 2 20 4x, + 2 +22 15 and yay BO Let x, and x, = number of units of liquid and dry product produced, respectively Min (ita east) Z = 35) 7 2p subject 0 () Sr #45210; Gi) 2, + 2n52 125 (i) 4+ 452 12 andr, 1, 20 Ans. Let x and x = number of serap (in ka) purchased from supplies A and B, respectively Min (tout east) Z =x, + 4, ay = Lexy = Sand Min Z = 13, 1s. 20, subject to 2.254, + 0.75x, 2 1,000 0.05e, + 0.10, 2 175 xy + a 2 2,000 and ye 20 ‘Ams. x; = 2,500, 4, = 500 and Min Z = 4,500, Letx, and x, = number of insertions messages for media A and B, respectively. Min (total effective audience) Z = 40,000x, + 55,000r, subject to (@) 1,000x, + 1500s, < 2,00,000; @ 51%; Gi) yes xp 20 ‘Ams. x)= 12, x, = 16/3 and Max Z=77,3333.33. Let x, and x, = number of radio and television programmes, respectively. Max (total audience) Z = (4,50,000 + 50,000) x, + (20,000 + 80,000)x, subject to (@) $0,000:, + 20,000x, < 2,00,000; and Gi x53; fii) 25 and xp 20 Ans. x, = 4, x, = O and MaxZ = 4 x 5,00,000. 1 Let xy, X, ¥; and x, = unit of models A, B,C and D to be produce, respectively Max (total income) Z ~ 8, + 15x + 2545 — xy subject 10 (@) xy+ 2x + 3x," 15; (il) 2x, +x + Sx, = 20 (ii) x, + 2 + +x, = 10 and Rp ty 20 Ans x, = 52, 4 = 92,15 = 52% Let xy. and xy = number of station wazons, minibuses and large buses to be purchased, respectively. Max Z= 15,000, + 35,000r, + 45,000, subject 0 +e ty $30 1,45,000x, + 2,50,000:, + 4,00,000, < 50,00,000 + 16x + 05x, $80 and aye ty 0 Let xy represents ith job and jh activity Max Z=275 (ry, +2) + 125¢4, +3) +225 (89) +252) subject to 1200 (4) + 49) + 1,400 49) + 239) + £00 (35, + x32) < 13.400 20 (54 + Hig) + 15 bey) + Xap) + 35 GH + tq) S400 100 (x, +842) + 60 (py 4 X99) + 80 (x5, + Hq) $1,050 1200512 + 1400%99 + 800r3p $ 9.200 Deja + 15aq + 35yq $250 100r}» + 60x, + 8x3, $ 650 and xj 2 0 forall i,j hairtypes 1, 2,3 and 4 to be produced, respectively. 0 and Max Z = 126, Leta, 3, sy and 5 Mar Z = 12x, + 20x, + 183 + 40x, subject to 4x, + 9x + Ty + 10ry < 6,000 xy + at Bx + 408, $ 4,000 and Xp tay ty 42 0‘Ams. x, = 4,000/3, x, = xy = 0,.x4 = 200/3 and Max 2 Rs 56,0008, 21. Let ay, and.x, = number of units of tomatoes, letuce and radishes 10 be produced, respectively, Max (profit) Z = Selling price — Fertilizer cost - Latour cost = (1 «2,000 - 050% 100 ~ 20% 5), ~ (0.73 * 3,000 ~ 6.50 100 ~ 20 * 6)x, + 1,000 ~ 0.50% 50 = 20% 5), 830x, + 2,080 x, + 1.875 25 subject to (i) x +1, +15 $100; (i) Sx, + 6 + Sry $400 and yi H2 0 WB. Leta ety funds allocated to edeation,fori- ture, automobile, home construction anc home repair, respectively. Max: 7 = 0.18x, +0.12s, + 0.09, + 0.10, + 0.075, subject to (i) x; +3; +x +x, +x, $1.5 (Funds available) Gi) x, 0.108, +, +x) yt x9) or 0.98, ~ O.lxy ~ O.lay ~ O.t3y — Dulas <0 (i) x, +x $020 tay + tH te) or O.8x, + 0.8x, — 02x, ~ 0.21, ~ 02x, 5 0 (iv) %5 2040, +5) or - 04x, + 06252 0 (9) 5202, 4 2) Hy tony +d or = 0.2, - 02x, - 0.2x, ~ 02x, + 0.85 > 0 and 20 for allj Ans. x, = 0.15; x)= 015; x, = 0.525; x, = 0.375; x, = 030; Max Z = 0.146 24, Let xj) «+s = mmber of tansformes produced in eeu | lar time each month My By + Yq = mamber of transformers produced on overtimes each month ‘number of transformers in stock at the end of each month Min Z = 18.x,+ 171, + 17x, + 85x, + Lrg + 19s, + 20y, + 199, + 1995 his by 2. Linear Progamming: The Simplex Method BEE + Diy, + 22y, + 22y_ + 051 + 0.54 +05), + 0.5, + 0.51, + 05/, subject to x, $50 for x S| ‘Capacity straints, 2} Caco 3, $20 for 1, = 15 +x, + y ~ $8 (Inventory for January) fy =f px; 7 Yj Orders for f= 2,3, 6 1, 2 8 (Final inventory) Ans. Month Regular Overtime Inventory Production Prodiction January 8 0 0 February 30 ° “4 March 50 u 41 April 50 0 2 May 50 6 0 June 48 o ‘Total cost = Rs 2,75,550. ‘An alternative basic solution exists that involves regular pro- ‘duction of 50 in January, and only 4 produced in overtime in ‘March. Ending inventories in January = 7 and February = 21 Let x}, 2) and x; = eubic metic of earth processed at plants 1. 2 and 3, respectively Min Z = 0.60x, + 0.36x, + 0.50%, subject to 058%, + 026x, + 021 036x, + 022x, + 0.263155 1,30.000 (Carat requirement) x; £ $3,000; x, < 3,10,000; 48,000 (Stone count requirement) x) $ 1,90,000 (Capecity requirement) Ans. x; = 61,700: x, = 3,10,000; x, = 1,50,500 and Min Z ~ Rs 2,23,880. 4.5 SOME COMPLICATIONS AND THEIR RESOLUTION We have discussed, through examples, how one can use the simplex method for solving both maximization ns that may arise in applying the simplex and minimization problems. In this ection some of the complicat ‘method and the resolution of these problems are discussed. 4.54 Unrestricted Variables Usually in an LP problem, it assumed that all the variables x,(j~ 1, 2, ‘An unrestricted variable in an LP model can have 1) should have non-negative values. In many practical situetions, however, one or more of the variables, can have either positive, negative or zero value. Variables that can assume positive, negative or zero value are called unrestricted varicbles. Since the use of the simplex method requires that all the decision variables must have non- negative value at each iteration, therefore, in order to convert an LP problem involving unrestricied variables into an equivalent problem having only restricted variables, we have to express cach of unrestricted variables as the difference of two non-negative variables. Let variable x, be unrestricted in sign, We define two new variables say x and x” such that nanan; xx" 20 If xf & 37 of then x,2 0 and if x7 SM—2 aes 1M } Initial Solution oy Mt aMen aes 0 In Table 443, c, ~2 is the largest positive value, we need to enter variable x, into the basis and remove variable x, from the basis. For this, apply the following row operations: R, (new) R, (old) + 5 (ley element) R, (new) > R, (old) ~ 4, (new). ‘The improved solution so obiained is given in Table 4.44.125 | 25 25 1 ws 0 |B xS 6 1" so ts 1 |ExS ara = [-ms+as sas 2 amat2s — aw sa | sss ess 6s 35 ee In Table 4.44, as c, — 2, in x’, column is still positive, introduce variable x’, into the basis and remove variable 4, from the basis. For this apply the following row operations: Ry (new) > R, (old) x (5/7) (Key element); Ry (new) > R, (old) ~ (2/5) Ra (new). ‘The improved solution so obtained is given in Table 4.45. Further, in order to improve the solution given in Table 4.45, we noed to introduce variable x; into the basis and remove variable x, from the basis. For this we must apply the following row operations: R, (new) > R, (old ) * (7/3) (key element), Ry (new) > R, (old) + (4/7) R, (new) Table 4.45 Improved Solution ‘The improved solution so obtained is given in Teble 4.46. > 3-3 z 1 1 7 143 ° is 3 a 3 1 43 © 2-08 y 3 198 1 Table 4.46 g-y | 0 os Susie S -0) Opimal Solution In Table 4.46, all ¢ ~ =) < 0, an optimal solution has been arrived at withthe values of the variables, as: x) = 113 or x,= 4) ~ #7 = 118-0 113;x, = 143 and Max Z= 47/3. 4.5.2, Tie for Entering Basic Variable (Key Column) A situation may arise at any iteration when two or more columns may have exacily the same ¢) ~ , value (positive or negative depending upon the type of LP problem). In order to break this tie, the selection for ‘key column (catering variable) can bz made arbitrary. However, the number of iterations required to arrive at the optimal solution can be minimized by adopting the following rules: @ If there is a tie between two decision variables, then the selection can be made arbitrarily. CopyrighteTavle 4.47 Inia! Solution Operations Research: Theory and Applications (i) If there is a tic between a decision variable and a slack (or surplus) variable, then select the decision variable to enter into basis first. i) If there is a tie between two slack (or surplus) variables, then the selection can be made arbitrarily. 45.3 Tie for Leaving Basic Variable (Key Row) —Degeneracy While solving an LP problem a situation may arise in which there is a tie between two or more basic variables for leaving the basis, i.e, the minimum ratio to identify the basic variable to leave the basis is not ‘unique or in which values of one or more basic variables inthe ‘solution values’ column (xg) become equal to zero. This causes the problem of degeneracy. However, ifthe minimum ratio is zero, then the iterations ‘of simplex method are repeated (cycle) indefinitely without arriving atthe optimal solution. Such a situation, of course, is very rare in practical problems. ‘The problem of degeneracy arises due to redundant constraint, ie. in the LP problem one or more of the constraints makes ancther unnecessary. For example, constraints such a5 x, <5, $5 and x, +m $5 in the LP problem make constraint x, + x, < 5 unnecessary (redundant). Degeneracy may occur at any iteration of the simplex method. In most of the cases when there is a tic in the minimum ratios, the selection is made arbitrarily. However, the number of iterations required to arrive at the optimal solution can be minimized by adopting the following rules. ( Divide the coefficients of slack variables inthe simplex table where degeneracy is detected by the ‘corresponding positive numbers of the key column in the row, starting fiom left 10 right. (Gi) The row that contains the smallest ratio comparing from left right columnwise becomes the key rv. Remark When there is a tie between a slack and artificial variable to leave the basis, the preference should be given to the artificial variable for leaving the basis. There is no nced to apply the procedure for resolving degeneracy under such cases. Example 4.13 Solve the following LP problem Maximize 23x, + 9x, subject to the constraints @y+4yS8 yt 2ys4 and 320 Solution Adding slack variables s, and s, to the constraints, the problem can be expressed as Maximize Z = 3x, + 9x, + 0s; + Os subject to the constraints Gx + 4, +5, =8, Gi x, + y+ 5, and xy.Xy sy $20 ‘The inital basic feasible solution is given in Table 4.47. As shown in Table 4.47, c) —2, ~9 i the largest positive value, therefore variable 2, is selected to be entered into the basis. However, both variables s, and ‘5 are eligible to leave the basis as the minimum ratio is same, ie. 2, so there is @ tie among the ratio in rows 5; and 3. This is an indication of the existence of degeneracy. To obtain the unique key row and for resolving degeneracy apply the following procedure: (@ Write the coefficients of the slack variables as shown in Table 447.Linear Programming: The Simplex Method 5 4 1 0 a 2 0 1 i) By dividing the coefficients by the comesponding element of the key column, we obiain the following ratios: 3 4 va=14 OM = 2 02-0 12-172 i) Comparing the ratios of Step (ii) from left to right columnwise, the minimum ratio occurs for the second row. Therefore, the variable s, is selected to leave the basis. The new solution is shown in Table 48. g3 3 9 0 0 0 4 Erste 0 ees 9 ie Pee iets cu o z= 18 28 ° 9R Table 4.48 g-5 |-2 0 0-92 Opiimal Soustion In Table 448, all c)~z, $0. Therefore, an optimal solution has been arived at. The optimal basic feasible solution is: x, = 0, x, = 2 and Max Z= 18. 4.6 TYPES OF LINEAR PROGRAMMING SOLUTIONS In this section, we shall discuss three different types of solutions in terms of the termi method of the simplex 4.6.1 Alternative (Multiple) Optimal Solutions ‘The alternative optimal solution can be obtained by considering the c)— z)row of the simplex table. We know that en optimal solution to a maximization problemis reached if all c,~z, £0. What will happen ifc,—z)=0 for some non-basic veriable columns in the optimal simplex table? Each entry in the ¢, ~ z) row indicates the contribution per uit ofa particular variable inthe objective function value if tis entered into the basis. Thus, ‘fa non-basic variable corresponding to which c,— = Ois entered into the bass, a new solution will bearrived at but in this case the value of the objective function will not chenge. Example 4.14 Solve the following LP problem Maximize Z=6r, + 4x, subject to the constraints (2x, +3y830, Gi) 3x, +2524, Gi) x +23 and ty 20. Solution By adding slack variables 5}, 95, surplus variable sy and artificial variable 4 in the constraint set, the LP problem becomes: Maximize Z = 6r,+ 4x, + Os, + Os + Os, ~ MA, subject 10 the constraints @ 2, +3x,+5,=30, GH 3x,72x+5=24 x, +x,~8)44,=3 and Xp Xp Sy. By Sy Ay20j a ‘Operations Research: Theory and Applications Table 4.49 Optimal Solution Table 4.50 Alternative Solution ‘The optimal solution for this LP problem is presented in Table 4.49. G7 64 o o o Oe aR cee ee 2B 0 13 0 4 0 2 0 oe a <0 Saree ie ‘The optimal solution shown in Table 449 is: x, = 8, x, = 0 and Max Z = 48. In Table 4.49, c, — z, = 0 coresponds to a non-basic variable, x, (i.e. x = 0). Thus, an alternative ‘optimal solution can also be obtained by entering variable x, into the basis and removing. s, from the basis. The new solution is shown in Table 4.50. 4 a5 0 1 0 asa ss | 0 0 - 1 6 i 125 1 ° Gs x ° Z-48 5 6 4 ° 2 0 Gy 6 0 o aig, o ‘The optimal solution shown in Table 4.50 is: x, = 12/5, x) = 42/5 and Max Z= 48, Observe that in Table 4.50, c — 2, = 0 and variable s, is not in the basis. This again indicates that aan alternative optimal solution exists. The infinite number of solutions that can be obtained for this LP problem are as follows: 8 x = 8% + (125) 2) a 0 re x) = 0% oo ea % 4 ° n=l Od- 5 5 395 tos semhey For each arbitrary value of 2, the value of objective function will remain same. 4.6.2 Unbounded Solution Unboundedness describes an LP problem that does not havea finite solution. For example, in a maximization LP problem, if c,— z)> 0 (c, - z; <0 for a minimization case) for a column variable not in the basis, and all entries in this column are negative, then for determining key row, we have to calculate the minimum ratio corresponding to cach basic variable having negative or zero valu in the denominator. Negative value in denominator cannot be considered, as it would indicate the entry of a non-basic variable in the basis with a negative value (an infeasible solution will occur). A zero value in the denominator would result in a ratio having @ value of + =, This implies thatthe entering variable could be increased infinitely with any of the current basic variables being removed from the basis. In general, an unbounded solution occurs due to ‘wrong formulation of the problem within the constraint set, and thus needs reformulation.image not availableOperations Research: Theory and Applications Example 4.16 Solve the following LP problem Maximize Z= 6x, + 4x, subject to the constraints O x+y S5, and Xp Hz 0, Gi) 28 Solution By adding slack, surplus and artificial varisbles, the LP problem becomes Maximize Z = 6x,+ 4%) + Os, + 05 — MA ‘subject to the consiraints @ xy ty+H=5, and Xp Ay Sy Say Ay 20 h (i) y- 54428 ‘The initial solution to this LP problem is shown in Table 4.53. . 1 0 0 6 0 af Table 4.53 6-5 0-24 0 =m 0 } Initial Solution 3 eration 1; Variable x3 enters the basis and x; leaves the basis. The new solution is shown in Table 4.58 97 6 ‘ 0 0 - 1 Table 4.54 -M 4, 3 Ei ° sr Ly i Optimal but : Infeasible 2-20-3M y 40M + atm ow Solution one pee 0 eae tone Since all ¢j— 5) ‘$0, the solution shown in Table 4.54 is optimal, But this solution is not feasible for the given problem since it has x, = 0 and x) ~ 5 (recall that in the second constraint x, > 8). The fact that arificial variable A, ‘constraint (x, 2 8) by 3 units is in the solution also indicates the fact that the final solution violates the second CONCEPTUAL QUESTIONS Define slack and surplus variaties in a inear programming problem. Explain the various steps of the simplex method Involved in the ‘computation ofan optimum ecto to a inearprogiammang prebiom. a) Give outines of the simplex method in near programming. {0} What is simplex? Describe the simplex method of solving linear programming problen. (c)_ What do you understand by the term two-phase method of sohing (4) Outine the simplex method in near programming. Why is Wt called 202 What do you mean by an optimal basic feasible solution to a linear progremming problem? ineae programming problem? (e) Exolain the purpose and procedure of the simplex method. 5. Given a general tnear programming problem, explain how you ‘would test whether a basi feasible solution is an optimal solution for not. How would you proceed to change the basic feasible ‘solution in casa it is not optimal? Explain the meaning of basic feasible solution and degenerate Solution in a tinea programming prottem. Explain what is meant by the tems degeneracy and cycing in linear programming? How can these protiems be resohed? Explain the tom articial variables ands use in near programing, ‘What are artical variables? ‘Why do we need them? Deserve the ‘wopphase method of solving an LP protiem with artificial varables. What is the significance of «| — z numbers in the simplex lable? Interpret their economic significance in terms of marginal worth,image not availableOperations Research: Theory and Applications 4. Soe the following LP protiems and remove the complication if any) () Max Z = 2x, + 31+ 10%, (i) Max 2 = 5x,— 2% + 3x (i) Max 2 = 5r, + 3x eubpct © x) aya 2 subject to 2x + 24- m2 2 evbect to ne nse et yet By- 4 8 Bx, + 2 < 10 and My Hy 20 toy ss 3x, + 84 < 12 and Me EO and My 2 0 (iv) Max 2 = 22%, + 90%, + 25x (v) Mac Z = 8% subject to 2x, + 2x 4 < 100 subject to x= 420 2+ Ht Hz 100 2k, + 9% 5-6 H+ Bay + 2xq < 100 and 2%, % unrestricted and HH mez 0 5. Solve the following LP problems to show that these have altemative optimal solutions. () Max Z = 6x, + 9% (il) Min 2= 2x4 + Bx (8) Max 22 4 + 2H + 3-H subject to Bytes 8 subject to 5x, +4 2 10 stbiec! tox, 4 2x + By = 15 Bx, + 3x < 18 2x + 2m 2 14 2k + % + Bry 2 20 See x + 4x, 2 12 tht ig ty 210 and Xe Mee 0 and Wye 0 and Xe Kaha 20 6. SoWe the following LP problems to show that these have an unbounded solution () Max Z = -2%, + 3% (i) Max Z = 8x; + 6x (i) Max 2= 107%, + x + 215 subject to xs5 subject to 8x, + 4x, 2 12 subject to 14x, + x ~ 6+ 3x4 = 7 2x, 8 <6 net S4 Tox, + 05% ~ 8% <5 and my M20 and mq 20 yy <0 and Hy %e20 (iv) Max 2 = 6x, ~ 2x subject to 2 - eS 2 med and No He 20 7. Solve the follwing LP problems to show that these have no feasible solution () Max 2 = 2x, + 9% (i) Max 2 = 4,4 95 dig Sey (i) Max 2 = 4) + 9x subject nome subject to 4x, + 6%- 5x + 4% 2-20 subject to ome 7 Krys 34) = 2k + 4% +x, 5 10 4-4 5-3 ns2 3x, — 2% + 4y + S10 and Xe me 0 and Xy 20 Bx, ~ Bx ~ By, + 2x, $20 and By Hy Hy Hy 20 (iv) Max 2 = 9x, + 2x) subject to yt s 2 By + 4x > 12 and Hy EO [Meenit, MSc (Maths), 1998) HINTS AND ANSWERS. 1 @ x)= 3.xp= 1 and Max Z= 11 (iil) xy = 1,2 = Ly = V2 and Max Z = 13/2 Gi) x)= 2.x, = 0 and Max Z = 10 (Gx) Divide the fst equation by 3 (coefficient of x,) Gi) x, = 0.x, = 100, xy = 250 and Max Z= 1,380 6) x, = 0.x = 0 and Max Z = 290 (is) Max 2" = ~ x, + 31, ~ 2x, where 2* (x) x)= 0, = 6, x, = 4 and Max Z= 6 x)= 4yx)=5,4,= 0 and 2" =I @ x, ~ 1x, = 0, and Nax Z= 6 (9) a7 $07, 34 ~ Oyxy~ 55/7 and Max Z~ 69577 (i) All ¢j—z) SO but Z =~ 5/4 (<0) and artifical variable Ww) = 00% and Max Z ‘Ay ~ 5/4 appears in the basis with positive value. Thus (id) = 03 1 and Max Z the given LP problem has no feasible soltion.(iil) x, = 5/4, x, = 0, x, = 0 and Min Z Gv) x)= 2,4) = 0 and Mex 2= 6 () x)= 0, = 5 and Max Z = 40 (i) Allg ~z,2 0, but Z= — 4(< 0) and artificial variable 4 ~4 appears in the tasis with a positive value, Thus the given LP problem has no fessible solution ( x= 3,4, =O and Max 25 0 (i) x, = 3/5, x, = 6/5 and Min Z = 12/5 (Gil) All ¢, ~ z= O anificial variable A, = 0 appears in the ‘basis’ wit zero value. Thus an optimal solution to the given LP problem exiss. (Ge), Introduce artificial variable only inthe third constraint 1 =0,x;=0,x;=0,.x5= 0 and Max Z= 4, © Sand Min 2=—11 (wi) and Min 220 “ O and Max Z = 3 (Gi). Degeneracy oceurs at the intial stage. One ofthe variable lige to leave the basis is artificial variable, therefore, there is no need of resolving degeneracy. Remove the antfcil varicble from the basis. 238, x, = 5.x = 0 and Max Z = 85/3 = 2,5, =0 and Mex Z= 10 Gi x, Linear Programming: The Simplex Method fiv) x, = 1008, x, = 50/3, x, = 50/3 and Mar Z = 1,650 (0) xf = 615 oF x, = ~ 65 oF xf = 615 or xy =~ 6/5 and Max Z =~ 44s O@x and Max Z = 24 (©) x, = 52, 2, =3, and Max Z= 24 ( (@) x, = 3246, x, = 106 and Min Z = 24 () 3, = 12. x = 0 and Min Z = 24 (At the curent solution: x, = 5, xy 9 and Max Z = 15, it may be observed that € — 2) = 3/2 but al elements inthe second eolur are negative. Solution is unbounded (i) At the curent solution: x, = 4, s;= 4 and Max Z= 24, it may be observed that ¢ ~ z, = 15 but all elements in the second column are negative. Solution is unbounded. (i) At second best solution, c; —z, = 113/3 but al elements {nthe third column are negave. Soluon is unbounded; x, 0," 13, 4; = $ and Max Z = 0. (iv) Optimal souton: x, = 4x, = 6 and Max Z= 12. Since {nthe inital simplex table ll the elements are negative {nthe second column, the feasible solution is unbounded but the optimal solution is bounded, (j= 2,270, 4,~2andMMax 2-9-4; Infeasible solution liv) x) "0,2," 2,4, =2 and Max Z=4—4M4, Infeasible solution. CHAPTER SUMMARY ‘The simplex method is an interactive provedure for reaching the optimal solution to any LP problem. It consists ofa series of rales that, in effect, algebraically examine comer (extreme) points of the solution space in a systematic wey. Each step moves towards the optimal solution by increasing profit or decreasing cost, while maintaining feasibility. The simplex method consists of five steps: (i) idemttying the pivot column, (i) identifying the pivot row and key element, (it) replacing the pivot row, (iv) computing new values for each remaining row; and (v) computing the and c) ~ 5, values and examining for optimality. Fach step of this iterative procedure is displayed and explained in a simplex table both for maximization and minimization LP problems. CHAPTER CONCEPTS QUIZ True or False os 2 “The major difference between sick and artificial vatiables is hat an aificial can never be zero. Wan optimal solution is degenerate, then there are alternate catimal solutions of the LP problem ‘An infeasible solution is characterized as one where one ‘constraint is viclated It the objecive function costicient in the «row above an artificial variabi is - M, then the problem is a minimization problem, In the simplex method, the intial solution contains only slack variable in the product mix Hf there is @ tie between decision variable and a slack (or surplus) vaviable, then solact the decision varable to enter into the first bass. ‘An oplimal solution to the maximized LP problem is reached if all q/- 2,20. Varies which can assume negative, positive or zaro value called unrestricted variabies, 10, Al the rules and procedures of the simplex method are identical ‘whether solving a maximization of minimization LP problem. Atitciel variables are added to a linear programming problem to aid in the finding an optimal selution. Fillin the Blanks 1. 12, 18, In the simplex method, the column contains the Variables which ro curranty inthe ealutien; tho values ef these variables can be read from the column, ‘A ——— variable represents amounts by which solution values exceed @ resource. The simplex method examines the points in a ‘ystomatic manror, repeating tho camo cet of elope of the algorithm until an solution is reached, ‘occurs when there i no soluton that satisfies all of the constraints in the linear programming problem.Operations Research: Theory and Applications 28. 28. at. ® 15, The vabe for the replacing row must te befow computing the vas for the owe 16. Eniies in the ¢)~ 2, rows are known as costs. 17. Optimalty is indicates for a rraximizaen problem wen all lomenis in tho ¢) =z, rows are vile for a nization probit al elements must be 18. In Big motoa, batic feasible soliton is ottaned by assigning velue to the original value. 18, In the two-phase method, an variable is never coradered fr re-entry ints tho basi. 20 cccurs whan there is no fis sottion in he L? preblem Muttiple Choice 21, The role of ail variables in the simplex method is (6) 10a in finding an ini! soution (©) to fd optimal dl pies inthe final simplex table {6) 10 san pnases of simplex meted (d) all of the above: 22, Fot a maximization problem, the objective function ccfcent for (b) — M (0) none ofthe above 23, 11 & negatve valve appears Inthe sluten vates (Xe) column of the simplex table, then (@) the soliton is optimal (©) tho soliton is infeasible (6) the solution is unbounded (6) al eh the above 28, At every eration of the simplex methos, Or minimzaton problem variabl in the curren’ bass i replaced wth another arable that has (a) m postive ~ 3, vale (©) a negate & ~ % valve 4-479 {@) none of the above 25, In the optimal sinpex table, 2; = 0 value indicates (@) unbounded ston (8) cif {6) atematve soution (6) infeasve sctuton 26, Fot maninizaton LP mode he simplex method Is terrinated when all values | (@g-%50 (@ q-z20 (4-70 (450 | 2F aT 4F 58 F &T cant 12, Susus 13. Extomo; optimal 17, non postive, non negative 2. (#22. @) 28.00) 2.) 25. fo) 26. Bt. (b) 32. (a) 33.(a) 4. 38. (c) ‘A variatle which does not appear in the basic variable (B) ‘column of simplex table lo (a) never equal to zero (b) always equal to zer0 (c) called a basic variable (d) none of the above It for a aiven soltion a slack variable is equal to 22r0 then (a) the solution is optimal (b) the soluton is infeasible (c) the entre ameunt of resource with the constraint in which the slack variable appears has been consumed. (4) all of the abore Ht an optimal soluton is degenerate, then (a) there are altemative optimal soluions (b) the solution is infeasible (c) the solution is of no use to the decision-maker (8) none of the above To formulate a problem for solution by the simplex method, we must add artical vaniebie to (a) only equalty constants (0) only ‘greater than’ constraints (c) both (a) and (b) (a) none of me above Hany value in xg ~Column of final simplex table is negative, then the solufon ie (@) unbounded (| infeasible (6) optimal (4) none of the above If allay values in the incoming variable column of the simplex table afe negative, then (a) soluton is unbounded (b) there are mutiple solutons (©) there exlst no solulon (c) the solutbn is degenerate Hf an artical variable is present in the ‘basic variable’ column ‘of optimal simplex table, then the solution is {a) infeasible (©) unbounded (6) degenerate (@) none of the above ‘The por unit improvement in the solution of the minimizaon LLP problem is indeated by the regatve value of {@) ¢j~2; in the decision-varable column (©) 6-2; in the slack variable column (©) 6-2 In be surplus vatable column (@) none of me above To convert 2 inequality constraints into equality constrains, we ‘must (@) add a surplus variable (®) subtact an artificial varable (2) eubstract a ouplus variable and an artificial variable (@) add a suplus variable and substact an artificial variable ZT 8T 8F OT 14 Infeasteitiy 15. Computed: remaining 18, Intal; 2670 19. Atificial 20. Undoundedn Br. (b) 28. (e) 29. (4) 80 (e)Linear Programming: The Simplex Method BEA CASE STUDY : State Electricity Board* The state clecticity board is planning 10 construct a new plant forthe next 10 years, It is possible to construct four types of electric power faciltes-team plants using coal for energy, hydroelectric plants with no reservoir, hydro- electic plants with small reservoirs (enough water storage capacity to meet daily fluctuations) and bydroeleetc plants with large reservoirs (with enough water storage to meet seasonal fuctustions in power demands and waters low. Consumption of electricity is based on three characteristics: The first isthe total annual usage ~ the requirement inthe area is estimated to be 4,000 billion kilowatt-hours by the 10th year. The second characteristic isthe peak usage of power — usually on a hot summer day at about 2 PM. Any plan should provide enough peaking capacity to meet ‘projected peak need of 3,000 million kilowatts in the 10th year. The third characteristic is guaranceed power output measured a: the averaged daylight output in midwinter when the consumption is high and water levels for hydroelectric power are low. The 10-year requirement is for 2,000 million kilowatts of guaranteed power ‘The various possible power plants vary in terms of how they can satisfy characteristics. Por example, hydro clectic plants with reservoirs are able to provide substantial peaking capacity, whereas steam plants and hydroelectric plants with no reservoirs are poor in this respect. ‘The characteristics of the various types of plants are shown in the ble below. Each is measured in terms of ‘2 unit of capacity, The unit of capacity is defined to be the capacity to produce | billion kilowatt-hours per year. Note that the types of plants vary substantially in their investment costs. The annual operating costs of the various {types of plants also vary considerably. For example, tbe cost of coal makes the annual costs of the steam plants quite high, whereas the annual cost of operating the hydroelectric plants are relatively les. The final column in the table shows the discounted total costs, including both the investments costs and the discounted annual operating costs. Case 4. CCharacteristies of Electic Plants per unit (1 biion kilowatt-hours) of Annual Output steam os 020 1200 2600 Hydroelectric: no reservoir 0.10 0.10 1600 1680 Hydroelectric: smal reservoir 0.10 040 2400 2550 Hydroeteoric: large reserve ono 090 000 ‘4400 ‘Questions for Discussio (a) Help the company in developing a 10-year plan that would deal the capacity of each type of plantthat it should build, (b) Develop an LP mosel and solve it to minimize the total discounted cost, However, there is restriction that no more than Rs 14,000 milion can be used for invesment in plants over the 10 yeurs, Case 4.2: Nationwide Air lines* Nationwide Ailines, faced with a sharply escalating cost of jet fuel, is interested in optimizing its purchase of jet fuel at its various locations around the country. Typically, there is some choice concerning the amount of fuel that can be placed on board any aircraf for ny flight segment, as long as minimum and maximum limits are not violated. The flight schedule is considered asa chain of fight segments, or legs, tha each airraft follows. The schedule ultimately returns the aircraft to its starting point, resulting in a ‘rotation’. Consider the following rotation: Delhi ~ Hyderabad ~ Cochin ~ Chena ~ Delhi This case isbased on, P. Masse and R. Gibrat, Application of Linear Programming to Investments in Electric Power Industry, Management Science, Jan 1957. ** This ease is based on ‘D. Wayne Darnell and Carin Loflin, National Airlines Fuel Management and Allocation Model, Interfaces’ Feburary, 1977,Fuel Requirements and Limits ((,000 gallons uniess specified otherwise) % ‘ 4200 8 20 7,500 Do 95. 25 BO 900 ‘The fue! for any one ofthese fight segments may be hought at its departure city or itmay be purchased aa previous city in the sequence and ‘tankered” for the flight. Of couse, i takes fuel to cary fuel, and thus en economic trede-off ‘between purchasing fuel atthe lowest-cos location and tankerag it all around the country must be made. Inthe table above the column ‘Regular Fuel Consumption’ takes into account fuel consumption if the minimum ‘amount of fuel i on board, and the column ‘Additional Fuel Burned” indicates the additional fuel burned in each flight segment per gallon of“tankered’ fuel caried; tankered fuel refers to fuel above the minimum amount. ‘The fuel originally carried into Delhi should equal fuel carried into Delhi on the next rotation, system to be in equilibrium, If f= Leftover fue! inventory coming ito city i(1,000 gellons) and x, = Ameunt of fuel purchased at city i1,000 gallons) then, (, + is the amount of fuel on beard he aircraft when it departs city i Questions for Diseussion Develop an LP model and solve it 1 suggest the optimal fuel quantiy to be purchased by the Ailines. order for theDuality in Linear Programming “A boat can't have two captains.” ~ Akira Mori Preview The chapter deals with how to find the marginal value {also known as shadow price) of each resource. This in tum helps the resource to be best utilized. Learning Objectives After studying this chapter, you should be able to ‘+ appreciate the significance of the duality concept. ‘© formulate the dual LP problem and understand the relationship between primal and dual LP problems. ‘© understand the concept of shadow prices. chapter Chapter Outline 5.1 Introduction 5.2 Fermulaion of Dual Linear Programming Prablem + Self Practice Problems A, © Hints and Answers 5. Standard Results on Duality 54 Marageral Significance of Duality 5.5 Advantages of Duality Conceptual Questions Self Practice Problems B + Hints and Answers (Chapter Summary Chapter Concepts Quiz ‘Appendixc Theorems of Duality pooOperations Research: Theory and Appications Dual LP problem provides. ussful economic Information about ‘worth of resources to be used ‘Shadow price represents increase in the objecive function value due to one/unit increase Inthe right nana side (resource) of ary. censtrant 5.1 INTRODUCTION ‘The term ‘dual’ in a general sense implies two or double. The concept of duality is very useful in ‘mathematics, physics, statistics, engineering and managerial decision-making. For example, in a two-person ‘game theory, one competitor’s problem is the dual of the opponent's problem. In the context of linear programming, duality implies that each linear programming problem can be analyzed in two different ways but would have equivalent solutions. Any LP problem (either maximization and minimization) can be stated in another equivalent form based on the same data. The new LP problem is called dual linear programming problem or in short dual. in general, it is immaterial which of the two problems is called primal or dual, since the dual of the dual is primal. For example, consider the problem of production planning. By using primal LP problem, the production ‘manager attempts to optimize resource allocation by determining quantities for each product to be produced ‘that will maximize profit. But through a dual LP problem approach, he attempts to achieve a production plan that optimizes resource allocation in a way that each product is produced at that quantity so that its ‘marginal opportunity cost equals its marginal return. Thus, the main focus of a dual problem is to find for each resource its best marginal value (also called shadow price). This value reflects the scarcity of ‘the resources, ie. the maximum additional prices to be paid to obtain one additional unit of the resources in order to maximize profit under the resource constraints. Ifa resource is not completely used, ie. there lack, then its marginal profit is zero. The shadow price is also defined as the rate of change in the optimal objective function value with respect to the unit change in the availability of a resource. To be more precise for any constraint, we have optimal objective function value Unit change in the availability of resource ‘The interpretation of rate of change (increase or decrease) in the value of objective function depends en ‘whether we are solving a maximization or minimization LP problem. The shadow price for élessthan or equal 10 (S) type constraint will always be greaier than or equal to zero, This is because increasing the right-hand side resource value cannot make the value of objective function worse. Similarly, the shadow price fora greater than ‘or equal to (2) type constraint will always be less than or equal to zero because increasing the right-hand side resonance value cannot improve the value of the objective function. ‘The format of the simplex method is such that solving one type of probiem is equivalent to solving the other simultaneously. Thus, ifthe optimal solution to one is known, the optimal solution of the other ccan also be read from the ¢, =, row of the final simplex table. In some cases, considerable computing time can be reduced by solving the dual. Shadow price = 5.2 FORMULATION OF DUAL LINEAR PROGRAMMING PROBLEM ‘There are two important forms of primal and dual problems, namely the symmetrical (canonical) form anc the standard form. 52.1 Symmetrical Form Suppose the primal LP problem is given in the form Maximize Z,=¢,x, + X97. + 6%, subject to the constraints yx, + ayx +... + a,x, Sy a3, +an%, +... + ay x, Sb, Aig Xy * Ogg Xp too F Oy 5, and Xppdy een hy 20. Then the corresponding dual LP problem is definedimage not availableOperations Research: Theory and Applications or £ (Units of resource /, consumed per unit of variable, x,) (Units of variable x) xt S Units of resource, i available and 20, forall j Dual LP Problem Minimize (cost) 2, ~ £ b,y;~ E (Unis of rsource, #) (Cost per unit of resource, #) fe subject to the constraints 5 anes oe (sits of a resource, j consumed per unit of variable y,) (Cost per unit of resource, i) a Prot pr unt for eacl aciviy (abi) 3 and ¥, 2 0, for all f r ‘i 4 From these expressions of parameters of both primal and dual problems, it is clear that forthe unit of measurement to be consistent, the dual variable (j;) must he expressed in terms of return (or worth) per unit of resource i. This is called dual price (simplex multiplier or shadow price) of resource i. In other ‘words, optimal value of dual variable associated with a particular primal constraint indicates the marginal change (increase, if positive or decrease, if negative) in the optimal value of the primal objective function from increasing (or decreasing) in the right-hand side value (resource) of a constraint. For example, if 9) = 5, then this indicstes that for every additional unit (up to a certain limit) of resource 2 (resource associated with constraint 2 in the primal), the objective function value will incresse by 5 units. The value ‘y= 5 is called the marginal (or shadow or implicit) value or price of resource 2. Similarly, for any two feasible primal and dual solutions, a finite value of objective function satisfy the inequality Z, < Z,, This inequality is interpreted as: Profit Worth of resources. Thus, so long as the total profit (return) from all activities i less than the worth of the resources, the solution of both primal and dual are not optimal. The optimality (maximum profit or return) is reached only when the resources have been com pletely utilized. This is only possible ifthe worth ofthe resources (i. input) is equal to profit (i.e. output). 5.2.3 Economic Interpretation of Dual Constraints ‘As stated earlier, the dual constraints are expressed as: Dew il Since cj represents profit (in Rs) per unit ofactivity x, therefore the LHS quantity jy, should also be in rupees per Unit but it represents cost (due to profit with negative sign). Since coefficients a, represents the amount of resource b, consumed by per unit of activity x), and the dual variable y, represents shadow price per unit ofresource 6, the quantity 2 a,,», (=z) should be total shadow price ofall resources required to produce cone unit of activity, ‘As weknow that fora maximization LP problem, if ~2,>0 corresponds to any non-basic (unused) activity (or variable), then the value of objective function can be increased. This implies that a variable, x, valuecan be increated fom zeroto a positive level provided ts unit profit (c) is more than its shadow price, (Shadow price of resources used per unit of activity, x) < Profit per unit of activity x, ey 20 5.2.4 Rules for Constructing the Dual from Primal ‘The rules for constructing the dual from the primal or primal from the dual, when using the symmetrical form are: |, Adual variable is defined for each constraint in the primal LP problem and vice versa. Thus, in a given primal LP problem with m constraints and n variables, there exists a dual LP problem with m variables and 7 constraints and vice-verse.Dalit in Linear Programming 2 The right-hand side constants by, by, ... b, of the primal LP problem becomes the coefficients of the ual variables y, ¥>y-.» Yin the dual objective function Z,. Also the coefficients c1,.¢3, .- ¢, of the primal variables x, x,,"... X, in the objective function become the right-hand side corstants in the dual LP problem. 3. Fora maximization primal LP probiem with all < (less than or equal to) type constraints, there exists, ‘a minimization dual LP problem with all 2 (greater than or equal to) type constraints and vice versa. ‘Thus, the inequality sign is reversed in all the constraints except the non-negativity conditions. 4, The matrix of the coefficients of variables in the constraints of dual is the transpose of the matrix of coefficients of variables in the constraints of primel and vice versa, ie. coefficients of the primal Yariables x4, xp, ..., 4, in the constraints ofa primal LP problem are the coefficients of dual variables in first, second, ..., nth, constrains for the dual problem, respectively. 5. If the objective function of a primal LP problem is to be maximized, the objective function of the dual is to be minimized and vive versa, 6 If the ith primal constraint vice versa. ‘The primal-dual relationships may also be remembered conveniently by using the following table: ‘Dual Feriables ‘Primal Varables Maximize 2, ‘ s Sr > 4 a 4 $6, i a iy e qi a % oy @q | Sb %n Sm no Sn) tm | $b Minimise | 26 Ba + 4 = | Cpuat ebjective tinction cvetficents TL pth cual constrain ‘The primal constraints should be read across the rows, and the dual constraints should be read across the columns. Example 5.1 Write the dual to the following LP problem Maximize Z= x, —x, + 3ry subject t0 the constraints Ontnty and XyXy4,2 0 Solution In the given LP problem there ere m= 3 constraints and n = 3 variables. Thus, there must be ‘m = 3 dual variables and m = 3 constraints. Further, the coefficients of the primal variables, c ¢; = 1, ¢, = 3 become righthand side constants of the dual. The right-hand side constants b, ~ 10, = 2, b,= 6become the coefficients inthe dual objective function. Finally, the dual must havea minimizing objective function with all> type constraints. Ify,, 5 and y, are dual variables corresponding to three primal constraints in the given order, the resultant dual is Minimize Z,= My, +2y, + 615 subject to the constraints Oy + +2521, @ ¥—¥,- Iz and By Inds 20 Example 5.2 Write the dual of the following LP problem Minimize Z, = 3x, ~ 2x, + 4x5 10, Gi) 2 Gil) 2x, — 2x, ~3xy $6 Gi) yy) ~ 9) — 3923 subject to the constraints @ 3x 45x +4r27, GD Gy + +324, Gi) Tx 2, - (iy) x, - 2x + 5x23, ) 4x, + 7-20 22 and xy xy; 20 crx)Operations Research: Theory and Applications Solution Since the objective function of the given LP problem is of minimization, the direction of each inequality has to be changed to 2 type by multiplying both sides by —1. The standard primal LP problem. s0 obtained is: Minimize Z, = 3x, - 2x, + 4x, subject to the constraints @ 34 45y 44427, @ 6x4 x +324, Gi) Ir, 42x, + 2-10 G) x)= 28) +5423, (W) Any + Tey 222 and Xp ty 20 If yp, Y2.JJg and yg are dual variables corresponding to the five primal constraints in the given order, the dual of this primal LP problem is stated as: Maximize Z,~ Ty, + Ay, ~ 10y, +394 + 2y4 subject to the constraints © 3, +6y,- T+ Yet AysS3, CVS, + Hy +2yy ~ Dy + Wy S-2 GD 4y, +3y, + yy + Sy4- 29,84 and Viv Var Vas Var Vs 20 Example 5.3 Obtain the dual problem of the following primal LP problem: Minimize Z =x, + 2x; subject to the constraints © 2, +4 5160, Gi) xx =30, (i) 4,210 and xy 20 Solution Since the objective function of the primal LP problem is of minimization, change all < type constraints to 2 type constraints by multiplying the constraint on both sides by -1. Also write = type constraint equivalent to two constraints of the type 2 and <. Then the given primal problem can be written, ne Minimize Z,= 1x, + 2x, subject to the constraint 0 ~2x, ~ 4x, 2-160, () x,-m 230 (@) x)- x $- 30 of —x, +x,2-30, @) x2 10 and p20 Let yyy yy 94, and yy be the dual variables corresponding to the four constraints in the given order. ‘The dual of the given primal problem can then be formulated as follows: Maximize Z,= — 160, + 30y,— 307, +10), subject to the constraints: © -M +t y—-YtS @ -M-y2 +582 a Yo IarYye M420 Let y = yy ~¥3 Wy, ¥3 2 0). The above dual problem then reduces to the form Maximize Z,= ~160y, +30) 10y4 subject to the constraints O -MtyFySI, @ ~4,-»s2 and Yp¥2 0; y being unrestricted in sign Remark To apply rule 6, rote that the second constraint in the primal is equality, therefore the comresponding second dual variable y (= y,~y,) should be unrestricted in sign.Duality in Linear Programming Example 5.4 Obtain the dual problem of the following primal LP problem: Minimize Z, = x, ~ 3x, — 2x5 subject to the constraints © By +24S7, (2,-4.2 12, Gi) —4x, + 3x, + BK, = 10 and X42 0; x, unrestricted in sign Solution Since x, is an unrestricted variable, therefore, it can be expressed as the difference of two non- negative variables, ic. x,— xi - xf, x5, x3 = 0. The given LP problem can then be written as Minimize Z, = x, ~ 3x, -2 (x5 - x) subject to the constraints @ 3x - x42 (x8 — af) $7 or ~3x +xy-2 (xh - af) B-7 @ %— 4x21, GH) 4x, $3x, 4 BCS — x)= 10 and 4m SF 20 Let y,,.¥, and y, be the dual variables corresponding to three primal constraints in the given order. ‘As the given problem is of minimization, all constraints can be converted to > type by multiplying both sides by -1. Since the third constraint of the primal is an equation, the thitd dual variable 4 will be ‘unrestricted in sign. Now the dual of the given primal can be formulated as follows: Maximize Z,=~7y, + 12), + 10) subject to the constraints @ -3y, +2451, @ HA +34 S-3 GH -29), +84 S-2 and Yin Ya2 0; yy untesticted in sign. Example 5.5 Obiein the dual of the following primal LP problem Maximize Z, = x, ~ 2x + 3% subject to the constraints @ -2x, + 2) +3x,=2, (i 2x, + 3x) + 4x, and Hy Xy xy 20 Solution Since both the primal constraints are of the equality type, the corresponding dual variables y, and, willbe unresticted in sign. Following the rules of duality formulation, the dual ofthe given primal LP problem is Minimize Z,=2y, + y subject o the constraints @ —2y, + 29,21, @ yj +3,2-2 Gi) 3y, +4, 23 and Yi. Yn unrestricted in sign. Example 5.6 Write the dual of the following primal LP problem Maximize Z= 32; +x, +2xy ~ xy subject to the constraints OAk-_tIyty=L — Myty- Hey and xp m2 O and xy, x, unrestricted in sign Solution Here we may apply the following rules of forming a dual of the given primal LP problem. (@ The x, and x, variables in the primal are unrestricted in sign therefore, the third and fourth constraints in the dual shall be equalities. (The given primal problem is of maximization; the first two constraints in the dual will therefore be > type constrains ince both the constraints in the primal are equalities, the comesponding dual variables y, and y, will be unrestricted in sign. Ify, and y, are dual variables corresponding to the two primal constraints in the given order, the dual of the given primal can be written as: i)Operations Research: Theory and Applications Minimize 2, = yy + 3y subject to the constrains © 2%, +723. @) y+, =-1 Yye Hp umrestricted in sign. and Gi) yy +221, Gi) 394-93 SELF PRACTICE PROBLEMS A ‘Wite the dual of the following primal LP problems 1 Max Z,= 2x + 5x, + 6% subject to (i) 5x, + 6% — x 53 | -2y + tay 24 | (i %)- 5% + 3y 51 | (Wy) =~ 3% + 7m 56 and Mt 20 | [Sambalpur MSc (Maths), 1996] | Min 2 = 74 + 9% + 8% | ‘subject to (Ox, + By + xy EO | (94,4 0g +4q24 | (i) 4 sm 4 Sy 24 (iy) m+ 8m +25 27 and Mn 20 Max 2,= 24, + 3% 4% subject t0 (I 4x, + 9 + = 8 (Kt 2m + Syed and sa 20 Mex Z,= 34, + % + 3% ~ X subject 10 (0. 2x; = m4 95) +X (i) Hem te and Ho fy M2 0 Min Z, = 2x, + 8% 4 4x Subject 10 (i) 2x, + Sky + 84 2 2 W) Ins a 4TH =D (i) my + ay t Oy S 8 and Hy Mp 2 Oy IS unrestricted 10, 0. Min Zp + 9+ subject 0 (i) x, ~ 3% + 444 = 5, (i) 2y~ med W%-% <3 ‘and i Me 2 0, % fs unrectricto, [Meenst, MSe (Maths), 1967) 7. Max 2, = 8%, + Sx subject 10 () x OE 2) BH + TH ‘and Hi 20, 8, Mat Z, = 8%, + 64, + 8x + 12x, supedt 10 () 30x, + 20% + 254, + 40x, < 800 (i) 25x, + 10x + 7a + 15x, < 260 () 4-20 w28 HM Hs MBO 9, Min Z, = 104, + 10% + 11% sutject 10 () 4x + Gly + Sxy 2 400 (W) 12% + 10x + 10% > 1,200 (i 105, + 16x, + 7's 1500 (i) 9g 2 50 W) x= es0 and Muto Hz D Min Z, = 2x, ~% + 34 sublet 0 (y+ 2m + % 212 (7 2y2-8 ) OS 42m 4 ye 26 ay Hy 2 0, uirestcted and and HINTS AND ANSWERS. Min Z,= 39) 4475 +9554 6% subject to. 5), —3y9 + 55394 22 (i) 6}, +3) ~S~ 3425 Gi =, +4), + 39, + 26 and Mp Ip Ye Ig BO Max 2) = 3), +45 4+ 2% nibject ©) & +3 + 44+ 487 Gi) 2, + 6y) + yy + 5% 53 GD yy + 492 + Sry + 2g 38. and Yay Yar Yar Wy ZO Min Z, = 6 + 4 subject to) 4, + 9,22, i) yy + Sie 1 Yy J untesrcted in sign, (+2423 and 4, Min Z, = yy + 3p subject ©) 29, typ 3 Gi ~y, ty, 21 Gi) 3) - yp 23 Wy tHe and Yiods unesticted in sign 5. Max Z, = 29, + 3y - 5y5 subject 10 ty, + 32-9 $2 ayy + aa ay 53 5, + Ty - 6 = 4 and ys J 2 O and yy unrestricted, 6. Max Z,=-5y ~ 393 + 43, subject © @) —y 9221, G39 +29, +2 ST 4 -yS1 and Nye Yy 2 0 and yy is unrestricted.Duality in Linear Programming Ea 7. Max Z, = 2%, — 4, ay, + 10, - 35 28 subject 10) ¥) +5 £8, Gil) ~6y, + Ty <3 2, + Ty toy 28 and 1; 2 0 and y, is unrestricted. 40y, +15, 212 & Min 2, ~ 800), + 25099.» y5 5y4 yp Np ® O and ys rg ae unrestricted subject to 30y, + 2yy + 4 28 5.3 STANDARD RESULTS ON DUALITY See appendix for detail proof of the following standard results: |. The dual of the dual LP problem is again the primal problem. 2. Ifeither the primal or the dusl problem has an unbounded objective function value, the other problem has no feasible solution. 3. Ifither the primal or duel problem has a finite optimal solution, the other one also possesses the same, and the optimal valve of the objective functions of the two problems are equal, ie. Max Z, = Min Z, This analytical result is known as the fimdamental primal-cual relationship. These results aie summarized as follows. ‘Dual Problem (Max) Feasible Max Z, 9 += Infessible Unbounded or infeasible 4. Complementary slackness property of primal-dual relationship states that for a positive basic variable in the primal, the corresponding dual variable will be equal to zero. Altematively, for a non-basic variable in the primal (which is zero), the corresponding dual variable will be basic and positive. 5.3.1 Principle of Complementary Slackness ‘The principle of complementary slackness establishes the relationship between the optimal value of the main variables in one problem with their counterpart slack or surplus variables in other problem. Thus, this principle can also be helpful in obtaining the primal LP problem solution when only the dual solution is, known. To illusirate this concept, let us consider the example of production planning. 1, Ifat the optimal solution of a primal problem, a primal constraint has a positive value of aslack variable, the corresponding resource is not completely used and must have zero opportunity cost (shadow price). This means that having more of this resource will not improve the value of the objective function. But ifthe value of slack variable is zero in that constraint, the entire resource is being used and must heve @ positive opportunity cost, i. additional resource will improve the value of objective function by allowing more production Since resources are represented by slack variables in the primal and by main cost variables in the dual, therefore, the principle of complementary slackness states that for every resource, the following condition must hold: Primal slack variable * Dual main variable = 0 In cases where the resources are not completely used for producing any unit of a product, the ‘opportunity cost of such resources exceeds the unit profit from that product. But, if few units of the product are made, the opportunity cost of resources used must equal the profit from each unit of the product, so that the surplus produet cost is zero. A primal variable in primal LP problem represents the quantity of product made, and the surplus variables in dual represent surplus product costs. In this case the principle of complementary siackness states thet for every unit of the product, the following condition must hold. Rn Primal main variable x Dual surplus variable = 0‘Operations Research: Theory and Appitations 54 MANAGERIAL SIGNIFICANCE OF DUALITY ‘The importance of the dual LP problem is in terms of the information that it provides about the value of the resources. The ecoromic analysis in this chapter is concemed about deciding whether or not one should secure more resources ard how much should one pay for these additional resources. ‘The significance of the study of dual is as follows: (The right-hand side of the constraint represents the amount of a resource available and the associated dual variable value is iterpretated as the maximum amount likely to be paid for, for an additional unit of the resource. However, such an interpretation is not always corect because of the two types of costs involved with each resource: (i) sunk cost that is not alfected by the decision made, this will be incurred no matter what values the dual variables assume, and (ii) relevant cost; which depends ‘on the decision made. This will vary with the values of decision veriabies (i) The maximum amount that should be paid for one additional unit of a resource is called its saadow price (also called simplex multiplier). (Gi) ‘The total marginal value of the resources equals the optimal objective function value, The dual variables equal the marginal value of resources (shadow prices). (iv) The value of the ith dual variable represents the rate at which the primal objective function value will increase by increasing the right-hand side (resource value) of constraint i, assuming that all other data remein unchanged. Remark When the cost of a resource is sunk, the shadow price represents the amount of money that is to be pid for the additional unit of resource. When the cost ofa resource is relevant, the shadow price represents the amount of money by which the value of the resource exceeds its cost. ies that when the resource cst is relevant, the shadow price represents the extra money over the normal cost that is likely to be paid for ‘ne unit of the resource. Example 5.7 The optimal solution simplex table forthe primal LP problem Maximize Z= 3x, + 4x +25 subject to the constraints @ x, + 2x, + 3x, $90 (time for operation 1), (i) 2x, + x, + x5 $60 (time for operation 2) ii) 3x, + x, + 2x, £80 (time for operation 3) and xy ty 20 is given below: 9? 3 4 1 0 0 0 4 % 40 0 1 16 5-10 * 10 1 ho nlewia' 18 4 Saemene ss 10 | 0 0 86 86 S06 0-26 106-23 (@)_ Find the solution, maximum profi, idle capacity and the loss of the total contribution of every one unit reduced from the right-hand side of the constraints. (b) Write the dual of the given problem and give the initial simplex table, Solution To illustrate the interpretation of the dual variables let us consider this example in a profit ‘maximization production context with constraints on time input resource. ‘The primal LP problem in this example is concerned with maximizing the profit contribution from three products say A, B and C, while the dual will be concerned with evaluating the time used in the three operations for producing the three products.oe 159 The productive capacity of the three operetions is @ valuable resource to the firm; the production ‘manager wonders whether it would be possible to place a monetary value on its Worth. If yes, then now much? His problem can be solved along the following lines, (a) The optimal solution of the primal problem can be read from the given table as: 2, = 10, x,= 40, x, = 0 and Max (profit) Z= 3 x 10 +4 x 40 +1 x0 =Rs 190 ‘Since the value of slack variable, s,, in the optimal solution is 10, the ile capacity in operation 3 is 10 units. Interpretation of dual variables The absolute values ofthe numbers in the c,~ =, row, of final simplex table, under the slack variables columns represent the values of dual variables, i. marginal value (shadow price) of resources, 1 and 2. The optimal solution of dual LP problem 0/6, y= 23, yy =O and Mix Z, ‘The value of y;, 72 and ys shows the worth of one hour in cach operation. That is, the optimal valuc of the dual variable y, represents the per unit price (worth or marginal price) of the first resource (ie. ime available in operation 1), viz., Rs 10/6 or Rs 1.66. This means that marginal contribution of operation 1 to the total profit is Rs 1,66. In ether words, if one productive unit is removed from the right-hand side of the first constraint, the total contribution would reduce by Rs 1016, or Rs 1.66. However, if one productive unit is removed up to the extent of idle capacity from the right-hand side Of the third constrain! (time available in operation 1), the total contribution would not be alfected. (©) Lety,,., andy, be the dual variables representing the marginal value of one unit of the resource that corresponds to constraint |, 2 and 3, respectively of the primal. That is to say that, the firm pays per unit price of y,, yp and y; for the time available in the three operations. The firm, however, wants to make sure that it makes as much profit as it would if it remained in business for itself. Then the dual of the given problem is stated below: Minimize Z = 90), + 60y, + 8), subject to the constraints Ot +HsE3, Wt yt WEA — Gi) By + yy F2VZBM and Vy Ia 20 Interpretation of dual constraints Suppose the competition dicates that the company should also produce product C, To see how can this be achieved, examine c,~z, value under column-x,, Producing Cwill be economical only if c, > z. This is possible by either increasing the profit per unit c, or decreasing the impated cost of the used resources 7, (= 3y; + y+ 2y,). Am increase in the unit profit may not be possible because the company wants to remain competitive in the market. A decrease in zis possible as this requires ‘making improvement in operation 3 by reducing the unit usage of operation times. Let ff, and ts represent the proportions by which the unit times of three operations are reduced. The company then needs to determine , ‘and f, such thatthe new imputed cost, of three operations falls below the unit profit, ie. 3 (1-1) ¥, +H, +2 (1-4) <1 Forgiven values of y,= 5/3,y3=2/3andy, weet 31-4) (53) + (1-4 )23)<1 or 54+ 24> 16 Thus, any values off, and fy between 3 and 4 that satisfy 54, + 21, > 16 should make the product profiteble. ‘However, this may not possible because this requires a reduction in the times of operation 1 and 2, which is impractical Example 5.8 XYZ Company has three departments — Assembly, Painting and Packing. The company ‘can make three types of almirabs. An almirah of type Iequires one hour of assembly, 40 minutes of painting and 20 minutes of packing time, respectively. Similarly, an almirah of type Il needs 80 minutes, 20 minutes and one hour, respectively. The almirah of type IMI requires 40 minutes each of assembly, painting and packing time. The total time available at assembly, painting and packing departments is 600 hours. 400 hours ‘and 800 hours, respectively. Determine the number of each type of almirahs that should be produced in order to maximize the profit. The unit profit for types 1, I and Ill is Rs 40, 80 and 60, respectively. Suppose that the manager of this XYZ Company is thinking of renting the production capacities of the three departments to another almireh manufacturer - ABC Company. ABC Company is interested in The shadow price Is represenied by the ¢)~ 2; value Under stack varablo ccdumns in the ‘optimal simplex tableOperations Research: Theory and Applications ‘minimizing the rental charges. On the other hand, the XYZ Company would like to know the worth of production hours to them, in each of the departments, in order to determine the rental rates. (a) Formulate this problem as an LP problem and solve it to determine the number of each type of almirahs that should ‘be produced by the XYZ Company in order to maximize its profit. (b) Formulate the dual of the primal LP problem and interpret your results. [CA, May 1994; Dethi Univ., MBA, 1999, 2002) Mathematical formulation The production data given in the problem may he summarized as below Types of Almira ~ Number of Hours Required per Unit Profit per Unit Assembly Painting Packing 1 1 2 0 ri 43 13 1 80 ry 28 28 2 © ‘otal availabilty (hrs) 500 400 800 Let x,, x; and x, be the number of units of the three types of almirahs, respectively, to be produced, ‘The given problem can then be represented as an LP model as: Maximize Z = 40r, + 80x, + 60x, subject to the constraints © + GB)x, + 2/3) $600, Gi) (213) x, + (1/3) xy + (2/9) x, $400 Gi). (13) x, + + 213) x, $800 and Hp Fy 20 If ABC Company pays the per hour rental price of y,, y; andy, for assembly, painting and packing departments, respectively, the company XYZ must get a total rent equal to: 600y, + 400), + 800y, Since company ABC wants to earn as much profit as possible, its objective is to Min (total rental charges) Z, = 600y, + 400», + 800), After knowing this rental value, the objective of ABC Company is toknow what minimum offer it should give to XYZ, The given data show that an almirah of type I requires 1 assembly hour, 2/3 painting hour ‘and 1/3 packing hour. The imputed cost (ie. rent) of time used for making almirah lis: y, + (2/3) + (V/ 3).yy, IEABC Company used that time to make almirah I it would earn Rs c, = 40 in contribution to profit, ‘and so it will not rent out the time unless: y; + 2/3)» + (1/3) ys ® 40. Similarly, the company will work for almirahs Il and IIL Since the total rent of all the departments should be greater than or equal to the profit from one unit of the almirah, the dual objective function slong with the constraints that determines for ABC, the value of the productive resources, can be written as: Minimize (total rent) Z, = 600, + 400 + 800), subject to the constraints O 1+ 2B)y)+ (1B)y9240, GH (43) y, + (13) +y5 280 Gi) (23) yy + @/3)y+ 2/3) y52 60 and Py Ins 20 Here the dual objective function, ie, the minimum total rent acceptable by the ABC Company is equal to the primal objective funetion, i.e. meximum profitthat could be earned by XYZ Company from its resources. ‘The dual variables y,, v5 and y, represent the rental rate of the various departments corresponding to the slack or unused capacity (called slack variables 5,, sy, and sy, respectively) of these departments and are called the shadow prices or marginal profitability of various departments.Duality in Linear Programming Example 5.9 A firm manufactures two products A and B on machines I and II as shown below: 1 30 20 300 n a 10 10 Profit per unit (Rs) 6 8 ‘The tote time available is 300 hours and 110 hours on machines I and I, respectively. Products A and B contribute a profit of Rs 6 and Rs 8 per unit, respectively. Determine the optimum product mix. Write the dual of this LP problem and give its economic interpretation. Mathematical formulation The primal and the dual LP problems of the given problem are Primal problem Dual problem 4%, and x, =number of units of A and B to | y; and y,= cost of one hour on machines [ and I, be produced, respectively respectively MaxZ,, = 6, + 8 subject to the constraints Min Z,= 300», + 110¥, subject to the constraints 30x, + 20x, £300 30y, + Sy 26 5x + 10x, < 110 2oy, + 10y, 28 and Xpme 0 and Py I BO ‘Solution of the primal problem The optimal solution of the primal problem is given in Table 5.2. Pe 6 5 0 0 © % + rae wo 10 SERIO 5 Se Pana. scene a Sa Optimal Solution 2% any, oo imo 60 of Primal Problem ‘Table 52 indicates that the optimal solution is to produce: x, = 4 units of product A; x, = 9 units of product B ang Z, = total maximum prof, Rs 96. Solution of the dual problem The optimal solution of the dual problem can be obtained by applying the Big-M method. The optimal solution is shown in Table 5.3. b> m0 ° 300 ” 10 1 0-120. Mo Seeds 410 Ya 610 ° 1 110 320, phat sonnet Z=% by 0 0 4 9 ‘of Dual Problem The optimal solution as given in Table 5.3 Rs 1/10 per hour on machine I; y= Rs 6/10 per hhour on machine IT and Z, = total minimum cost, Rs 96, The values of y, =Rsl/10 and y, = Rs 3/5, indicate the worth of one hour of machine time of machines and I respectively. However, since the machines hours can be increased beyond a specific limit, this result, hholds true only for a specific range of hours available on machines | and Il Copyrighted materialTable 5.4 Operations Research: Theory and Applications Comparison of the solutions For interpreting the optimal solution of the primal (or dual), its solution values can be read directly from the optimal simplex table of the dual (or primal). The method can be summarized in the following steps. 1. The slack variables in the primal, correspond to the dual basic variables in the optimal solution and vice versa. For example, sis the slack variable of the first primal constraint. It corresponds to the first dual variable y,. Likewise s,,, corresponds to the second dual variable y,, Moreover, sy and 5), are the dual surplus variables and they correspond to x, and x,, respectively of the primal problem. ‘The correspondence between primal and dual variables is summarized in Table 5.4. 2 The value the cj ~ 3, row under columns of the slack/surphus variables, ignoring the negative sign, directly gives the optimal values of the dual/primal basic variables. For example, cy ~ z, = 1/10 and 4—z,= 6110 values in Table 5.2 under primal slack 5, and s,, correspond to solution values of dual variables »; and y in Teble 5.3 and vice versa, The primal-dual relationship for these problems which are feasible can be summarized in Table 5. “+ Values of the basic variables + (—5) values in the non-basic sephs variable columns + (G)~=) values in the non-basic + Value of the basic variables slack variable columns ‘The optimal value of the objective function is the same for primal and dual LP problems. ‘The dual variable y, represeats the worth (dual price or shedow price) of one unit of resource /. For example, for y, the worth of ime on machine I, equals Rs 1/10 and fory, the worth of time on machine TI, equals Rs 6/10, as shown in Table 5.3. ae Example 5.10 company wishes o get atleast 160 million ‘audience exposures’ the numberof times one of the advertisements is seen or heard by a person. Because of the nature ofthe product the company wants at Iecast 60 million month and at east 80 million of the exposures to involve persons between 18.and 40 years of age. The relevant information pertaining to the two advertising media under consideration—magazine and television is given below: ‘Magazine Television . 0 200 ‘+ Audience per advertisement (million) 4 0 ‘Audience per advertisement with monthly income ‘over Rs, 10,000( million) 3 10 ‘+ Audience (per advertisement) inthe age group 18-40 (milion) 8 10 ‘The company wishes to determine the number of advertisements to be released each in magazine and television so as to kezp the advertisement expeneiture to the minimum. Formulate this problem ss a LP problem. ‘What will be the minimum expenditure and its allocation among the two media? Write ‘dual’ of this problem. Solve the ‘dual’ problemto find answer to the problem.Dually in Linear Programming [itaxy Solution The primal and dual LP problems ofthe given problem are: Primal Problem Dual Problem xp) = mumber of advertisements in magazine |,’ = shadow price (or worth) of one unit of and television, respectively. advertisement over audience characte- ristics, respectively. Minimize Z, = 40x, +200x, Maximize Z,=160y, +60y,+ 80), subject to 4x, +40r,>160 subject to 4y, + 3y,+8y, $40 3x + 10x260 40y, + 10y, + 1095 200 Rr, + 10,280 and 493.9520. and 4,20. Example 5.11 XYZ manufacturing company operates a three -shift system at one ofits plants. Ina certain section of the plant, the number of operators required on each of the three shifts is ¢s follows: Day (6am.t02 pm) 50 ‘Afternoon (2 p.m to 10 psn.) a ‘Night (10 pam. 106 p.m) 10 ‘The company pays its operators at the basic rate of Rs. 10 per hour for those working on the day shift. For the aflernoon and night shifts, the rates are ene and a half times the besic rate and twice the basic rate, respectively. Inagreement with each operator at the commencement of his employment, he is allocated to one of three schemes 4, B or C. These are as follows: ‘A: Work (on average) one night shift, one afternoon shift, and two day shifts in every four shifts B : Work (onc average) equal number of day and afternoon shifts, C+ Work day shifts only. Inschemes A and B,it is necessary to work strictly alternating sequences of specified shifs, as long as the correct proportion of shifts is worked in the long run. (©) Formulates tincar programming model o obtain therequired numer of operators at minimum cos. (®) By solving the dual of the problem, determine how many operators must be employed under eack of the three schemes. Does this result in over-provision of operators on any one of the three shifts? Solution The primal and dual LP problems ofthe given problem are: Primal Problem Dual Problem 2, ¥;and x, = number of operators employed under | _ y,,., and y, = shadow price (or worth) per unit scheme 4, B and C respectively. of resources—operators in three 1 ie sil shift respectively. Minimize Z,= 20> bat h(Gat5) MaximizeZ,=10y, +241, +50), 1 subject to the constraints ~ (i2)yy + U2)ys = 2572 +25, 4103 7 st0 subject to the constraints and Ppl ds20. (Uia)x, 2 10, (1)x, + (V2)x 2 24 (12)x, + (/2)x, 3,250 and yxy 20. 5.5 ADVANTAGES OF DUALITY 1. It is advantageous to solve the dual of « primal that has a less. number of constraints because the ‘number of constraints usually equals the number of iterations required to solve the problem.(Operations Research: Theory and Applications 2 This avoids the necessity for adding surplus or artificial variables and solves the problem quickly (the technique is known as the primal-dual method). In economics, duality is useful in the formulation of the input and outpat systems. It is also useful in physics, engineering, mathematics, etc. 3. The dual variables provide an important economic interpretation of the final solution of an LP problem. 4. It is quite useful when investigeting changes in the parameters of an LP problem (the technique is known as the sensitivity analysis), 4. Duality is used to solve an LP problem by the simplex method in which the initial solution is infeasible (the technique is known as the dual simplex method). CONCEPTUAL QUESTIONS Detine tne dual of a near programming problen. State the | 7, functcnal properties of duality, What is @ shadow price? How does the concept felate to ne dual of an LP problen? 2. Ee he pn ara A. Howcon scoot yb ek menage deco EWhere tone day WEG Tees | A Sn tt ie len Kann Ya ai ga 5, What duality? Whet is the significance of dual variables in an | 10. Prove that the necessary and sufficient condition for any LP ees Covet Se meme ard ei ce ea Le 6. State the general rules for formulating a dual LP problem from | ‘both have feasible solutions. eed | SELF PRACTICE PROBLEMS B i al aac ale i Ta conpad pests pcdicI ad alles we eyes Mourn reer et ee subjed to (i) $x,+6%-%53 (il) -2q+m+4ysd below: (iil) ay Sky+ B%yS1 (iv) — 3x, — 3% +7 56 Raw Matelale Number of Unite of Raw Material Required and Fete te ‘to Produce One Unit of Product 2 hn z.a74.+34 ee pe Po Riek he vers ensuaeraees stood io ah Sipe ye8 21rd Tesntpat coin df pocaes, ¥rd ZnS ms Ms Ro Ma and 50, respectively. The number of units of raw material available 4. Max 2, = 34 + 3 + By — Hy are 36, 60 and 45, respectively. Socio Oe ee eKEUM a IR aRE to ches Se eect i nasi ge ne i) Can te teal lst we ne sant nee wee ase gp ane aergca | Trot we metas comely Re S0 aa gee: |. 2s agree oe ae i ae Sc aa aa redial | calories per unt. respectively and 200,900 and 500 prctein units per fy M2 Or My ‘unit, respectively. It is required to find the minimum-cost det shies eWay ehye8 Wx 2m <9 Fan an sovo fe enone poe Wis foe sreray ner aenana stare wa” ALEo mle weaiced 40. Acoma pocies tres pecs P, end Fo ar Te scummy, jaa:| CRE Oo eee een 7. One unit of product A contributes Ris. 7 and requires 3 units of raw: ‘3 units of B. A unit of product Q requires 2 units of B and 5 units of C ‘material and 2 hours of labour. One unit of product 8 contributes Fs. 5 and requires one unitof raw mateialand one hour ef labour ‘Availabilty of raw material at present is 48 units anc there are 40, hours of laour. fa) Formuate this problem asa linear programming preblem, (0) Write is dual ()_ Solve the cual by the simplex method ard find the optimal product mix and the shadow prices of the raw material and labour. and.one unitof product A requires 3 unis ofA, 2 urits ofB and 4 urs (fC. The company has 8 units of material A, 10 units of material B ‘and 1Sunis f material C avaliable ot Profs per unit of products P, Oand R ara Rs. 3, Rs 5 and Rs. 4. respectively (@), Formulate tis problem as an LP problem. (@) How many unite of each product should be produced to ‘maximize profit? (€), Wate the dua of his problem,1 13, A diet corscious housewife wishes to ensure certain minimum Intake of vitamins A.B and C forthe family. The minimum daily (quantity) needs of the vtamns A, B and C for the family are respectively 30, 20 and 16 units. For the supp ofthese minimum vitarin requirenents, the housewife relies on Wo fresh food | and I. Tho fire! one provides 7, § and 2 units of the three vitamin per gram respectively and the second ene provides 2,4 and unis of the same three vitamins per gram of the foodstut,respectiely. The fst fcodsttf costs Fs. 3 per gam and the second As. 2 per gram The protiem s now many grams of each fcodstut should the housewife buy everyday to keep her food bil as low as possble? (@) Formulate this problem as an LP model. (b) Write and then solve the dual problem. A menufacturing firm has discontinued procuction of a certain unpreltabla product ine and ths has crested consicerable excoss production capacity. The management i considering to devote ‘hie excose capacty to produce one or more of three produc 1, 2 and 3. The evallable excess capacity on the machines which right limit output, is summarized in the fllowing table Machine type Available Excess Capacity (Machine Hours per Wee!) Ming mactine 250 Lathe 150 Grinder 50 The number of machine hour for each unit of ho product ls given below. epective Capacity Requirement {Machine-hours per Unit) Machine Type Product Product2_—‘Product 3 ‘Miting mechine 8 2 a Latne 4 a 0 Grindor 2 0 1 ‘The per unit contribution would be Rs. 20, Rs. 6 ard As. 8, respectively for products 1, 2 and 3 (a) Formulate this proslem as an LP problem, (©) How much of each of the three products should the firm reduce and sellin order to maximizo the contibution? Determine the maximum contribution (©) Weta tho dual of tho above problem ard givs its economic interpretation Alarge disributng company buys cottee seeds trom four diferent plantations. On these plantations, the seeds are avalable only in a blend of two fypas A and B. The company wanis to market a blend consisting of 30 per cent of type A and 70 per cent of type B. The percentage of each type used by each plantation and the sling Prices per 10 kg ofthe blends of each plantation are as follows: Type Plantation Desired 1 2 8 4 a 40% 20% 60% 80% 30% 8 60% 80% 40% © 20% «70% Selling price perskg ASS As.2 As. 120Rs. 1.60 What quantity of coffee seeds should the company buy ftom each plantation so that the total mixture will contain the desired percentages of A and B and at fie same tme keep the purchasing ‘ost at a minimum? Aleo write he cial of the problem, ‘The XYZ Pastic Company has just received a covernment conract 1b produce three aiftrent plasic valves. These valves mus. be highly heat and pressure resistant and the company has developed 46. ”. Duality in Linear Programming, 2 three-stage production process that will provide the valves with tho necaseary peopartas involving workin Free diferantchambare, ‘Chamber 1 provides the necessary pressure resistance and can ‘procoss vaves for 1,200 minutes each wook. Chamber 2 provides heat resistance and can process valves fo" 900 minutes per week. Chamber 5 tosis the valves end can work 1,900 minutes per ‘week. The three valve types and the time in minutes required in each charrber ere Time Rlequireain Vawve Type Chimber1 Chamber 2 Chamber 9 a 5 7 4 8 3 2 19 ° 2 4 5 “The government wil buy all the valves that can be produced and the company wilreceive the following proft margins on each vaive: A, Rs, 19; 8, Ps. 13.80; and C, Fs. 10. How many valves of each type should the company produce each weekin order to maximize profs? Write the dual of fe given LP problem and give its economic interpretation? ‘The procurement manager of a company that manufactures spacial gasoline adives must determine the proper amounts of each raw material to purchase for the production of one ofits Froduets, Throoraw ratorials are avalablo. Each iro the finiohod Froduct must have atleast a combuston point o!222'F. Inadditon, the gamme content (which cause hydrocarbon polation) carnot exceed 6 par cent of volume. The zela content (whichis goods for ‘ceaning the internal moving parts of engines) must De atleast 12 per cent by volume. Each raw material contains three elements in ‘varying amounts, as shown in fie given table Faw material A costs Rs. 9.00 per ire, whereas raw materials B and C cost Rs, 6.00 and As. 7.50 per lire, spectively Raw Material A 8 c ‘Combustion pcint, 200 180-280 Gamma content, % 4 3 10 Zeta content, % a 10 8 ‘The procurement manager wishes to minimize the cost of raw ‘materials per lire of product. Use linear programming to find the ‘optimal preporions ef each raw material to use ina lite of the ‘ished preauct. Also wnt tne dual ef me given problem ana give its econome interpretation. ‘A medical scienist calms to have found a cue for the common Cold that consiss of three drugs called K, S and H. His results inicale that the minimum daily adult dosage foreffectiveteatent is 10 mg ef drug K, 6 mg of crug $ and 8 mg of drug H. Two substences are ready available for prepzring ols for distibuion to cold eutforore. Both substances contain all tree of the required tugs. Each unit of substance A contains 6 mg, 1 mg and 2 mg of drugs K, $ and H, respectively and each unit of substance B contains’ 2mg, mg and 2mg of the same drugs. Substance A {Costs AS. 3 per unit end substance B cosis Fis. 9 par ni (@) Find the least-cost combination of the two substances that will yield a pill designee to contain the minimum dally recommenced adult dosage, Suppose that the costs of the two substances are interctanged so that subsiance A costs As. 5 per unit and substance B costs Ris. 3 perunit Find the new optimal soluton. [Dei Univ, ME (HCA), 2602) ‘The XYZ company has the option of producing two products duting the period of slack actvy. For the next petied, production has been scheduled co thatthe milling machine is tree for 10 hours and slled labour wil have & hours of time avaliable. 0)E14 Oporations Rosoarch: Thoory anc Applicatone: 10. 19. Product Machine Time Skid Labour Profit per Unit — Time per Unit Contibution er Unit(s. A 4 2 5 5 2 2 3 SoWve the primal and dual LP poblens and bing ot the fact hat ito opfmum soliton of ore can be sblaned tom the other. A180 ‘explain in the cont of ho exam, what you understand by ‘hadow pcos (or dual prices er marina vate) of oscuro. [uae Uni, MBA 1206) A company produces tives products P, Qand A whose pices par tnt ae 9,3 and 4 respectively, On unt of product P requires 2 Units cfm, nd @ uns of mA tn of prt Orequres 2 unis of im and 5 unt cl m, and eno ut of prods Rroqures 3 units of mm, 2 Unt of 2 4 uns of my The company has 8 unis of ioral, 10 ints of material mand 15 units of meer m, Bralane tot (a) Formulate the problem as on LP medal {b) How many unis of each product should be prosuces to mreniize revenue? (c) Write the cual problem. ‘A person consumes two types offood A and B everyday to obtan 8 Units of protsins, 12 units ol carbohycates and 9 units of fats which Is his dally minimum requirements. 1 kg of food Acontans 2, 6 and 1 Unts of probin, carbonates and fas, respectively. 1 kg of food B ‘contains 1, 1 and 3 units of proteins, carbohydrates and fats, respectively. Food A costs Rs. 850 per kg, wha B costs Rs. 4 per kg. Determine how many kg of each food should he buy dail to Iminimze Fis cos of food and stil meet the minimum requiremets. Formulas on LP problem mathematically. Write its dual and soWve he dual by the simglex method. [Gujarat Univ. MBA, 1896) A fm produces three artis X,Y; Zat a total cost of Ri. 4, Ps. 3, and Rs. 8 per lem respectively. Total number of X and Z tiem 2 produced shouldbe atleas! 2 and number of Y and Z together be ‘atleast 5, The fer wants to minimize tha cost. Formulate this problem as an LP problem Wrie its dual Solve the dual by fe ‘Simplex method. Can you point out the solution of the primal problem? Ifyes, what isi? ‘Atirm produces mee types of tisculs A, 8 anc C: i packs them inassertments of wo sizes| and I. The size I contains 20 biscuts of type A, 6) of ype B and 10 of type C. The size I contains 10, biscuits of type A, 80 of type Band 60 of type C. A buyer inondsto buy atleast 120 tisculs of type A, 740 of type Band 240 oftype C. Determine the least number of packets he should buy. Wie te dual LP problem and interpret your answer Consider the following product mix problem: Let x, doncte ‘umber of units of Product 1 to be produced dally and x, the ‘numberof units ct Product 210 be produced daly. The production. ‘of Product 1 requires one hour of precessing time in department D,. Production of 1 unit of Product 2 requires 2 hours of processing time in department D, and one hour in department ,, The number of hours avalable in department D, are 32 hours and in department D. 8 hours. The contribution of one Unit of Product 1 is Fs. 260 and of Product 2 is Rs. 200. ‘The solution to thie LP model ie givon below: gq _200 300 Basic Sottion| mG Varies Vaues 200 x eft 2 1 0 0 5 8 o 1 o Ff Z= 6400 @-% [10 200 0 6400 ‘Given the dual oF he primal model, Obtain the optinam solution to the dual LP model from the above table. interpret the dual variables. HINTS AND ANSWERS Min Z,= 39,449, +546 subject to (i) $y,— 23+ 59 ~ 39422 Gi) 6+ yy S95 ~ 3yg2S Gi) (i) 2x, +3yS8 Gi) Sxj+2x,+25<10 | Gi) 55, 14515, and AytyH2 0 Dual Min Z~ 8), + 10y3 + 15y3 subject to) 2y,+5y,23 Gi) 3y,# 29,4 4y,24 and YI W20. 11. Primal Gi) +5) 25 | 1x, and.x,= units of fresh food I and Ito be bought, respectively Min 2=3x, + 2x, subject () 724,230) Sx, + 4x,220 Gi) 2x, + 8,216 and xpme 0. Dual _yyeJp and ys ~worth perunitof vitamins A, B and C, espe 10 the bady. Max Z=30y, +20), + 16y, subject to) Ty, +42 t2y5 83 and Yin ar¥9 20, (b) The optimal solution to the dual problem is: y, = 20/52, GH 2y F494 8 <2 12, Product | = O unit: Product 2= 50 units: Produet 3 and Max profit = Rs. 700, 13, Primal Xp ty and ry sO uits | quantities of coffee seeds, respectively which te company buys from exch plantation. Min 35+ 2m +1204 154 subject io () Ody, 024, + O64, 08,2 03 | (a) 06x, + 0.8, + 04x, + 0.22 0.7 | sod remnine | Duality in Linear Programming Tag Dual 1p; and y,= werth of coffee seods A and B, respectively Max Z,~ 0.33, * 07), subject to) Oy, +0643 ii) 0.29, + 08,2 ii) 0.6y, +04y,< 12 (iv) 08y, +029) 51.5 and 9220 14. Primal xj.% and.y.= number of valves of the types A, B and C, respectively to be produced. Max Z,= 13%; + 13.5%, * 10x, subject to) $x, 3x +2xy€ 1,200 (Ix, + 2xj 44x, 900 (ii) 4x, + 10x, + 5x5 1,300 and 0. Ams. x, =22,400/217, x9 = 5800/62, x5 = 0 and Max Z,= Rs. 58,125/31 Dual Yd} And y,= Work of Chambers |, 2 and 3 production capacity, respectively Min Z, = 1,200y, +900), + 1,300, subject to () $9, +%2-+ 4yy215 () 3y, #29) #1052 135 Gi) 2444+ Syy210 and Wn ZO. Ans y,=0,9)=— 4135 =~ 645/62. 17. Primal 2X). = number of units of product A and B respectively to be produced Max Z,=5r, + 3x, subjectto ()) 4 +25) 10 ) 24258 and 2320 Ans. x, and Max 2, = 14 Dual Yj. ¥5 = wort (o shadow price) per unit of resource operators in three shits respectively Min Z, = 109, + 8, subjectto (4, +2,25 i) 2429523 and FoIrz 0. Ans.y, = 1,y2 1/2 and Min Z, = 14 18. Primal 2,44 = numberof units of product P,Q and R wo be produced, respectively. Max Z,=31) * 549+ 4 subject to () 24 +3ry 58 (i) Sx, 42x +2510 (ii) 5x, +4, 515 and apa EO Ans. x, =4/41, x5 = 59141, x, = 80/41 Dual Yip Np Jy = Worth per unit of material m,, my and ms respectively,Operations Research: Theory and Applications Min Z, = 8), + 10) + (5y, Max Z,=2;+% anecte ) B+m23 MD Anrsy2s — | sabato 21042120 5Oqje 80420 y+, 44,24 (i). 1, + 60552 240 and, Vedra 20. and p20 19, Prinat | Ans, =2,.)= Band Min Z, = 10 xj.fy~ aumber of unis of food A and B to be consumed, Dual respectively yi3bs93 = worth per unit of biscuits of type A. B and C. Mas Z,= 8.505, +4, | respectively subject (i) 2x, +28 Gi) 6, +212 MinZ, = 120y, + 740, + 240 Gi) 54329 subject to) 2, +509) 107551 and Gi) 10 +809, + 609551 Ans 5" 1.8 = and ds20 Duel 22, Primal Yin Ys Worth per unit of proteins, carbohydrates and fats Max Z,~ 200s, + 300% respectively. subject to (i) x, + 2x, $32 Gi) Oxytx $8 Min Z, = 8, + 1273+ 95 and 20 subject © 6) 2%, +69, 495 $ 880 Gi) 9, 493 +3955 4 Dua and Yarnrs20 MinZ,~32y, + 8 Ans. y,=31/8,9,= 18,5 =O and Max Z, = 657%: subject 10 ( y5+04220 Gi) 2 +9, 2 300 20. Primal and Jn 3220 Yt. %)= mumber of units of artes X, ¥ and Z respectively ‘The optimum slution tothe dual LP problemis:y,= 200,)=05) Max Z_~ Airy + 3 + 6ry = Ox, = 100, 2 = Rs. 600 subject 9) x +4522 @ y+n25 ‘Note that, y= marginal increasein profit for an addition I hour and Hoty 20, of epacity. Since y, = Rs, 200, profit becomes Rs. 6400 + 200 Ans, ~ 0,2," 3.4 =2 with Min Z=21 = Rs. 6600. Also y3~ marginal increase in profit given an addition mat hou ofcepacity in D, Since y,=0, there would be no change in ss the profit Ya Ja = worth per unit of resources to be used. respectively {In primal LP problem solution, s, = 8 indicates that there is an Min Z, = 2y, + 535 excess capacity of 8 hous in D. Thus, a dual solution we have y=. subject io () yi <4 (i) p83 ual constraint (i) is writen as: y, + Oy, 2 200 or Dy tHS 6 total value of hours in department D and D; and yu3gS 0 required to produce | nit of Product | Ans, =3,¥)= 3 with Max 2" =21 unt profit received from manufacturing 2 Pa (ming enttPeaet | xj.32~ ntmbor of assortments ofsize [and I, respectivel CHAPTER SUMMARY In this chapter we discussed the relationship between an LP problem and its dual. The method of formulating a dual LP problem of the given LP problem was explained with several examples. The main focus, while solving a dual LP problem, was to find for each resouree ils best marginal value, also known as dual or shadow price. This vaiue reflects the maximam additional price to be paid in order to cbtain one additional unit of any resource in order to maximize profit (or minimize cost) under resource constraint, CHAPTER CONCEPTS QUIZ True or False 1. Ifthe objective function of the primal LP problem is maximized | 3. The dual of dual is a primal probiem. then the objective functon of the dual is to be minimized. |, The value in the ¢ ~ z) fow under columns of slack/sumplus 2. A dual variabe is not defined for each constraint in the prima variables ignoring negative signs do net give the direct optimal LP problem, values of tne duavormal basic variables.image not availableBQ.) Operations Research: Theory and Applications 4 1" 16, 19. a 92, TO o8F OT AF 6&F 6T 72T 8T OT WT Primal, dual 42, Dual 13. Simplex 14. Untounced; infeasible 18. Equality; unrestricted Complementary slackness; equal 417. Shadow price 18. Slack, basic Decision variables; constrainis; objective function 20. Dual primal; minimization (©). (0) 23. (a) 24. (a) 25. (&) 28. (c) 27. fa) 2B.) 2B. (co). (@)_— 31 CC) @) 3B @) 34) 35. (a) APPENDIX: THEOREMS ON DUALITY Theorem 5.1: The dual of the dual is the primal Proof: Consider the primal problem in canonical form: Minimize Z, = ex subject to Ax2b; 120 ® where A is an m xn matrix, b’ © E” and ¢, x” € E". Applying the transformation rules, the dual of this problem is: fa Maximize Z, = by subject to Aly sels y20 ® Tais dual problem can also be written as: ie Zt= (-W)"y 8% subject to, CATy 2(- €); y20, where 2; =~ If we consider LP problem (3) as primal, then its dual can be constructed by considering x as the dual variable. Thus, we have Maximize Z* = (- "7x = (ox subject to CAN x <(-b7) or A) s (-b) and x20 ® But LP problem (4) is identical to the given primal LP problem (1). This completes the proof of the theorem Theorem 5.2: Let x* be any feasible solution to the primal LP problem Maximize Z, = ex subject to Ax Sb; x20 and y* be any feasible solution to the dual LP problem Maximize Z, = b"y subject to ATy 2e7 y>0 of the above primal problem. Then prove that ex* ; y unrestricted in sign. Theorem 5.5 (Unboundedness Theorem): If either the primal or the dual LP problem has an unbounded objective function value, then the other problem has no feasible solution. Proof: Let the given primal LP problem have an unbounded solution, Then for any value of the objective function, say + e, there exists a feasible solution say x yielding this solution, i.e. ex — =. From Theorem 5.2, for each fessible solution y* of the dual, there exists a feasible solution x* to the primal such that ex* < b”y*, That is, b’y* > + =. As b is a constant number and y* has to satisfy the constraint A7y* < e7. Hence the dual objective function Z, = by* must be finite. This contradicts the result b’y* — ©, Hence the dual LP problem has no feasible solution, A similar argument can be used to show that when the dual LP problem has an unbounded solution, the primal LP problem has no solution. ‘Theorem 5.6 (Duality Theorem): Ifeither the primal or the dual problem has a finite optimal solution, then the other one also possess the same, and the optimal values of the objective functions of the two problems are equal, Max Z, = Min Z,, Proof: Let the following LP problem represent a primal problem: Maximize Z—ex subject to Ax Sb or Ax+Is=b: x20 ‘where T is the identify matrix of order m and s is the slack variable. Let x, = B"'b be an optimal besic feasible solution to this primal problem and ¢, the cost vector ‘of the basic variables. Then according to the optimality condition, we have, ¢,~ z, < 0, for any vector a, of A but not in B, Therefore, ~e,(B'a) <0, forall)EEE Operations Research: Thoory and Applications: This is equivalent to € < ¢,B™'A, when written in metrix notation. Let y* = ¢,B', Then ¢ < ¢gB™'A becomes ¢ < y*A orc! < y* A!. Hence y* is a feasible solution to the dual of the given primal because it satisfies the dual constraints. The corresponding dual objective function is given by Z,= y*b" = ¢,B"'b? = cy3, =Z, Hence, it has been shown that x, and ¢,B"? are the feasible solution to the primal and dual LP problems, respectively for which Max Z, = Min Z, Theorem 5.7 (Complementary Slackness Theorem): Ifx* and y* be feasible solutions to the primal and dual LP problems, respectively, then a necessary and sufficient condition for x* and y* to be optimal solutions to their respective problems is, By %yei2O TL Zee and mej FQ where, ., is the ih slack variable in the primal LP problem and y,,., the jth surplus variable for the dual LP problem, Proof: Let x* be the feasible solution to the primal LP problem, Maximize Z, 3 gs, pale sunjetto B ayay tage “85 4/20 o and y* be the feasible solution to the dual of the above primal. Minimize Z,= £ 4,y, ® a subject © Bays nay “G5 EO z ‘Multiplying constraints in (7) by yj, i= 1, 2, .... m, and then adding we get 5 Lagat ani $3 y,5,5 f= 12am © ‘Subtracting (9) from the objective function of (7), we get Substituting the value of Z, and y,,..; from (8) to (10) we get, ~ Barnes Erte Z-% ay If x* and y* are optimal solutions to the primal and dual problems, then we have ex* = b’y* (or Z, =Z,) Thus, from (11), we have Haw italy rete Blo eri pp ead Sa 5 aa jaca in the dual is an equation. Also x, = 0 implies y,,,=0and 3 a,y, ><, 7 nth Rowers Silty foreach > 0,46 12s...) implies, =Oand £ ays, iF 1» ke. fh constraint in the primal is an equation. Also y, 0 implies x,,,>0 and ¥ ayx, mix (Gf no ratio is positive)} iginal value, ey Per unit improvernent value Input — Output coefficient in the variablerow ay where Improvement ratio = Remark While performing sensitivity analysis, the artificial variable columns in the simplex table are ignored. Any exchange of quantities between basic variables and an artificial variable makes no sense, because an artificial variable has no economic interpretation. Thus, improvement ratios using coefficients corresponding to artificial variables should not be considered. Gase lit Change in coefficient of a non-basic variable in a cost minimization problem ‘The procedure for calculating sensitivity limi of a cost minimization LP problem, where the objective function coefficients are unit costs is identical to the Case I discussed above. In this case, the unit cost coefficientBEL Operations Research: Theory and Applications ‘can be increased to any arbitrary level but it cannot be decreased by more than the per unit improvement value, ‘without making it eligible so that a non-basic variable can be entered into the new solution mix. The sensitivity limits can be calculated as: riginal value c, - Absolute value of per unit improvement value (c) ~z)) Infinity (+ =) Alternative method Ifwe add Ac, to the objective function coefficient ¢, of the basic variable xj, then its coefficient cg, will become cj, = Cp, + Acgy- The new ¢, ~ 2% can be calculated as follows: {3 ay + (Can + Sean) Ye, isk Leg y = i me bey + Meme yy) = (6 ~ 2) ~ Bee ry For a carrent basic feasible solution to remain optimal for a maximization LP problem, we must have $0. That is, G- FG 2)~ Mery $0 oF 6-35 Dew ry (2S Aen when yy > 0 (6-2)i4j 2 cy when yy <0 Hence, the value of Acpy that satisfies the optimality erierion can be determined by solving the following system of linear inequalities: Min] 27 b> deg, 2 Max — 5 is a " at 2 Here it may be noted that y, sare entries in the non-basic variable columns (i.e. variables) of the optimal simplex table. The new value of the objective function becomes: Denko + (ae +m a4 = 2 coetsi + Seo tee = Z+ Sep, Soy Zz Hence if dcp, satisfies inequality (1), then the optimal solution will remain unchanged but the value ‘of Z will improve by an amount Aci, Xgy Example6.1. Solve the following LP problem: Maximize Z=3r, +51, subject to the constraints @ 3yt2_Z18, GD mZ4, Gi) Zw and xyty20. (@)_ Determine an optimal sclution to the LP problem. (b) Discuss the change in con the optimality of the optimal basic feasible solution. ‘Solution (a)The optimal solution ofthe given LP problem is shown in Table 6.1. From Table6. 1, the optimal solution can beread as: and Max 2 = 38, (b) Here c= (cy, ¢>, ¢s Cys Ce), and the cost coefficients associated with basic variables x,, x», and, are 4" (61, ¢p.¢4) = G, 5,0). The ckanges inc, can be classified as under: () Changes inthe coeffictents (he. and es of non-basic variables, and sy Lets adc, and Acy tothe objective function coeficents and, Thenthe new objective function coeficient will become c3 = + Aes and cs = c,+ Ac,, Since c, =0 and c5=0, cf = Ac, and cj = Acs. Thus, the new values ofc —z5 and z,will become Ac, ~ I and Ac, ~3, respectively. Now in orderto maintain optimality, we musthave ‘Bcy—150 and Ac,—350(Maximizationcase) or cy and Acg<3Sensitivty Analysis in Linear Progamming BRAg @ Change in he coeffciems cj (.e.¢y cand, of base variables x, and sy: The value of additional increment cy, t= 1,2, 4in the coefHcients, ¢,andc, which satisfy the optimality condition can be determined by solvingthe following system oflinearinequaties refer to Eqn, (1)]- 74 we 5° 0 0 2 x 2 1 0 18 0 23 ° a 0 0 o 23 1 a3 8 % 6 0 1 o 0 1 Table 64 Z=36 o=% 0 0 a 0 Optimal Solution Fork=1 (je. basic variable x, in row 1), we have =3 a1 win { $3 fen = man (=, G35 or 922hq23 Here it may be noted that y,,= (43 = 1/3, y,5=—2/3) are only for those columns corresponding to which ‘variables are not in the optimal basis (ie. non-basic variables}. The current optimal solution will not change as long as 9 15 (2#2}ae20-3 ot 2e,20 Fork=3 (ie. basic variable x, in row 3), we have = 3 sacl neal flocs wean Hence, the current solution will not change as long Example 6.2 A company wants to produce three products: A, B and C. The per unit profit on these products is Rs 4, Rs 6 and Rs 2, respectively. These products require two types of resources, manpower and raw material. The LP model formulated for determining the optimal product mix is as follows: Maximize Z = 4x, + 6x, + 2r, subject to the constraints © x, +x, +4, 53 (Manpower required), (ii) x, + 4x) + 7x, $ 9 (Raw material available) and Spay 0 where x) x, and x, = number of units of products A, B and C, respectively, to be produced. (® Find the optimal product mix and the corresponding profit of the company. () Find the range of the profit contribution of product C (ie. coefficient c, of variable x,) in the objective function, such that the current optimal product mix remains unchanged. (© What shall be the new optimal product mix when per unit profit ftom product C is increased from Rs 2 to Rs 10? (© Find the range of the profit contribution of product A (i. coefficient ¢, of variable x,) in the objective function such that the current optimal product mix remains unchanged. Solution (a) Convert the given LP mode! into the standard form by introducing the slack variables 5, ands. Maximize Z = 4x, + 6x, + 2r, +05,+ 0s, subject to the constraints Ox tmtytsy—3 Gx +4yt Byte and Xp Xp Ty Sy 520 ‘The optimal solution obtained by applying the simplex method is shown in Table 6.2 5: (S +09) 2 2 (5-3) ore, 22,"> 4 6 2 o o 4 7 1 1 0 =I 43 13 6 % 2 0 1 2 sia 1B Table 6.2 z= 16 4 6 8 108 Ey Optimal Solution 9, 0 ° =6 103-28. ‘The optimal sotution is: x, = 1, 1) = 2 and Max Z= Rs 16, (b) Effect of change in the coefficient c, of non-basie variable x, for product C In optimal simplex Table 6.2, the variable x, is non-basic and its coefficient c, ~ 2 is not listed in the eg column of the table, This ‘means a further decrease in its profit contribution c, (= Rs 2) per unit will have no effect on the current optimal product mix. But, if cy is increased beyond a certain value, the product may become profitable 10 produce. Hence, there is only an upper limit on c, for which the current optimal product mix will be effected. 97 4 6 2746 ° ° ay a 1 -l Taga 6 a 2) 0 1 2 “13 Table 6.3 Table with a Acy y 4 6 103 9-5 6 o 108 Table 6.4 ‘As wealready know, a change Ac, inc will cause a change in z, and c3 ~ 3 values in the x-column, ‘The change in cy results in the modified simplex Table 6.3. For an optimal solution shown in Table 6.3 to remain unchanged, we must have: Ac, ~ 6 $0 or ey 5 6. a dep ils Regret abi il con't emi wwe get c; - 2 £6 of cy <8. This implies that as long as the profit contribution per unit of product C is less than Rs 8 (i.e. change should not be more than Rs 8) it is not profitable to produce it and therefore the current optimal solution will remain unchanged, (©) If the value of ¢, is increased from Rs 2 10 Rs 10, the new value of ¢, ~ 23 = (€,~ €s) will be =i gos-10-6 6[ 4]or-c4e1y-260 ‘Thus, if coefficient of variable x; is increased from Rs 2to Rs 10, the value of ¢, ~2, will become positive as shown in Table 6.4. The variable x, (corresponding to product C) becomes eligible to enter into the basis. Hence, the solution cannot remain optimal any more. 57 4 6 wo 0 o an low 2] 0 Copyrighte| Sensitivity Analysis in Linear Programming BRED Applying the following row operations to enter variable x, into the new solution mix R, (new) —> R, (Old)/2 (Key elememi) ; Ry (new) > R, (old) + R, (new) The new optimal solution after entering non-basic variable x, into the solution is shown in Table 6.5. 1 2-18 y g=% Table 65 Since all ¢/~ 4; $0 in Table 6.5, the new optimal solution obtained is: x ~ 2, xy ~ Oand.x3 ~ 1 and Max Z=Rs 18. (a) Effect of change in the coefficient c, of basic variable x, for product A_ In the optimel simplex Table 62, the variable x appears in ‘Basis B’. This means a further decrease in its profit contribution c, Rs 4) will make it less profitable to produce and therefore the current optimal product mix willbe affected. Also, an increase in the value of c,, beyond a certain limit, will make the product A much more profitable and may force the decision-maker to decide to only produce product A. Thus, in either case, the current optimal product mix will be affected and hence we need to know both the lower as well asthe upper litt, ‘on the value of ¢,, within which the optimal solution will not be affected. Referring again to Table 6.2, the range of change in the value of ¢, (and/or also in ¢,) that does not affect the current optimal product mix cen be determined once again by calculating the values of ¢; ~ zj, ‘values that correspond to non-basic variables x, s, and 3» respectively. For this reproduce Table 6.1 once again with unknown velue of c, as shown in Table 66. go thse, 6 2 ° o athe, 1 1 0 1 43 -13 6 a 2 o 1 2 “13 re Table 66 Z=12+@ + de) = | ata 6) BO, «WAT A4Ac 3 23 —Ac/s Table with a Acy ons | 0 0 By 6 103 4ey3__bey/3—-23_ change in c For the solution shown in Table 6.6 to remain optimal we must have all ¢, ~ z; $ 0. That is Ae 6 £0 gives Ao, <6 acy -Z- Eso gives Ac 2-52 dy 2 , 3 SF <0 ives dey 52 Thus, the range of values within which c, may change without affecting the current optimal solution is: 3 3 2.8 be 5 She, $2 or 4-5 From these calculations, it may be concluded that in the case of competitive pressure or in the absence ‘of any competition, the decision-maker knows to what extent the prices can be adjusted without changing, the optimal product mix. 3 S442, ie. 5 $056 Copyright erial4 Opsrations Research: Theory and Applications Example 6.3 Given the following LP problem Maximize Z= x, +2e)—x, subject to the constraints @3x4x)-x,S10, (i) —1, +4, 4,26, Gill) xy+x, $4 and By kyxy20. Determine the effect of discrete changesin c)(j=1, 2,.. 6) onthe optimal basic feasible solution shown in Table67, 6° 2 4 0 0 0 -~M ° 4 6 Bc qi renege) Saji 20 Fo Memes oe 2 xy 4 Ue NE ee te age Sat 0 o 3 10 Libeat Caestinaedane tena Ae onae Table 6.7 Z8 2 z Or seieme eat TO aren ON ‘Optimal Solution “4 SEO Ca, 4s eal Solution In Table6.7, variable, isa non-basic with c= -1. If, further decreases, x, will not enter the basis, and hence will nat afect the optimality ofthe solution. Thus there is no lower limit onthe valueofc,. However, if, inereases and exceeds a certain value, it may become profitable to have.x, in the basis. Thus, there should ‘bean upper limit on the value ofc. Solution shown in Table 6.7 will remain optimal provided largest value of @ <0, ic. 3 €,-.2,0)]0/<0 or eS0. 1 ‘The variablex, in Table 6.7is abasic variable, the solution will remain optimal so long as nor-basic variables remain non-positive, ie, 3 =10,6,,0)/0|S0 or <0 1 ] 2 €5=-1-@.c,,0)] 1 |s0 or 2-1 3 =I = 0-(,6,,0)| 1]s0 ot e520 4 0 M~(0,¢,,0)| 0|<0 or MZ0 -1 is enon-asie variable, and to find its upper limit, we musthave c, $0, i. a €5-,2,0)| 1]50 or ey-250 or e552, 3 ‘The variables; in Table 6.7 is abasic variable andi limiting value is obtained from largest value ©),23,% and &; of cy, ¢4,¢4 and c, respectively as follows: ‘The variablex, in Table 6. ; 7‘Sensitivity Analysis in Linear Programming : -1-¢40)| 1 |s0 1425-28005? [s 7 ceyain | aleumene aeons 3 o 3 =-M-~(cy.2,0) | 0/0 or - M0, which is tue, ft Hence, limits for ¢, are: ‘The variable s, in Table 6.7 isa basic variable and its limiting value is obtained from largest value Z, 2), 2% 1B $e $32 and & of ¢, cy and c, respectively as follows: F=-1- 0,2) |0]s0 or $0 or c2-1, li 2 B=-1-0,2c)| 1)s0 or -1-(2+3c,) <0 or cg2-1, 3 % =0-(0.2,6)| 1]s0 or 0-@+4e,)<0 oF % =-M-(,2,¢) | 0/s0 or -M+e5<0 or cg SM. Hence, limits for cg are:~1/2- bby = Max bi sf * tial 2. The range of variation in the availability of resources (b,), can also be obtained by using condition of feasibility ofthe current optimal solution, ie. x3 =B~'b > 0, where B~! = matrix of coefficients corresponding to slack variables in the optimal simplex table ‘Abj = amount of change in the resource k Xp = basic variables appearing in B-colunn of simplex table Remarks 1. If one or more entries in the xp-column of the simplex table are negative, the dual simplex ‘method can be used to get an optimal solution to the new problem by maintaining feasibility 2 A resource whose shadow price is higher in comparison to others, should be increased first to ensure the best marginal increese in the objective function value Example 6.4. Soive the following LP problem Maximize Z = 4x, + 649+ 2x5 subject to the constraints @ytytySh Gx F+4yt+ IQ s9 and yy 20 (@ Discuss the effect of discrete change in the availability of resources from [3, 9]" to [9, 6)’. (©) Which resource should be increased (or decreased) in order to get the best marginal inerease in the value of the objective function? [AMIE, 2004) Solution The given LP problem in its standard form can be stated as follows: Maximize Z = 4x, + 6x, + 2r, + 0.5, + 0.5 subject to the constraints OxtHtHts, and Ryn y By (i) x, +44 + Tr +5 =9Operations Research: Theory and Applications Applying the simplex method, the optimal solution so obtained is shown in Table 69. The optimal solution is: x, = 1, x, = 2 and Max Z = 16. (@) If the new values of the right-hand side constants in the constraints are (9, 6 ]’, then the new values of the basic variables (xg = B-' b) shown in Table 6.9 will now become: lL: elf) « s-maes 67 ‘ ‘ 2 o o 4 5 1 1 o See I 6 es z ° U Bo IB Table 6.9 Z=16 : = 4 6 8 103 8. Optimal” Solution vee ° 0 26) S10 2) Since the value of x, is negative, the optimal solution shown in Table 6.9 will become infeasible. To remove infeasibility, we use the dual simplex method. By reproducing Table 6.9 with new values we get the values shown in Table 6.10, ES Table 6.10 Since x, ~~ 1, the second row is the key row and x, is the outgoing variable, For identifying the key ‘column find the following ratios: =f) neat = Spf 719 columa sy) Hence column ‘s,"is the key column and variable, will enter into the basis. Now revise the simplex table ‘using the following row operations: Ry (new) > R, (Old) *-3 Ry (new) > R, (old) ~ (4/3) Ry (new) The new solution is shown in Table 6.11 As all ¢) ~z) $0 and all 6, > 0, the solution given in Table 6.11 is optimal. The optimal solution is: 21 = 5, 4) = 0, X = 0 and Max Z = 24. In Table 6.11, slack variable s,, is present at positive level in the solution mix. Thus, the sensitivity limits for the corresponding right-hand side, ic. first constraint, are determined 2s follows: Lowerlimt =9-3=6 Upper limit‘Senstivity Analysis in Linear Programming Eg o> 4 6 2 oo 4 4 6 1 4 ° 5 3 RE 2-34 ae 4 16 Table 6.11 RS See Ee rapper Oi Cote S268 Te een Optimal Solution () In Table 6.10, cg —=4 = 10/8 and 0, ~ 25 ~ 2/3 valuee (ignoring negative sign) comeeponding to 55, and s columns represent shadow prices of resources 1 and 2 respectively. Thus increasing the amount of resources 1 and 2 will increase the value of objective function Z by Rs 103 and Rs 2/3 respectively. Now in order to know how much these resources may be increased so that each additional unit continues to increase the objective function by 10/3 and 2/3 respectively, till the optimal solution remains feasible, we proceed as follows. Let Ab, be an increase in fist resource (RHS of first constraint) so that Xy=BTb20 [i | “ [ 43 El Is] era ps 6 or »| [43 2 2-(13)Ab, ie, 1+ (43)Ab, 20 or Ab, 2-3/4 and 2 ~ (1/3)Ab, 2 0 oF Ab, $3 From these inequalities, we have: -3/4 Ab, 53 or 3-(/4)S,534+6 of 94R, (old) x-5 Ry (new) -¥R, (old)—(7/5) R, (new) ‘The new solution is showa in Table 6.15. Since b, = 3/5, the fist row isthe key row and.x, is the outgoing variable. To identify the key column, find the following ratios: Minf(c,z)/,, (< 0)} = . Hence ‘x,'-column becomes the key column and variable.x, will enterinto the basis. Revising the simplex table using suitable row operations. The new solution is shown in Table 6.15. ‘opyrightSensitivity Analysis in Linear Programming Eg an 4 6 0 5 a She es 1 70 1 aaa 1 Table 6.18 ae: ere een Optimal Solution Since all¢,~z; < 0 and b, 2 On Table 6.15, the solution is optimal. The optimal solution sy 3, Mak Z 12. (@ To find the resource that should be increased (or decreased), write the dual objective function with the help of given LP model as 2, = Sv, +2y, where, y,=29/5 an v,=2/5 are the optimal duel variables (Table 6.13). Since value of y, is higher than the value of y,, the first resource should be increased as each additional unit ofthe frst resource increases the objective function by 29/5. Further to find how much the firstresource should be inereased sothat each additional unit continues to increase the objective function by 29/5. This requirement will be met so long as the primal problem remains feasible. Let A 6, be the ‘increase in the first resource, so that 22] pipe [25 -5] [544m] _ [os +200/5-25]_rees2aqys), 3 vs 2s|| 2 sis+anis+4is |~| oranys ie, 8+4b,200rA 2-8 and 9446, 20 oF 442-9 incex, and x, remain feasible (20) forall values ofA 6, 2 0, the first resource can be increased indefinitely while maintaining the condition that each additional unit will inerease the objective function by 29/5, ‘The second resource should be decreased as each additional unit of the second resource decreases the objective function by 2/5. Let A b, be the decrease in the second resource, so that 2] pty [25 US] 5 ]_flos-2is+ab/syy e+any5). x us 15 ||2-46, |” [5'5+4/5-284/5 |] (9-205, ¥5]~ ‘Thus. remin positive nly sn longan9—24 b,2or A by 87 1EAb> 97. becomes agave sndimt leave the solution. Example 6.6 A factory manufsctures three products 4, B and C for which the datas given in the table below. Find the optimal product mix ifthe profitunitis Rs. 32, Rs. 30 and Rs. 40 for product 4, B and Crespectively, Product ~ tvalable Resources 4 r e Materia required (kg/unit) 3 4 3 2,500 kg Machine hours equired/unit = 2 3 1 1.275 hows Labour hours require/anit 3 2 4 2,100 hours (8), Find the optimal solution if machine hours available become 1,380 instead of 1,275, (b) Find the optimal solution iflabour hours become 2,000 instead of 2,100. (©) Find the optimal solution if 10 units of product 4 are to be produced. Solution Let us define the following decision variables: xy, and x,= Number of units of product 4, B and C to be produced, respectively. ‘Then the LP model based on problem data is written iBEE] Operations Research: Theory and Applications Table 6.16 Iniial Solution ‘Table 6.17 ‘Optimal Solution Maximize 2-32, +30x) + 40x, subject to constraints @ Sx, +44,+3x52,500, (2x,+3 +x, 1275, (ii) 3x, +2 +44 $2,100 and ay20. Initial basic feasible solution: x,=0,, 275,55" 2,100 and Max Z=0 is shown in Table 6.16 ° 5 2500 Selo ia geet aro no aa ° 4 1215 Beene Oe is 0 as 2100 3 Bes 0. 0 1 | 21004 > z=0 y Tge USA ear 20 | Ce se ee ; ‘The initial solution shown in Table 6.16 is updated to obtain the optimal solution shownin Table 6.17 using suitable row operations: 5 2 ww wo 0 Oo o 0 4 12s Bie ore S09 Shh ay ay 30 = 300 Bee ath. 30) lo BieraomriG 40 oh a5 BBS 0s ae AO 2-24, 000 St gia enes 9 E ES UeUe. anon wen ee! ‘The optimal solution shown in Table 6.17 0, x, = 300 units, x, = 375 units and Max Z = Rs. 24,000. Since s, = 0 and s, = 0, this implies this while producing product B and C all the machine hours, and labour hours available are consumed, However, , = 125 kg implies this only 125 kg of material remains ‘umutilized. (@ In Table 6.17, sy-column shown that if one machine hour is increased, 1 kg of material will remain ‘unconsumed and product B will increase by 2/5 unit and C will reduce by 1/5 unit. Thus, if machine hours available are increased by 75, the product B (= variable x,) will increase by (2/5) = = 30 units and product C (= variable x,) will decrease by (1/5) x 75 = 15 units. Thus the optimal sclution will be: 330, x,= 360. Further, fone machine hour remains unused, there isa loss of Rs, 4. In other words, for every additional machine hour used, there is a gain of Rs. 4 and hence profit will increase by Rs. 75 4 = Rs. 300 to attain a value of Rs. 24,300, which is equivalent to the new solution value: Rs. (0 x 125 + 30 x 330 + 40 x 360) = Rs. 24,300, (®) In Table 6.17 sy-column shows that if one labour hour is decreased, it would change the value of x, by— 1x (1/10) = (1/10) units and of x, by ~ 1 (3/10) =— (3/10) units. Therefore, a reduction of 100 labour hours will increase value of variable x, by (1/10) * 100 = 10 units and decrease value of variable, x, by (3/10) = 100 ~ 30 units and the new optimal solution would be: x, ~ 0, x ~ 310 units, 24 = 345 units and Max Z= Rs. 23,100. (© Im Table 6.17 x,-column shows that when one unit of product 4 is produced, it would reduce the production of B and C each by 1/2 unit each. Therefore, if x, ~ 10, x, ~ 300 (1/2) » 10 ~ 295, x= 375 — (1/2) x 10 = 370. The profit contribution of Rs. 3 by product A (coefficient of variable x,) will also reduces and hence it will reduced to Rs. 30 and become Rs, 23,970, which is equivalent to the new solution: Rs, (10 * 32 + 30 * 295 + 40 = 370) ~ Rs, 23,970,| Sensitivty Analysis in Linear Programming 6.2.3 Changes in the Input-Out Coefficients (a,'s) Suppose that the elements of coefficient matrix A are changed. Then two cases arise @ Change in a coefficient, when variable is a basic variable, and i) Change in a coefficient, when variable is a non-basic variable, Case I: When a non-basic column a, ¢ B changed to aj, the only effect of such change will be on the optimality condition, Thus the solution will remain optimal, if G7 tf =Gq- CyB taj SO ‘therwise the simplex method is continued, after coluran k of the simplex table is updated, by introducing the non-base variable 2, into the basis, However, the range for the discrete change Aa, in the coefficient of non-basic variable x, in the constraint, / can be detzrmined by solving following linear inequalities: Here Bi isthe ith column B1. If ¢yB,~0, then Aa, is unrestricted in sign Alternative method The change in the coefficiens (a,’s) values associated with non-basic variables {nthe optimal simplex table can be analysed by forming a Corresponding dual constraint from the original set of constraint: & ayy, 2c); forx, non-basic variable im “The value of dual variables y;’s can be obtained from the optimal simplex table. The reason behind this dual constraint formulation is that an activity is considered as fally undertaken provided shadow price (or imputed cost) of all resources needed to produce one unit of an activity become equal to its per unit contribution to total profit. Case Il: Suppose a basic variable column a, € B is changed to aj. Then conditions to maintain bath feasibility and optimaliy ofthe curent optimal solution ate ~* me * Bk (@ Max |__| < a, < Min } —_——## ___ te? [Xp Bmp Bu > 0 120 [Sp By — yp By <0 ee ene © ~2)By — sade te By —3y & | Example 6.7 Solve the following LP problem Maximize ® rs x, +3 - subject to the constraints © B-y_t%yS7 GD - 2, +4e,<12, Gil) — 4x, + 3x, + RK, S10 and xpXy 420. Discuss the effect of the following changes in the optimal solution. (@) Determine the range for discrete changes in the coefficients a, and ay consistent with the optimal solution of the given LP problem. (b) ‘x,'-column in the problem is changed from [3,~2, ~4 ]"to [3, 2,—4]". (©) ‘ry’-column in the problem is changed from (2, 0, 8 J" to [3, 1, 6]". [UME, 2005] Solution The given LP problem in its standard form can be stated as follows: Maximize Z =—x, + 31 — 2x, + 0.5, + 0, + 0.5; subject to the constraints @ 3r,-y +My +4=7 6 and Xp Xpe py Sp 8520 ny + bey + 52 = 12, Br, +5 = 10 189Operations Research: Theory and Applications Table 6.18 ‘Optimal Solution Applying the simplex (Big-f) method, the optimal solution so obtained is shown in Table 6.18. 4 “fe 0 6 o SE ae i 0 45 US MO SA aeay ° 1 mS MO Gore ° 0 10 ic Sear Zul ate ete ee 0 o oO -ms sas ‘The optimal basic feasible solution shown in Table 6.18 is: x, = 4, x, In table 6.18, the inverse of basis matrix, B is 25 M0 0 vs mo 0}= 1-121 5.x, = 0 and Max Z= 11 By, By Bs) Thus, we have eg B, =~ 12/5) +3(/5) + 0) = U5 epBy ~- 101/10) + 3/10) + 172) -8/10 ep By =~ 100) +300) + 0(1) =0 Since variables x,, rz and s, are in the basis (column B of Table 6.18), therefore any discrete change in coefficients, belonging to any of these column vectors may affect both the feasibility as well as optimality .of the original optimal basic feasible solution, whereas any discrete change in the non-basic variables (ie. 2Xy 5, and s,) column vectors may affect only the optimality condition. (@) Ranges for discrete change in coefficients a), and a;, in the x,-column vector of Table 6.18 are computed as: 205 Max = Max ayy oF Aayy > —12 {esa} Ne Fie] 1200, Gi) 3x, +40e, +H, +14 < 800, and Happy Hy X420. 10 Go 4 10 3 30 0 o as =1/130 “ns 25 nos sors Table 6.19 23 0 58 3B Optimal Solution (@) ‘x,'-column in theproblem changes from [7, 3]" to[7, SI". (b) ‘x,'-column changes from (7, 3]" to [5, 8)’. Solution (2) The variable x, is a non-basic variable in the optimal solution shown in Table 6.19. The upper limit &; the coefficient, of; iscalculated as follows: ‘ 7 1~€ nd =4—egB a, =, Fidy where = 45, a i : and 5 =cgB" 0,100 | 5, Pa ] - 3. es 154 10 ‘a I[t]-=-(te Since value of @, isnegetive, the current optimal solution remains optimal for the new problem also.Operations Research: Theory and Applications Table 6.20 Optimal Solution Or eg mey=egne,~ 4,-45-[ 2.2] ] 45- (428) ~ Since value of @; is positive, the current optimal solution can be improved. Also sore le eh 6 so o o 30 8003 40 73 ans -1s | 10008 100. le 4083 10-130 1502s | 20027 + 7= 28.00% 168 103011 MHOS S078 i 6 Aygo 88 ab > ld t Table 6.21 Optimal Solution Introducing non-basic variable x, into the basis to replace basic variable x, using suitable row operations: In “Table 6.21, all ¢,~2,$0, the solutions optimal with: x, 2000/27, x,=0,x,~ S6W0/27 and Max Z=86,000/9. go 0080 0 0 0 30 aS 0 409 1 om] a= aS 1090 a = 86, 00/9 4 3503585108 0-303 et em 109 6.2.4 Addition of a New Variable (Column) Let an extra variable x,,,, with coefficient ¢, , be added in the system of original constraint Ax ~ B, x20, This in tum creates an extra column a, ,, in the matrix A of coeffcients. To see the impact of this addition on the current optimal solution, we compute Yast and Geet Fat ‘Two cases of the maximization LP model may arise: (@ If c,,1~ 2,41 $0, then Xp = 0, and hence current solution remains optimal. (b) If). — Zpe1 > 0, then the current optimal solution can be improved upon by the introduction of a new column a, , , into the basis to find the new optimal solution. Example 6.9 Discuss the effect on optimality by adding a new variable tothe following LP problem with column coefficients (3, 3, 3)’ and coefficient 5 in the objective function. Minimize Z = 3x, + 8x, subject fo the constraints @ mty-200, Gi) 41580, Gil) 260 and 420 Solution The given LP problem in its standard form can be expressed as:Minimize Z = 3x, + 81, + 0s, + 0s,+ MA, + MA, subject to the constraints © AFH +4, =200, Gi) x, +5,=80, Gill) ys + 4,= 60 and Xp psp ty Ay 4,20 By applying the simplex method the optimal solution that we get is shown in Table 6.22. The optimal solution is: x, = 80, x, = 120 and Min Z= 1,200. aes 3 8 o oom 0 60 0 neuen 1 1 3 4 80 1 0 1 ° s 120 0 1 at ° 1 Ee Table 6.22 2 1200 y 3 ek See ie Primal-Dual an% 0 0 5 te aes Relationship To see the changes in the optimal solution, we are given that c, = 5 ard the column, a Thus, to see the change we calculate: -1 1 Of) cy- 7, =e — eg Bt a,=5-(,3,8)| 1 0 0|/3 -10 dh —11 py] fs a; -Bta,-| 1 0 of]3)-|3 -10 1/3 oO) Since c ~ 2; = -4, which is a negative number, the existing optimal sclution can be improved upon. For this, we start with the optimal solution shown in Table 622 and by adding entries corresponding to variable x, we get values as shown in Table 6.23. 47 3 8 0 Oo mM 5 o s 60 0 boas 1 1 @® | «3+ 3 * 80 BEES C6 150) Loe ee Beans 8 4 120 Opt apcopetl A0, ee ashe 20, = Z= 120 sega: S80. Sw aes Table 6.23 o o s 2 EAS New Column with , Variable x, Added ‘The variable x, in Table 6.23 must enter into the solution and s, should leave it. The new solution is shown in Table 6.24 go 3 ‘ oo M 3 ‘ 5 =I Isa a 1 3 i 1 0 Pee o 8 ys 0 1 “re 1 0 Zz 3 4-18 43-45 5 Oo 6 ms a3 sas ° Table 6.24 ‘opyrightad materialOperations Research: Theory and Appications. Since all cj ~ z, 2 0 in Table 624, the solution is optimal, with x, = 20, x, = 120, x, = 20 and Min Z= 1,120. 6.2.5 Addition of a New Constraint (Row) After solving an LP model, the decision-maker may recall that a particular resource constraint was overlooked during the model formulation or perhaps he may desire to know the effect of adding ¢ new resource to enhance the objective function value. The addition of a constraint in the existing constraints will cause a simultaneous change in the objective function coefficients ()), 25 well as coefficients aj, of « corresponding nom-basic variable. Thus, it will affect only the optimality of the problem. This means that the new variable should enter into the basis only if it improves the valve of the objective function Suppose that a new constraint eet Tht Bnet. 27 tn XS Ome is added to the system of original constraints Ax—b, x30, where b,,, ; is positive, zero or negative. Then the following two cases may arise. 1. The optimal solution (xg) of the original problem satisfies the new constraint If this is the case, then the solution remains feasible as well as optimal. This is because the new constraint either reduces or leaves unchanged the feasible region of the given LP problem 2. The optimal solution (xg) ofthe original problem does not satisfy the new constraint. In this case the ‘optimal solution to the modified LP problem should be re-obtained. Let B be the basis matrix for the ‘original problem and B, be the basis matrix for the new problem with m+ | constraints. That is, matrix B, of order (m + 1) is given by a-[2 ' ‘where the second column of B, corresponds to slack, surplus or artificial variables added to the new constraint ed @ = (4p) » 2 == Ge t,n) 8 A TOW Vector containing the coefficients in the new constraint and corresponds to variables in’ the optimal basis. In order to prove that the new solution sie['D ] ssa sere is abasic feasible solution to the new LP problem, we shall compute the inverse of B, using partitioned methods, a8 given below pi=| BU er ee Since each column vector inthe new LP problem is given by a = (ay a af are given by . Bt olf a ~plat ~ i sore [Se he] “Feng rane ‘The eniries in the ¢;—2* row of the simplex table for any non-basic variable. in the new problem are computed as follows: % ja oe) A 41,5 8) in yj) the new columns -ayf ~ (BY, + Same 1 m1 ~Sam-+1 2%) where cp,,. is the coefficient associated with the new variable introduced in the basis of the new LPprobiem Ifthe new variable introduced in the basis of the new LP problem is slack or surplus variable, then Hence, we baveSensttivity Analysis in Linear Programming ‘That is, the entries in the ¢, ~ z tow are the same forthe initial as well as the new LP problem. The value of objective fanction is given by: Z=Lep a8 ecu “iz, ‘This shows that the optimal simplex table of the original problem remains unchanged even after adding the new constraint. However, the slack or surplus variable appears with a negative value because the optimal solution of the original problem does rot satisfy the new constraints, Thus, the dual simplex method may be used to get an optimal solution, Remarks If the new constraint added is an equation and an artificial variable appears in the basis of the new problem, then the following two cases may arise: 1. Ifan artificial variable appears in the basis at negative value, a zero cost may be assigned to it. Apply the dual simplex method in order to obtain an optimal solution, 2. If the artificial variable appears in the basis at a pesitive value, then ~ M cost may be assigned to Apply the usual simplex method to obtain an optimal solution, Example 6.10 Consider the following LP problem Maximize 2= 3x, + 5x, subject to the constraints © 3yt2Q518, G)yF2_S4 Gi) SO and xy me0 Obtain an optimal solution of the given LP problem. (@) Suppose variable x, is added to the given LP problem. Then obtain an optimal solution to the resulting LF problem. It is given that the coefficients of x, in the constraint of the problem are 1, 1 and 1, and that its coefficient in the objective function is 2 () Discuss the effect on the optimal basic feasible solution by adding a new constraint 21, +x, < 8 to the given set of constraints. [AMEE, 2005] Solution Theoptimal solution: x;~2,13~6 with Max Z~ 3610 the gi is shown in Table 6.25, Table 6.25 Optimal Solution (G) Aftor adding the new variable, say x, in the given problem, the new LP problem becomes: Maximize Z= 3r, + 5x, + 2x suiject to the constraints @ By ty tagS18, GD aytagsd, — Gi) tay <6 and yaya 20 Given that the column vector ass wwe compute ed with variable, x, 1, 1,1), with the help of Table 6.25, v3 0 -23]f1] [713 a 1 4a|lif=] 93 oo 1 fli 1 Since in the current optimal solution, eg = (3, 0, 5), we have -V3 -G,0,5)] 53 |=-260 1‘Operations Research: Theory and Applications As ¢5 ~ %6 $ 0, the optimality of the current solution remains unaffected with the addition of xg. (b) The optimsl basic feasible solution given in Table 6.25 does not satisfy the additional constraint 2x, + x $8. Thus, a new optimal solution is obtained by adding slack variable s, to this constrain, ‘Then, itis written together with the entries in Table 6.25, as shown in Table 6.26. a 0 2) Pal i x Table 6.26 z= 36 o-% In Table 6.26 it may be noted that the basis matrix B has been disturbed due to row 4. Thus, the coefficients in row 4 must become zero. This can be done by using the following row operations. Ry (tw) > Ry (Old) ~ 2R, ~ Ry We get a new table as shown in Table 6.27. Tapie 8.27 Since the solution given in Table 6.27 is optimal but not feasible, apply the dual simplex method to get an optimal basic feasible solution. Introduce s, into the basis and remove s, from the basis. The new solution is shown in Table 628, q> 3 5 0 0 ° ° 3: oo 1 1 0 0 Oe 12 0 as 2 0 Crile) 1 1 -1 5 % 6 (Seis ae (eA 1 ° Table 6.28 Oa are 3 @ DBAs) fo ea eee Optimal Selution 2-3 Se OPO RU pr eee Since all cj ~ z; $ 0, the solution shown in Table 6.28 is the optimal basic feasible solution: x, = 1, x= 6 and Max Z 33. It may be noted here that the additional constraint has decreased the optimal value of the objective function from 36 to 33.‘Sersitivty Analysis In Linear Programming Ena Example 6.11 Solve the following LP problem Maximize Z= 3x, + 4x, +2, + Try subject to the constraints @ Br, +3x +44, + 2457, (i) xy + 6x, + x + SmyS3 (ii) =) + 4x, + $x, 428,58 and Xp Xp Ay 420 (®) Discuss the effect of discrete changes in (i) RHS of constraints, (ii) Coefficients of variables in the objective function, (iii) Structural coefficients in the constraints on the optimal basic feasible solu- tion of the LP problem. (© Discuss the effect on the optimal solution of the LP problem of adding an additional constraint: Dey + 3x +25 + Sky SA. Solution The optimal solution of the LP problem is shown in Table 6.29. 9? 3 4 1 z ° oo S x ing | 1 98 12 0 SR 0 a ee sp | 0 2g oD 0 0 s 126119, 0 59mg ne BB 1 7809 5 sy pANIAR Jap conor ips 0 Table 629 6-3 | 9 698 120-088-5338 Optimal Solution The optimal solution is : x; = 16/19, x, = 0, x 5/19 and Max Z = 83/9. (@) (The range of variation in RHS velues 6,(k = 1, 2, 3) can be obtained by solving the following system of inequalities: oom om 3430) hn<0 For k= 1, b, = 7 we have 16/19 3/28 5/29 -126/19 seh enn ar} = 3215 sb <5 7-325) < Aas So = 1, 2,3 and column j = 6 (-53/38)(1/38) < Any, or oy 2-53 (-53/38)(5338) $ agg oF Ady, 2-1 (53/38) € Aayg or —— 2 ayy <= The effect of change in coefficient a, of basic variables x,, x, and s, on the optimal solution is determined by solving the followings system of inequalities. Any change in the coefficients of basic variables may affect both feasibility and optimality of the solution, Considering a change in the element belonging to basic variable x, = By i.e. day Feasible condition: Tomaintain the feasibility of the solution we need to solve the following inequalities: c For row For row x | — tx ng, < Min} vsfam hia seem anctaal ‘where Bi are the entries under column s,, s, and s, in the optimal simplex table.Sensitivity Analysis in Linear Programming Since a1 Bag ~ Xn Bas = X01 Bra ~ Xz2 Box (16/19) (4/19) ~ (15/19) (-1/19)= 7/38 Bao = Boo — X52 Bax 126/19) (4/19) — (15/19) (15/38) = 3/2 ‘Thus the range of variation in the element ay, maintaining feasibility is given by ns Bag ~ 16/19 -126/19 vn [ELBE an ar 9 tng Optimality condition: Further, consider the discrete changes in ay, belonging to column vector. x47 By in the basis matrix B. 8(_ 169) 21/53 Bas t64=4)~ rato = 3(—2)-B(3] 1.47 Bao(ex— 20 ~ 324 Gn Ba 5-3) -O{ 5p] 0105 2 33) _ 3) =0.078 8 Bas es=29~ resem Boe 3-35 8(_53)_4(53)_ Bop (C5 - 26) — ¥06 &p By aR) a(S) = 0.293 ‘Thus, the range of variation in the element a;,, maintaining optimality, is given by 1/38 _ 53/38 1/2 } é $ ayy S Min 0.078" 0.293 } a { = 0.105 .= 4.760} $8 a34 0, therefore the current solution is optimal. The new solution value are: x, = 33/38, x) = 0, x5 = 0.x, = 1/19 and Max Z= 113/38. CONCEPTUAL QUESTIONS ‘write @ shor rele on senstivty analysis ©. Given that the problem: Mex Z= ex such thal AX = b, x 20 hhas an optimal solution, can one obtain a linear programming 2. Dscuss the role of sensitvty analysis in near progremming. Under what circumstances i it needed, and under what conditions | Problem which has an unbounded solution changing B alone? ‘do you thir it & not necessary? 7. Consider the LP problem: Max Z= ex; subject to AX = b ard Jain how a change ina hputoutputcoeticient can affect x2 0, whore ©, x7 ¢ E", b? © E and A is mx cooficienis oon eaten ‘nat, Batoce how much ca components the cost veer (€) How can a change in resource availabilty affect a solution? be changed wihout affecting the optimal sluon of he LP 4. What do you understand by the term ‘sensitvly analysis"? e Dscuss the eftect of) vation of «, (i) vation of b, and i) addtion of & new constant Discuss the changes in the coetiionts a) for tho given LP problem: Max Z= ex, subject to Ax= b, £2 0 Find the limts of variation of element a, so that the optimal feasible soluion of Ax = b, x 2 0, Max Z = ox remains the opimal feasble solution when () a, ¢ B, (i) % € B. SELF PRACTICE PROBLEMS Ina LP problem, Max Z = ex, subjectto Ax = b, x>0, obtain re Variation in coeftcents ¢ which are permitted witout changing ne coptimal soluion. Find this forthe following LP problem 2. Discuss the effect of changing the requirement vector from [6 4 24}t0[6 2 12] on Me optima’ soluion of he folowing LP problem: Max Z = 3x, + 5%) Max 2=8x, + 6% + % ‘ubjectto (Ix, + % <1; (2m Hayed subject to () 44 ZO) (DH eS ES ard Kh 20. [Meerut Univ, BSc, 1990] i) %+5%+ S28 and Ky Be X20.Discuss the effect of discrete changes in the parameter by(i= 1,2, 9) for the LP problom. Max Z= 3x, + 4%, 4354 7, subject to () 8x, 18% + 4y+ %457: | () 2x, +8 +35 +8x, 53: lil) +44, + 8x +24, 58 | and Xe %% 20 [oon Un, MSe (Math), 1902) | Given the LP problem Mex 2= x, + 2%) — subject 10 (i) 3, 4% = 1S 10; i) =X, +4 Hz 6; fi) meme ard XH G20. Sl SSS daceth a eeecon sSeantenabe necaton Seen maa ae ee ern Max 2 = 3x, + 5x | come Oetayss use mace | Discuss the effect on the optimally ofthe solution when the | pean fash mien nan LP problem, Max Z= ex, subject Ax = b x2 0, obiain the | ‘ataton in efits ¢ wich are permitted without changing te ‘optimal sluton. Condit sensivty analysis onc forthe flowing UP protiem, MaxZ = x, + x subjectto () %,#%5 1 (i) 2x, 4351 and xi %20, [Meerut Univ, BS, 1980) 1 Gren te LP problem MaxZ=%,+2% + % subjectto (i) Sy 4% —%S 10; fl) y+ + HSE W) ytays4 and yy X20. (a) Determine optimal solution tothe problem. (&| Determine the effect of dizcrete changes in hove components ‘ofthe cost vector which corresponds to the basic variable. Find the optimal soluticn tothe LP problem Max 2= 15x; + 46x subject to (i) % + 18x, <250, (I) Sx, + 2x, < 162, i) 455 ‘and 9520. Max 2 = 3.6 xy /= 1,2 and cp is kept fixed at 45, determine how rmucn can c, be crangea winéut aecing te opimal solution ot the problem. (Bombay Uni, BSc (Maths), 1981] ‘Given the LP problem Max Z=— % + 2% — | aiblectto () 24% — S10, Dy sdyeyrs — | ‘Sensitviy Analysis in Linear Programming ) myemsa and hy Hy 8520. ‘Determine the range for dscretechengesin the rescurcevalues +b, = 10 and b, = Gof theLP rodel s0.asto maintain optimality of he ccient solution 10. Consider the following LP problem Max Z= 4x, + 6x subject io () x +2558, and Xs ¥20. (a) What is the optima soltion? (©) It he frst constraint is altered es: x, + Sx < 8, does the copimal solution change? 11. A stainless steel uensil manufacturer makes thee types of tems. ‘The restrictions, pofis and requirements ae tabulated below: (i) 6x, #4524 sensi Type t mW! ° a 3 " Faw materia requrernent fg per unt) ‘Welding and finishing ime hours per unit) 3 4 Profit per unt (Rs.) 1 Hf stainless steo! (raw materia) availabilty is 25 kgand welding ard finishing time available io 20 houro par day, thon the optimum ‘reduct mix problem is expressed asi Max 2 = de + 4 4 subject io 4, + 8% +54 5 25 (raw material restriclon) 8x, + 4% +8 $ 20 (time restiction) ‘ang ym 20 where (j= 1,2, 3)s the numberof units ofthe fn type ofthe item to be pioducsd. Get the optimal simplex table and encicle the appropriate answer to the folowing quastions. (a) The second type of utensil would change the current optimal ‘basi if be profit per unk ie: = 1; = 1.02;> 1.06; 22; > 3 The simplex muliplior associated wih the machine time ‘esircton of 20 hours Is (~ 9/9). Thus, the multiper remains ‘unchanged for the upper limit on the machine time avalabiliy of 25; 27.5; 35; 425; 32.5 hours. ‘The increase in the objective function fer each unt avalabilty ‘of machine tine higher than the upper limit indicated in part (b) is .. (show calculations). ‘The proit ofthe third type of utensil is Rs. 4 per unit. The ower uri on its profitability such thatthe curent basis is stl optimal s 4; 9;2.5 2: <2. 12, The folowing LP model apps for hast & Sons Wood Furniture Company which makes tables (), chairs (C) and beok shelves (8) along win otner tens. Max 2 = 200%, + 150%, + 150%, subject 10 1Oxy+ Bxp + 10xy £ 100 (Wood) 5x74 5x,+ 5x5 60 (Labour) ©) © @ ‘and Xo *Be 0 ‘The optimal simplex table is shown below: 4 2001601500 o ‘Solution eee Values biexy oar 1 ° 1 17-670 2077 o 1 0 -m oan 200 «180 «20007 ° 0-50 sor -807 ‘where &y ard 6, are the slack variables for unused word and labour, respectivelyOperations Research: Theory and Applications. (@) Determine the sensitnity mits for the avaliable wood ang labour within which the present product mix wil remain optimal (©), Find the new optimal solution when the available woed is 90 board feet and labour i 100 hours. 1, Refer to the data in Practice Problem 12, (a), Determine tho conctiviy inte for unit profits within which tho Current optimal solution will remain unchanged. (©) What ie the total proft when each table yields « profit of Rs, 180 and each chai yields a profit of Rs. 100 14. A.company makes two products A and B. The production of bth products requires processing time in two departments | andl. The hourly capacty of anc I, unt profi fr products A and B and the processing time requirementsin| andl ee giveninthe folowing tai: Department Product Capacity a 8 (Hours) 1 2 2 o i 8 Unit profit (Rs.) 200 300 Now the company is considering the addition of anew product C1o its tine, Product C requires one hour each of departments | andl What must be preduct C's unit profit in order to profitably add ito the fis product ine? 415. A company produces three products A, B and ©. Each product requires two raw matetals: steel and aluminium. The folowing LP ‘model describes ine company's prodvet mx probit. Max 2 = 30%, + 0x + 50% sibject to 6444+ 349 45%0 £ 450 (Stee!) 8a, + 44g + 8x¢ 5 300 (Aluminium) Units of nutrient B ard 50 units of rutrient ©. Would ths. ‘change the optinal mx of ead? If yes, how? 17. A company sells two diferent products A and 8, The seling price ‘and incramertal cost information is as follows: Product A Product Selling price (Ra) 60 40 Incremental cest (Fs.) 30 10 Incremental profit (Re.) 30 20 ‘Tho two products are produced in a comman producton procece and are sod in two different markets. The production process has ‘2 capacty of 30,000 latour hours. It takae thro hours to produco 4 unit of A and one hour to produce a unit of B. The market has ‘been surveyed and.company oficial feel thatthe maximum number ‘of units of A that can be sold is 8,000; and that of Bis 12,000 units. (a) Find the optimal product mix (©) Suppose masimum number of units of A and B that can be Sold is actualy 91000 units and 19,000 units, respectively Insiead of as given inthe problem, what effet does this have (nthe solution? What isthe effect on the profit? What is the ‘shadow price forthe constraint on tho eal liit for bot these products. (©) Suppoce thor are 31,000 Iebour hours avaliable nstoad et 30,000 as in he base case, what effect does this have onthe ‘olution? What i the effect on the prof? What isthe shadow rice forthe constraint on the numberof labour hours. 18. Anorganizaton can produce a particular component for passenger cars, Jeeps and trucks. The production ofthe component requires tutlzaton of sheet metal working and painting facie, the detais of which are aiven below. Consumption (in hra) to and ete? 0 Rocourco Avalebity| ‘The optimal production plans given inthe following table: Produce a Unit or (his) o> 3 Ww 5 0 O Passenger Car Jeep Truck ‘Unit Front Varabies Soutin | x, pe %> 8 ‘sheet metal in Basis Values working 02s 1 os 2 e, a bexy Painting 08 12 0 a ee ee The protits inat can be earned by the three catagories of 50 % eo | 35 45 1 0 18 components, 2. for passenger cars, jeeps and trucks are Ris. where s, and sq are the slack variables for unused steel and ‘aluminiuin quantly, respectvaly. (a) Suppose an adaltonal 300 tonnes of steel may be procured ata cost of Re. 100 per tonne. Shoulé the company procure the addtional steel? (b)_ Unit profit of product Ais Rs, 30. How much should ths price be increased so at A is produced by the company? 16. Apig farmer is atempting to analyse his feeding operation. The ‘minimum daly requirement of the three nutrtional elements forthe pigs and the number of units of each ofthese nutrtional elements Intwo feeds is given inthe following table: Required ‘Unt of Nutrtionat Meimun ‘Nuttional Elements (i kg) Requirement Benen Food 1 Food 2 a 20 30 200 B 40 25 350 c 30 45 430 Cost per kg (Fie) 5 2 (a) Formulate and solve this problem as an LP model (b) Assume that the farmer can purchase a third feed a a cost f is. 2 per ka, which will rovide 36 units of nutiont A, 80 £600, Fs, 1,400, and Rs. 1,900, respectively, (@) Find the optimal product mix for that organization. (b) What acditional profit would be eamed by increasing the Availabilty of()) Shoot metal working chop by an hour only, and (i) Paintng shop by an hour only. (©) What would be the effect on the profit earned if at least one, ‘component fcr jeep had to be produced? 19. A firm uses three machines in the manufacture of three preducts, Each unit of product | requires 3 hours on machine 1, 2 hours on ‘machine? and 1 hour on machine 3, Each unit of product Il requires 4 hours on machine 1, 1 hour on machine 2 and 3 hours on ‘machine 3, Each unit of product Il requires 2 hou's on machine 1, margin cf the tree products s As. 30, Rs. 40 and Rs. 35 per unt, respectively. Available ‘or scheduling sre 90 hours of machine 1 time, 54 hours of machine 2 me, and &3 hours of machine 3 time (2) What isthe optima production schedule forthe fem? (b) What is the marginal value of an addional hour of ime on machine? (c) What is the oppertunity cost associated with product 12 What interpretation stould be given to this opportunity cost? (a) Suppose thatthe contrtution margin for products increased ‘rom Ris. 30 tos. 43, would this change the ootimal production plan? Give reasors.‘Sensitivty Analysis in Linear Programming [xuRy HINTS AND ANSWERS 1. = $e, $3+(13; $-(12)Seq$5+0 of WBSc, Se ‘The currenthasie feasible solution consisting of, ed x remcins feasible andoptimal atthe new values x and, =O with Min 2= 26, 3. SBE —205; BASHA E54; Aby= 4. by <0; -52.<6,26 5. Initial optimal sclution:x, = 2,1» = 6and Max 2= 36. Where ¢, 4%=—Landes—25=—3. After having cy =3 aud c= 1 ealculate now valuesofe,~2,~ 1 and cy ~2.=2/3. Ths, introduces, inthe tasis to getthe new solution: x)= 2,x, = 6 and Max Z = 12. 7. () %)= 0,4) = 4,45 = Oand Max Z=8 (b)—(12)Sc 8. (I5—195/16) $c $ (15+ 195/16) 9. by < 10; 325d, 26 12619 12, @36Sx, £120, 505, 5003 @xp-S47% 407, 5e=40/7; 1007; 2=18,3007, 23,0000 1, 18SxpS50; 65 x¢5 20; —-S19520 TT. (@) = 6000, x)= 12000, Max Z~Rs.5,40000 ©) Asx, £8,000isnotbindingbecause x, ~ 6,000 inthe slain, “hus, am inereas inthe init on will have noefTeet on the solution or prof. Shadow prices er. For increase inthe limit OB, the new solutions, x,~ 5,70 and.xg~ 13,000 aad Max Z= Rs, 5,60,000. The shadow price is (560,000 — 540,000) = Rs. 20,000. Now solution is, ~ 6330andsy = 1,000 with MaxZ=Re, 550,000. The stadow price is (5,50,000 ~ 540,000) = Rs. 10,000. ° CHAPTER SUMMARY Sensitivity analysis provides the sensitive ranges (both lower and upper limits) within which LP model parameters can vary without changing the optimally of the current solution, Formulae to conduct sen: vity analysis followed by numerical examples Ihave been solved to illustrate how changes in the parameters specially, (i) the change in the objective Function coefficients right hand-side value of a constraint and (ii) input-output coefficients in any constraint of a LP model, affect its optimal solution. CHAPTER CONCEPTS QUIZ True or False 1. The solution to the dual yields shedow prices. 2, The absolute values of the numbers in the 6 ~ row under the ‘sack variables represent the solutions 10 the duel problem. 3. The transpose ct the primal constrants coetfiients become the dual constraint coefficients, 4. The range over which shadow prices remain valld is calle right. hand-side ranging, 5. The stadow price is the value of additonal profit margin trom an acivity to be conducted, 8. The solution in optima as long as all G)~ z)2 0. 7. Non-basic variables are those that have a value of zero. 1 Testing basic variables does not requre reworking the thal ‘smplex table during sensitivity analysis, 8. Ina sinplex table, ial ofthe substitution rates in the key column are negative, then it indicates an infeasbie solution. 110, Naking changes in the resources value result in changes in the feasible region and fen the optima solution, Fill in the Blanks 1. The of tie resouree 12. The range of significance are values over which a coetficent can vary without causing @ change i solution mia 13, The range of optimality are values over which a ‘eofficent can change without causng a chango in the optimal solution mis. 14. ighthand-side rangng method 1s used to find the range over whieh remain vali. {for @ constrain isthe value of an edkltienal unit the optimal 15, Inthe optimal simplex table the absolute value of cj ~2j eumbers ‘ceraepands to slack variable column reprocont of the available resources. 16, The adiion of a consirain In the exsting constrains wit cause a change in the objective function covffcients. 17. The adition of new constraint the feasible region of the given LP preblem. 418. Sensitty analysis provides the range within which a parameser ‘may change without offering 48. When an addtional variatle is added in the LP problem, the soluton can be improved if g,~ 6 = 0. ‘A non-basic variable shoud be brought into new solution mix 20, provided its 18 G <6, + - 6). ‘Multiple Choice 21. Sensitity analysis, {is also called pestogtimalty analysis as itis cared cut afer the optimal solution is obtained (©) allows he decsion-maer to get more meaningful information| atout changes in the LP model parameters provides the range within which a parameter may change ‘wihout affecting optimality (4) allof the atove 22. When an additonal variable is added in the LP model, the cating optimal solution can farther be improved # (@) G-z20 @) g-3s0 (6) Bath (@) and (&) (8). fone of the above 23. dition of an aditioral constraint inthe existing constraints wil cause‘Operations Research: Theory and Applications (0) change in coeticients ay {(c)_ both (a) and (b) (0) one of the above 30. In sensitivity analysis of the coeficient of the non-basic variable In a cost minimization LP problem, the ugper sensivty limit is (a) orpinal value + Lowest postive value of improvement rato (©) orginal value — Lowest absolute value of improvement rato (a). change in ebjectve function coetcnts (<) 24. 1 the actienal constnint added isan equation and an artical | (e) postive Infinty variable appears in the basis of the new problem, the new | —() negative infty ‘optimal solution is obiained by . 31. It the RHS value of a constraint is changed to another value within fa) ‘sg ro cast coficint fo the arial vale fi ‘specified limits, then the new optimal solution would have chosen fe basis at nepate value ees ne eee ne (0) signi ~ M coat cnfcot th ail varie 1 Fest apecbes tein vue appears In he basis at postive vale {c) &tigher objective tuncton value (6) eer (2) and () (same set of variables i tho soliton mix (4) nore oF the abore 232, A change in the abjecive function for a non-basic variable an 25, Te siete beat maa asa the atacve ton | “thet valve a resource value mey be increased whose shadow prce | (a) ¢j~z, value of al nonbasicvarabos accent ‘ei aaa | () ¢~ aval of al base varables 1h IO ay worse Od ae ancic9, | Steg apemenaieteens 26, Anonctasic variable shoulé be brought info the new soluton mix | (d) none of the above provided its contibution rete (q) 1S | 33, The 100 per cent rule for RHS velues says that ifthe sum of the dager Od mae tga) | ~~ percentage change i ee than or equal fo 100%, then (© dede—a) (rane ofthe above |) Sinn nether 27, While performing sensitivity analysis, the upper bound infinity on vend cel estes many the vaio of the rigthang side ofa contrat mano tat | (6) Bo (2 and) ay oor (a) there is slack in the consiraint poe (bj the constraint is redundant (34. The 100 per cent rule for objective function coefficients says that (c} the shadow price for that constraint is zero it the sum of the percentage change is less than or equal to 100%, (3) none of the abore then 26 Tr ertwing vale n he sony ana of ebjoctve | (2) onal aun des change function coaicients says | 2 clerent optimal 0 em {a} decision variable (b) ron-basic variable Sh bot (a ane ‘may occur, {e| basic variable (8). lack varieble | ‘none ofthe above 28, To maintain optimality of curert optimal sluton fer a change | 95- The igt-nand sie range is ote refered 0 asthe range of gin Be coeficent , of on-base variable, we must ave | (2) Improvement (2) feaiity Wace nae, (6) infeaibity {optimality ytaq=y a2 y | oo Caer eal wT 27 37 47 6F 6F 77 8F 8F wT 11, shadow pico 12. nontbasc varable 18. basic vafebla 14. shadow pce 18. shadow price 16. simutaneois 417. leaves. unchanged 18. optmaly 19, existing optimal 20. contrbuten rato 2. @) 22 (a) 23. (0) A.) 25.) 28. (6) BF () BAS) -2B. fo) 9D. (6) 9H. (SY 92, (c) 33 (a) 34. (a) 35. (b)Integer Linear Programming “The key is not to prioritize what’ on your schedule, but to schedule your priorities.” = R. Covey Preview In this chapter, Gomory’s cutting plane method, and Branch and Bound method have been discussed for solving an extension of LP model called linear integer LP model. In ¢ linear integer LP model one or more of the varigbles must be integer due to cerain managerial considerations. Learning Objectives After studying this chapter, you should be able to ‘+ understand the limitations of simplex methed in deriving integer solution to linear programming problems. ‘+ epply cutting plane methods to obtain optimal integer solution value of variables in an LP problem, © apply Branch and Bound method to solve integer LP problems. ‘© appreciate application of integer LP problem in several areas of managerial decision-making. chapter ey Chapter Outline mM 12 13 14 18 18 77. Introduction ‘Types of integer Programming Problems Enumeration and Cuting Plane Soluton Concept ‘Gomory’ Al integer Cutting Plane Method 4 Salf Practice Problems A + Hinte end Answers Gomory's Mixed-Integer Cutting Plane Method Branch and Bound Mettod Applications of Zero-one Integer Programming ‘+ Conceptual Questions + Self Practice Problems B +) Hots and Answers Chanter Summary Chapter Conceots Quiz ase StudyOperations Research: Theory and Applications. Classification of 7.4 INTRODUCTION In linear programming, each decision variable as well as slack and/or surplus variable is allowed to take any real or fractional value. However, there are certain real-life problems in which the fractional value of. the decision variables has no significance. For example, it does not make sense to say that 1.5 men will be working on a project or 1.6 machines will be used in ¢ workshop. The integer solution to problem can, however, be obtained by rounding off the optimum value of the variables to the nearest integer value. This approach can be easy in terms of economy of effort, time, and the cost that might be required to derive ‘an integer solution, This solution however may not satisfy all the given constraints. Secondly, the value of the objective function so obtained may not be the optimel value. All such difficulties can be avoided if the given problem, where an integer solution is required, is solved by integer programming techniques. Integer LP problems are those in which some or all ofthe variables are restricted to integer (or discrete) values. An integer LP problem has important applications. Capital budgeting, construction scheduling, plant location and size, routing and shipping schedule, batch size, capacity expansion, fixed charge, etc, are few problems that demonstrate the areas of application of integer programming. 7.2 TYPES OF INTEGER PROGRAMMING PROBLEMS Linear integer programming problems can be classified into three categories: (i) Pure (all) integer programming problems in which all decision variables are restricted to integer values. Mixed integer progranming problems in which some, but not sl, of the decision variables are restricted to integer values. (ii) Zero-one integer programming problems in which all decision veriables are restricted to integer values of either 0 oF | @ The broader classification of integer LP problems ané their solution methods are summarized in Fig. 7.1 In this chapter, we shall discuss two methods: (}) Gomory’s cutting plane method and (ii) Branch and Bound method, for solving integer programming problems. Integer Programming, Problems ———e Linear Integer Programming Non-linear Integer Programming Proplemt Problems Pure Integer Mixed Integer ot Polyzomiat General Problems Problems Problems Programming Non-linear Problems Problems (© Catting Plane (Cutting Plane Pure Ineger Mixed Ineger ‘Method Methed Problem Problem (i) Enumeratve (or (il) Enumeratve Branch and Method Bound) Method ii) Balas Additive aaa is Generalized Penslty Function Method ‘The pure integer linear programming problem in its standard form can be stated as follows: Maximize 2= 0, x + 6%) +... + 69% subject to the constraints ayy, + ayy ky +o. + yg hy = 5, Ce ee a a Gi 8+ a2 > * han Fy = Em and pepe ove 4p 20 and are integers.Integer Linear Programming Aig 7.3 ENUMERATION AND CUTTING PLANE The cutting-plane method to solve integer LP problems was developed by RE. Gomory in 1956. He developed this by using the dual simplex method. This method is based on the generation of a sequence of linear inequalities called a cut. This cut cuts out only a part of the feasible region of the corresponding, LP problem while leaving out the feasible region ofthe integer linear programming problem. The hyperplane boundary of a cut is called the cutting plane. LUTION CONCEPT Mlustration Consider the following linear integer programming (LIP) problem 4x, + 16x, subject to the constraints @ 4x, #3y)5 12, (i) 6x, + 81 524 yy 2 Oand are integers, Maximize Z: and Relaxing the integer requirements, the problem is solved graphically by using Fig. 7.2. We obtain the optimal solution to this LP problem as: x, = 1.71, ,=1.71 and Max Z=51.42 This solution does not satisfy the integer requirement of variables x, and x,, The feasible region (solution space} formed by the constraints is marked by OABC in Fig. 7.1. Rounding off this solution to xy solution is infeasible. The dots in Fig. 7.1, also referred to as fattice points, represent all ofthe integer solutions that lie within the feasible solution space of the LP problem. However, itis difficult to evaluate every such point in order to determine the value of the objective fundtion Oata-2 © si th=2 ILP solution QA, woes 2240) 3 ee LPsctason Ge ET LT, 251.42) Be ILP solution Ga 3n=0 2-2) Figure 7.2 suggests that we can find a solution to the problem when the problem is formulated as an LP problem (which as a matter of chance could contain integers). It may be noted that the optimal lattice point, point C, lies at the comer ofthe solution space OABC, cbtained by cutting away the small portion above the dotted line. This suggests a solution procedure that successively cuts down (reduces) the feasible solution space until an integer-valued comer is found. ‘The optimal integer solution to the given LP problem is: x, = 0, x53 and Max Z = 48, Notice that its lattice point is not even adjacent to the most desirable LP problem solution comer. Remark Reducing the feasible region by adding extra constraints (cut) can never give an improved objective function value. Usually it makes it worse and if Zp represents the minimum value of objective function in an ILP problem and Zp the minimum value of objective function in an LP problem, then Zp > Zp 7.4 GOMORY'S ALL INTEGER CUTTING PLANE METHOD ‘A systematic procedure called Gomory’s all-integer algorithm will be discussed here for generating ‘cuts’ (additional linear constraints) so as to ensure an integer solution to the given LP problem ina finite number Cf steps. Gomory’s algorithm has the following. properties, 2 = 2 does not satisfy both the constraints and therefore, the” ‘A ut isthe Inear ‘constraint added to the given LP probiem constraints Fig. 7.2 Concept of Cutting Plane Hyperplane ‘cutting planeOperations Research: Theory ana Applications ’n integer LP froblem is usually worse in terms of Higher cost or lower rotit Constants called Gomory cuts lead to a smaller fessible region that includes all feasible integer values @ Additional linear constraints never cutoff that portion of the original feasible solution space that Contains a feasible integer solution to the original problem. (Each new additional constraint (or hyperplane) cuts off the current non-integer optimal solution to the linear programming problem. 7.4.1 Method for Constructing Additional Constraint (Cut) Gomory’s method begins by solving an LP problem without taking into consideration the integer value requirement of the decision variables. Ifthe solution so obiained is an integer, ie. all variables in the *xg” column (also called basis) of the simplex table assume non-negative integer values, the current solution is the optimal solution to the given ILP problem. However, if some of the basic variables do not have non- negative integer value, an additional linear constraint called the Gomory constraint (or cut) is generated After having generated a linear constraint (or cutting plane), itis added to the bottom of the optimal simplex lable so that the solution no longer remains feasible, The new problem is then solved by using the dual simplex method. If the optimal solution, so obtained, is again a non-integer. another cutting plane is generated. The procedure is repeated until all basic variables assume non-negative integer values “The procedure of developing a cutis discussed in what follows. In the optimal simplex table, we select one of the rows, called source row for which basic variable is non-integer. The desired cut is developed + by considering only fractional part of the coefficients in source row. For this reason, such 2 cut is also referred to as fractional cut. Suppose the basic variable x, has the largest fractionsl value among all basic variables, restricted to be integers. Then the rth constraint equation (row) from the simplex table can be’ rewritten as: Xp EF bJ= 13, + (ay Hayat +E a, x o feet where xj (j= 1, 2, 3,...) represents all the non-basic variables in the rth constraint (row), except the variables x, and 6, (=x,,) are the non-integer value of variable x,. Let us decompose the coefficients of 4p X, Variables and xp, into integer and non-negative fractional parts in Eq. (1), as shown below: [edt = 1404+ F tlay lt Sesh ty 2 ‘where [rp] and (a, denote the largest integer obtained by truncating the fractional part from x, and a, respectively Rearranging Eq. (2) 50 that all the integer coefficients appear on the left-hand side, we get Ft Wd —¥— Flares }= % fey ey @) where /, is stritly a positive fraction (0 < f, < 1) while /, is a non-negative faction (0 5 fy, < 1) Since all the variables (including slacks) are required to assume integer values, the terms in the bracket ‘on the leftchand side as well as on the right-hand side must be non-negative numbers. Since the left-hand side in Eq, (3) is f, plus a non-negative number, we may write it in the form of the following inequalities: GS Efi ts ® « Bfoxs oF fe S> Bs 6 where s, is a non-negative slack variable and is called the Gomory slack variable, Equation (5) represents Gomory’ cutting plane constraint. When this new constraint is added to the ‘bottom of optimal simplex table, it would ereate an additional row in the table, long with a column for the new variable s,. 7.4.2. Stops of Gomory’s Alll Integer Programming Algorithm ‘An iterative procedure for the solution of an all integer programming problem by Gomory’s eutting plane ‘method can be summarized in the following steps. Step 1: Initialization Formulate the standard integer LP problem. If there are any non-integer coefficients in the constraint equations, convert them into integer coefficients. Solve the problem by the simplex method, ignoring the integer requirement of variables.image not availableaa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.image not availableimage not availableBEEN Oporations Research: Theory and Applications eration 3: Remove the variable s,, from the basis and enter variable s,, into the basis by applying the dual simplex method. The new solution is shown in Table 7.10. g> 2 20 -0 0 oOo oO 9 Table 7.10 Optimal’ Solution Since all variables in Table 7.10 have assumed integer values and all ¢,~z, $0, the solution is integer optimal solution: x, = 2, x, = 0,x, = 2 and Max Z=~ 16. Example 7.3. The owner of # readymade garments store sells vo types of shirts ~ Zeeshirts and Button- fy ad xy integer 7.5.2. Steps of Gomory's Mixed-Integer Programming Algorithm ‘Gomory’s mixed-integer cutting plane method can be summarized in the following steps: ‘Step 1: Initialization Formulate the standard integer LP problem. Solve it by simplex method, ignoring, integer requirement of variables. Stop 2: Test of optimality (@) Examine the optimal solution. If all integer restricted basic variables have integer values, then terminate the procedure. The current optimal solution, obtained, in Step 1 is the optimal basic feasible solution to the integer LP problem. (b) If all integer restricted basic variables are not integers, then go to Step 3. ‘Step 3: Generate cutting plane Choose a row r corresponding to # basic variable x, that has the highest fractional value f, and generate a cutting plane as explained earlier in the form (Ret. Eq. 13} S fe a, fet % agry + 1 3 wr (fi Stop 4: Obtain the new solution Add the cuiting plane generated in Step 3 to the bottom of the ‘optimal simplex table as obtained in Step 1. Find a new optimal solution by using the dual simplex method and retum to Step 2. The process is repeated until all restricted basic variables are integers. Example 7.4 Solve the following mixed-integer programming problem: Maximize Z=~3x, +x, + 3, subject to the constraints @ -x42y+%,<54 (i) 24-G2)xj<1, Gi) y— 3+ 2S and 1%, 2 0, x, non-negative integer. Solution Given integer LP problem can be expressed in its standard form as: Maximize Z = — 3x, +.x, + 3x, +05) +05, + Os; subject to the constraints Ox t2_tH+s,=4 Gi) 24-GARtY and Xp Hp Sy. Sp 2 0, xy non-negative integer. (il) x)- 3x) + 2ey+5; Step 1: Ignoring the integer requirement, the optimal solution of the problem using the simplex method is obtained, and is given in Table 7.16. 97> 3 1 3 0 0 o 1 es sa | -3n 1 0 0-10 0 5 au on 0 o om 1 107 able 716 3 4 137 sis o fleas eas Optimal Nor- or integer Solution 2247 9-3 [34 o 0-07 0. 3aInteger Linear Programming Ew) ‘The non-integer optimal sotution as shown in Table 7.16 is: x, = 0, x, = 5/7, x, = 13/7 and Max Z= 44/7. Step 2: Since, basic variable ry is required to be integer, constructing Gomory’s mixed-integer cut with the help of x,-row we get Bedue mets 3s Fa EMT TH HT (eysoure row) Since the coefficients of non-basic variables x, s, and s, are positive, therefore factorizing the coefficients using rule (15), we get: 6) 1 iis. 8 3 2 (8) pute s Hor }sea+0m4{0+3)s a{o+2}s Sa Sipe its 6 3, 43,42 6 S4-m) = Sx +int2s, ie 2s 4. Gtd-m) = Satiatin, ie Fs Sut (On aulding slack variable s, we obtain Gomory’s mixed integer cut as follows: i 2. 2 8 ag a Fp Fn Fa + 5 =F Mixed imeger cut 1 ‘Adding this constraint to the bottom of the Table 7.16. The new values so oblained are shown in ‘Table7.17. 4° -3 1 3 0 0 0 0 1 ay si Bye a 0 Oh eet eee 0 3 487 7 0 60 Um 1 8 3 4 Bq S00 tae are 10s. reared, 0 i a | -si4 00-37 GD 1 ——_——t___—_+ - Table 7.17 zea G-y. [Hsia a0) 50) 11 Oa ‘Optimal but Ratio: Mint, — 3)yy(<0)] SVS — — NB — SR — infeasible io: Min(c, ~ =)ry (< 0) a ‘Solution eration I: Remove the variable s,, from the basis and enter the variable s, into the basis by applying ‘the dual simplex method. The new s6lution so obtained is shown in Table 7.18. 4? 3 1 3 0 0 60 0 i % a7 “4 1 0 mm 0 0 -m © 5 .w7 igse 0 20; “econ eee comes 3 oa 1 0 o 1 o 0 o t 0 ie 3 Be 0 on aac wh ieee mee7is Z= 297, es ee ee ee ee Optmal Solution Since variable x, has assumed integer value and all c,~ z; <0, the optimal mixed integer solution is: 3) = 0,4) = 87, x= | and Max Z = 2977. Example 7.5. Solve the following mixed-integer programming problem Maximize Z =x, +) subject to the constraints @ 342555, Gi) <2 and XX, 2 0, x, non-negative integer.‘Operations Research: Theory and Applications Solution After converting the given LP problem into its standard form, obtain an optimal solution, ignoring the integer restriction on x, by the simplex method. The optimel solution is given in Table 7.19. got 1 o o 1 e 1B 1 0 1B i % 2 0 1 0 Optimal’ Solution Zi @-45 [0 0 =18 Since all cj ~ z;$ 0, the optimal non-integer solution is: x, = 1/3, x, = 2 and Max Z = 7/2. In the current optimal solution the variable.x,, which is restricted to take integer value, is notan integer, therefore, we generate Gomory cut considering x,-row as follows: 1 2 FT ATP F_ Grsouve row) Since the coefficient of s, is positive, by applying rule (15) we get: ete fi Ai =003 22 ‘The coefiint of is also negative, and so by applying rule (15), we got: S134 + (ned integer cut 1 ‘Tous, Gomory’s mixed integer ext becomes Scan 3473 where 5, 8 Gomory’s stack variable By introducing this Gomory cut al the bottom of Table 7.19, we get @ new table, as shown in Table 7.20. > 1 1 ° o 0 tig, = Table 7.20 Optimal but Infoasiblo Solution 92 1 1 0 0 0 1 a ° 1 ° heer 1 1 cS 2 0 1 0 1 Table 7.21 & i : 2 2 Optimal Solution Z=2 G=% o o o 0 Since all ¢,~ 2; $ 0, therefore the required mixed integer optimal solution, given in Table 7.21, y= 0, y= 2,5, = and Max Z = 2Integer Linear Programming 7.6 BRANCH AND BOUND METHOD | ‘The branch and bound method was first developed by AH Land and AG Doig, and it was further studied by JDC Little ota, and other researchers, This method can be used to solve all-integer, mixed-integer and | zero-one linear programming problems | The concept behind this method is to divide the enti feasible solution space of an LP problem into | smalier parts called subprobiens and then evaluate corner (extreme) points of cach subproblem for an optimal solution. This approach is useful for solving LP problems where there isa large number of feasible solutions and so the enumeration tecomes economically impractical. The branch and bound method starts by imposing bounds on the value of objective function that help to determine the subproblem tobe eliminated from consideration when the optimal solution has been found. If the solution to a subproblem does not yield an optimal integer solution, a new subproblem is selected for branching. At a point where no more subproblem can be created, an optimal solution is arrived at. ‘Thebranch and bound method for the profit-maximization integer LP problem involves the following steps: The Steps of the Algorithm ‘Step 1: Initialization Consider the following all integer programming problem Maximize Z= cx, +¢,%; +... + Cy, subject to the constraints yy Hy Hayy Xp +... Hayy x, = by Sy Ey aa ea tk ag hy be (LP-A) ig, Ky # Ogg Bat ot Op and x, 20 and non-negative integers. Obtain the optimal solution of the given problem ignoring integer restriction on the variables. (If the solution to.this LP problem (say LP-A) is infeasible or unbounded, the solution to the given all-integer programming problem is also infeasible or unbounded, as the case may be. (Gi) If the solution satisfies the integer restrictions, the optimal integer solution has been obtained. If ‘one or more basic variables do not satisfy integer requirement, then go to Step 2. Let the optimal value of objective function of LP-A be Z,. This value provides an initial upper bound on objective function value and is denoted by 2 (Gi) Find a feasible solution by rounding off each variable value. The valve of objective function so obtained is used as a lower bound and is denoted by Z, Step 2: Branching step © Let x, be one basic variable which does not have an integer value and also has the largest fractional valu Branch (or partition) the LP-A into two new LP subproblems (also called nodes) based on integer values of x, that are immediately above and below its non-integer value. That is, i is partitioned by adding two matually exclusive constraints: SE) and x2 [4] +1 to the original LP problem. Here [x] is the integer portion of the current non-inieger value of the variable xy. This is obviously is done to exchide the non-integer value of the variable x,. The two new LP subproblems are as follows: LP Subproblem B LP Subproblem C Max z= 3, oxy ox subject to Ean, =), subject to Sax 6, S04) 2 Ly) +1 and x20, and 420. Branch and bound method breaks the {feasible solution region into smalior regions until an ‘optimal solution is obtainedOperations Research: Theory end Applications Fig. 7.4 ‘Step 3: Bound step Obtain the optimal solution of subproblems B and C. Let the optimal value of the ‘objective Function of LP-B be Z, and that of LP-C be Z,. The best integer solution value becomes the lower ‘bound on the integer LP problem objective function value (Initially tis is the rounded off value). Let the lower bound be denoted by Z,. ‘Step 4: Fathoming step Examine the solution of both LP-B and LP-C (@ [fa subproblem yields an infeasible solution then terminate the branch. Gi) If a subroblem yields a feasible solution but not an integer solution then retum to Step 2. (ii) fa subproblem yields a feasible integer solution, examine the value of the objective function. If this value is equal to the upper bound, an optimal solution has been reached. But if it is not equal to the upper bound but exceeds the lower bound, this value is considered as new upper bound and retum to Step 2, Finally, if it is less than the lower bound, terminate this branch. Step 5: Termination The procedure of branching and bounding continues until ro further sub-problem remains to be examined. At this stage, the integer solution corresponding to the current lower bound is the optimal all-integer programming problem solution. Romark The above algorithm can be represented by an enumeration tree. Each node in the tree represents a subproblem to be evaluated. Each branch of the tree creates a new constraint that is added to the original problem. Example 7.6 Solve the following all integer programming problem using the branch and bound method. Maximize 2 2x1 + 312 subject to the constraints @ bx + 5n< 25, @ m+3n<10 and sm & 0 and integers [Jammu Univ, BE (Mach) 2000) Solution Relaxing the integer conditions, the optimal non-integer solution to the given integer LP problem obtained by graphical method as shown in Fig. 7.4 is: x; = 1.92, x2 = 2.69 and max Z; = 11.91. The value of Z; represents initial lower bound es: Z, = 11. Selecting variable 2» for branching. Then to divide the given problem into two sub-problems and to eliminate the fractional part of x; = 2.69, two ‘new constraints 1» S 2 and x» 2 3 are created and added to the constraints of original LP problem as follows: LP Subsproblem B LP Sub-probiem C Max Z= 25; +3x2, Max 2= 2x1 + 3 subject to (@) 61 +5x £25, (i)xi+ 3x10, | subject wo @ 6r+5u<25, Gi)m +3xr<10 @) <2, Gi) 223, and x1 422 O integers. and Xi,x2 2 0 and integers. Sut-problem B and C are solved graphically. The solutions are: Sub-problem B : xj = 2.5, x» =2 and Max Z= I Sub-problem C:x, =|, x73 and Max Z= 11 Graphical solutions of sub-problem B and C are shown in Fig. 7:54) Opens soon s 75 1 2 Bs In the solution space of LP sub-problem C, variables x, and x2 are integers, so there is no need to branch this sub-problem further. The value Max Z;,~ 11 is a lower bound on the maximum value of Z for future solutions. 5 4 OG “einai sf a a ay 2 Mac 2-11 FA ® — optimal stuion eas @ ween haze 102 E % a ay Fig. 7 1." 3 4 § LP sub-problem 8 is further subdivided into two LP sub-problem D and E (shown in Fig. 7.6) by taking variable xj = 2.5. Adding two new constraints x; < 2 and.x; >3. Also Max Z = 11 is also not inferior to the Z, = 11. LP Sub-problem D LP Sub-problem E Max Z = 2x, + 3x, Max Z = 2x + 3x2 subject to @) 6x; + 5xz $25, (i) ay + 3xz < 10] subject to @ 6x, + Se $25, Gi) 1 + 3m < 10, Gi) 2s (wy) ns? (i) 222, (vy) 423 and u1,%2 2 0 and integers. and x12 2 O and integers. Sub-problems D and £ are solved graphically. The solutions as shown in Fig. 7.6 are Sub-problem D: x; =2,x2=2 and max Z,= 10 Sub-problem £ : xy ~ 3,49 ~ 14 and max Z5 ~ 10.2 ‘The solution of LP sub-problem D is integer feasible but is inferior to the best available solution of LP sub-problem £, Hence the value of lower bound Z; = 11 remains unchanged and sub-problem D is not considered for further division. Since the solution of sub-problem E is non-integer, it can be further branched with x; as the branching variable. But the value of its objective function (= 10.2) is inferior to the lower bound and hence this does not give a solution better than the one already obtained. The sub-problem £ is also not considered for further branching. Hence, the best available solution corresponding to sub-problem C is the integer optimalOperations Research: Theery and Applications solution: x; ~ 1,2 ~ 3. and Max Z~ 11 of the given integer LP problem, The entire branch and bound: procedure for the given problem is shown in Fig. 7.7. Fig. 7.7 ‘Complete Branch ‘and Bound Solution Example 7.7 Solve the following all integer programming problem using the branch and bound method. Maximize Z= 3x + subject to the constraints @ 2y +4525, Gi) <8, Gil) 2 $10 and Xp p20 and integers. Solution Relaxing the integer requirements, the optimal non-integer solution of the given integer LP problem obtained by the graphical method, as shown in Fig. 7.8, is: x, = 8, x, = 2.25 and Z, = 35.25. The value of Z, represents the initial upper bound, Zy = 35.25 on the value of the objective function. This ‘means that the value of the objective function in the subsequent steps should not exceed 35,25. The lower bound Z, = 34 is obiained by the rounded off solution values to x, = 8 and x, = 2. The variable x, (= 2.25) is the only non-integer solution value and is, therefore, selected for dividing the given LP-A problem into two subproblems LP-B and LP-C, % @21+4n525 @ x8 10. @ ms10 * i ® 6 . @ 2 Optimal non ineger ston BA (eB =225, 23525) 2 Feasibe Fig. 7.8 Regioe Graphical hi Solution of LP-A # 2 4 6 8 wb BR ow ® In order to eliminate the fractional part of 2.25 (the value of x,), two new constraints x, < 2 and x,2 3 are created. These two constraints are then added to the LP-A problem as shown below: LP Subproblem B LP Subproblem C Max Z= 3x, + Sr Max Z= 3x, + 5x3 subject to @ 2x, + 4x, £ 25, @ x, <8 _ | subject to @ 2x, + 4x, < 25, ii) 2x < 10 (redundant), iv) x, $2 fii) 2x, $10, @ %s8 (iv) x23 and xp X20 and integers. and x;,X 2 0 and integers.Intogar Linear Programming Onss % © 2+ 4n<25 # @ ms10 Problem C solution y= 65,45 45) 2 Problem B solution f= Ma=2,2,=24) ‘Subproblems B and C are solved graphically using Fig. 79. The solutions are: ‘Subproblem B : x, 2, and Max Z, Subproblem C : x, ~ 6.5, x; ~ 3, and Max Z;~ 34.5 Notice that both the solutions yield a value of Z that is lower than that of the original LP problem. The value of Z, establishes an upper bound on the values of the objective functions Z, and Z, Since the solution of the subproblem B is an all-integer, we stop the search of this subproblem, ic. no further branching is required from node B. The value of Z) ~ 34 becomes the new lower bound on the integer LP problem's optimal solution. A non-integer solution of subproblem C and also Z, > Z,, both indicate that further branching is necessary fiom node C. However, if Z, would have been < Z, then no further branching would have been possible, even from node C. The second upper bound takes on the value 34,5 instead of 35.25 at node A. The subproblem C is now branched into two new svbproblems: D and B, the valve of which are obtained by adding constraints x, $6 and x, 2 7 (x, is the only non-integer-valued variable in the subproblem C solution). The two subproblems are stated as follows: LP Subproblem D LP Subproblem E Max Z= 3x, + Sx Max Z= 3x, + 51 subject to () 2x, +41, $25, (ii) x $8 (redundant) | subject to (2x, +4r,<25, ii) x) 58 i) 2x < 10, iv) 3323, () =) 56 Gi) 2x £10, Gv) 33, and Xp % 2 0 and integers () x27 and ay, %; 2 O and integers. Subproblems D and E are solved graphically using Fig, 7.10. The solutions are: Subproblem D : x, = 6, x, = 3.25 and Max Z, = 34.25. Subproblem E.; No feasible solution exists because constraints x, 2 7 and x, 2 3 do not satisfy the first constraint, So this branch is terminated b Omrms2 Ouss Oms0 ‘f o| © @x O23 Ouse 4 fe) @x27 Fig. 7.9 Graphical Solution of Problems B andC Fig. 7.10 Graphical Solution of Problems Dand EOperations Research: Theory and Applications The non-integer solution obtained at sub-problem D yields an upper bound of 34.25 instead of 34.50 and also greater than Z, (an upper bound for sub-problem B). Once again we create su-problems F and G from sub-problem D with two new constraints x5 <3 and 2 4, as shown in Fig. 7.6 LP Subproblem F LP Subproblem G Max Z = 34, + 5x Max Z = 3x, + 53 subject to @ 2x, + 4x, $25 subject to @ 2x, + 4x, £25, (ii) x, £8 Gi x, $8; iii) 2x, $ 10 (redundant) ii) 2x,£10, (iv) x, 3 (redundant) fy) 23, W)x,S6 (<3 W456 Wi) Qed and ps % 2 0 and integers and jy 20 and integers The graphical solution to subproblems F and G as shown in Fig. 7.11 is as follows: oes 2 Sorento See Seta ots Ete eos 4 @ mrss Ouse @ 2s10 10} Pn Gin @ as2 @xur3 @ nee 42553"S) @ uz? @uss Ours > \ CQ) VHT 5 ont Fig. 7.11 4 are tt —,-® ses : c SS a oe le - A o 2% 6 8 oO Rw ‘The branching process is terminated when new upper bound is less than or equal to the lower bounds of previous solutions or when no further branching is possible. Although the solution et node G is non-integer, ‘no additional branching is required from this node because Z, 4 maha 328 has % oe a 3. Re c » ‘sy, m= 55.=3 G ivi yams pereey - + 4-5 Complete Branch a and Bound bo olution ‘ Infeasile Example 7.8 Solve the following all-integer programming problem using the branch and bound methed Minimize Z= 3x, +2.5x, subject to the constraints @ x,42e,2 20, (ii) 3x, 2x, 250 and Xp Fp 20 and integers.Integer Linear Programming Pa) Solution Relaxing the integer requirements, the optimal non-integer solution of the given integer LP problem, obtained by the graphical method, is: x, ~ 15,x, ~ 2.5 and Z, ~ $1.25. This value of Z, represents the initial lower bound, Z, = 51.25 on the value of the objective function, ic. the value of the objective function in the subsequent steps cannot be less than 51.25. The variable x, (2.5) is the only non-integer solution value and is therefore io selected for dividing the given problem into two subproblems ~ B and C. In order to eliminate the fractional part of 2.5, two new constraints x5 $2 and x, > 3 are created. These two constraints represent the two new parts of the problem as shown below: LP Subproblem B LP Subproblem C Max Z= 3x, + 2.51, Max Z= 3x, + 2.5%, subject 10 @ x, +2x)220, (il) 3x, +2x,250| subject to @ xj +2r220, (ii) 3x, +2250 ) x,$2 Gi) 423 +X) 20 and integers. and jv 8p 2 O and integers, and Subproblems B and C are solved graphically. The solutions are: Subproblem B:x,=16, x =2 and Min Z, = 53. Subproblem C : x, ~ 14,66, x5 = 3 and Min Z, = 51.5. Since the solution of subproblem B is all-integer, we stop the search of this subproblem, ie. no further ‘branching is required from the node B. The value of Z,= 53 becomes the new lower bound. A non-integer solution of sub-problem C and also Z, < Z, indicates that further branching is necessary from node C. However, if Z, woald have been 2 Z,, then no further branching would have been possible from node C. ‘The second lower bound takes on the value Z, = 51.5 instead of Z, = 51.25 at node A. ‘We divide subproblem C into two new subproblems: D and E which are obtained by adding constraints 4, 4 [4 and x, 2 15. The two subproblems are stated as follows: LP Subproblem D LP Subproblem E Max Z= 31, 2.51) Max 2 31, + 2.51) subject to () x, +2x220, Gil) 3x, #2x 250] subject to @ x,+2x220, Gi) 3x +2e,250 Gi) x, 23, (iv) x, 514 () 23, () 4215 and xj.) 2 O and integers. and yy %, 2 0 and integers. Subproblems D and E are solved graphically. The solutions are: Subproblem D: x, = 14, x5 = 4 and Min Z, = 52. Subproblem E: x, ~ 15,13 ~3 and Min 2, ~ 52.5. ‘The solutions obtained at node D and E are beth all-integer and therefore branch and bound algorithm is terminated. The optimal integer solution to the given LP problem is at node D, where the value of the objective function is the lowest amongt the values at nodes B, D and E. ‘The entire branch and bound procedure for the given problem is shown in Fig. 7.13, B xy 216.2 423 6 m= Opimat a Z%-92 | Sounion y= 15.39 = 25 so z= 5125 es) a2} c 213 E at 25-928 Fig. 7.13 Completa Branch and Bound SolutionOperations Research: Theory and Applications, 7.7 APPLICATIONS OF ZERO-ONE INTEGER PROGRAMMING A large number of real-world problems such as capital budgeting problem, matching problem, sequencing problem, scheduling problem, location problem, travelling salesman problem, etc., require all or some of the Gecision variabies to assume the value of either zero or one. A few such problems are discussed here: ‘The zero-one integer programming problem can be stated as: Minimize Z= § ej, re subject to the constrains Bia) 205 1=1,2,--0m and x)= O or 1. 7.7.4 Capital Budgeting Problem Such problems cover the problem of allocating limited funds to various investment projects in order to maximize the discounted net retum. Example 7.9 A corporation is considering four possible investment opportunities. The following table presents information about the investment (in Rs thousand) profits: Fae Present Vals of Capital Required Vear=vine by Projects a Year 1 Year 2 Year 3. 1 6300 700 330 00 2 75900 350 $50 350 3 2250 300 1530 100 4 3500 350 200 _ Capital evailble for investment 1,200 200 400 In addition, projects | and 2 ere mutuelly exclusive and project 4 is contingent on the prior acceptance of project 3. Formulate an integer programming model to determine which projects should be accepted and ‘which should be rejected in order to maximize the present value from the accepted projects. [Delhi Univ,, MBA, 1998, 2004) Model formulation Let 1 if project jisaccepted 1 0 it project jis rejected Integer LP mode! Maximize (Total present valus) Z~ 6,500x, + 7000s, + 2,250x, + 2,500x, subject to the constrainis (i) Expenditure in years 1, 2 and 3 (@ 700s, + 850x, +300rs +350, < 1,200 (i) $50x, + 550x, + 150r, +200x, < 700 (i), 400x, + 350, + 100r, $400 @) my tmeh Wy-ysl and 4) = 0o0r L 7.7.2 Fixed Cost (or Charge) Problem Incertain cases, while undertaking a particular set of activities, the fixed costs (fixed charge or setup costs) are incurred. In such cases, the objective is to minimize the total cost (sum of fixed and variable costs) associated with an activity Let us define the following decision variables 1x, = level of activity j F, = fixed cost associated with activity xj > 0 atiable cost associated with activity x > 0Integer Linear Progamming Then the general fixed cost problem can be stated as: Minimize Z= $ (c;x;+ Fjy;) subject to the constraints Bayt) 0 and inogors Nax 2 = 51, + 4x, sibiec to) + EA (HSK, 43% 5 15 Gi) Ox, + Bmp 5 18 and ys %p2 0 and intogers 5. Max Z=-34, +X + 9% siblect 0) x, +#2y +H S4, (I 2R-15y <1 Gi) 5 9m + By 59 and % % 2 0; % nonnegative integer. 6 Max Zax sibject tof) 2x, + 5x2 16, (i) Bx, + Sx $30 and 20 2 pornegalive integer. 7. Min Z= 4x, + 85 + 5x sibject to) 2h, ~ 2x 4+ 427, () 24, + 6m -2y25 and X20: x, 5 nonnegative integers: 8 Max Z= 110%, + 100%, subject to) 6x, + 5x 5 28 (44, + hy 2 48 ard 2X % = 0 and integers.Integer Linear Programming 9. firm is considering investing in plant modernization and plant expansion. All of these proposed projects would be completed within 2 yoaro, with varying rogquiroments of money ard plant ‘enginearing. The management is willag fo use the folowng data in electing the best set of proposals. Three resource limitations ‘The cost to buy the products are : Product ee ie 3 4 Costunit (in Fs): 310 260 475 «225 Formulate this problem as a zeroone integer programming ae pret {Dat Un, MBA, 189%, 2000] Fs year exendture Rs 4.20000 Second year expendive : Re 440,000 Engineering hours 15,000 hours | inchs Machion: a8 ¢ oO © F Projet Dasrotin —Etpenaire a Pasert Eee +004 002 092 0 008 008 8 ae 2 0 O01 095 015 0.08 0.08 ie year andar 3 002 000 0 008 9.02 Oe ae a oe 4 008 00s ors 0 0005 Sider ste tor ta) B® Eaapment fnew 11, ABC Censtston Company i facing fe pobl of determining momonne = 8 whch pojecs shoul underaio over the next yeas. The «eens menace a 3 nlomation about es! “en oo ow ele folowing table Qs Ifrmation about each projet Prat sibconact ese ota | hae eames Zope Rapuara processing plant 195, ° 7 2 o_o _ ee uy els Bd a 3,00,000 400,c00 4.00.00 500,000 conanas © «2 « ° 6 120.000 80,000 | 0 0,000 dinate ease en | @ 300,000 2.00,c00 2.00000. 200,000 Bete ee eee |e 200,000 400,000 4,00000_1.00,000 ‘the mew shop floor. The company may not want to buy or build Fund available 6,50,000 800,000 8,00,000 5,090,000 raw materia prozessing faites. Formulate the given problem 2s an integer linear programming problem in erder to maximize fh pal present value or Hie trocor Formulate a zero-one progremming model » maximize estmated 10. ABC Manufacturing company faces the folowing problem: Should they make or buy each of Mer soveral producie? The company | plicies specify that they will ether make or buy the whole it [Dein Urw., MBA, 2000) 12, Wrte constraints to satisty each of the following conditions in 1 projec selection model. The prejects are numbered 1, 2, 3, ‘of each product ina complete lol. The company hes fcur ‘to. products to make or buy with six machines involved in making () Exactly one project from the set (1,2, 9) must be selected, these products, If they are made In the shop. The tme per unit i) Project 2 can be selected only ‘number 10 © setcted, (ie hours) required are as follows: However, 10 can be selected without 2 being selected Forty Tours are avaiable on each machine. One hundred ten (i) No more then one project rom the set (1, 3, 5, 7, 8) can Units of each product are needed. The costs of makng the be selected products are. listed. below: (iv) number 4 is selected, then number 8 cannot be seecte. dacs d 4 ee (W) Projects 4 and 10 must both be selected or beth be ig rected. [Oehi Univ, MBA, 1998) Cosiunit(in Rs); 225 2.22 460 1.90 HINTS AND ANSWERS 2,4) = 2 and Mex Z= 14 x, ~ Sy > 0 and Max 2 ~ 35 87, x5 © 1 and Max Z = 29/7 = 1.25, x, = 1 and Min z= 20.75 x, = 0.x, =3 and Max Z= 9 xy —3.ay = 0 and Max Z = 15 x, = 4. = 6/5 and Max Z = 2615 y= 4.x = 1, Max Z = 540 CHAPTER SUMMARY {In this chapter, important extension of linear programming, referred to as integer linear programming, was introduced where one ‘or more of the variables must be an integer. Ifall variables of a problem are integers, then such problems are referred to es all- integer linear programming problems. If some, but not necessarily all, variables are integers, then such problems are referred to as mixed integer linear programming problems. Most integer programming applicetions involve 0-1 or binary variables, ‘The number of applications of integer linear programming continues to grow rapidly due to the availabilty of integer linear programming software packages. The study of integer linear programming is important for two major reasons: (i) integer linear programming may be helpful when fractional values for the variables are not permitted and rounding off their values may not provide an optimalimage not availableimage not available(Operations Research: Theory and Applications Bharat Chand & Co. also provides 1,000 units of steel free of cost if at least 1,500 units of steel and 1,400 units of aluminium are purchased from this vendor. Seth & Co. has a minimum quality standard of 98 per cent for steel and 95 per cent for aluminium. In addition, ithas amaximum acceptable detivery standard (in days) of steel as 7 and aluminium as 7.5. The total requirements for the coming month is 3,000 units of steel and 3,000 units of aluminium, Questions fordiscussion Develop a mathematical model and specify the company’s production schedule Akko advice the company to choose which seller it should purchase its two raw materials from and the quantity ‘of the raw materials that should be purchased Case 7.2: Apparel Industry odiacis well known brand inthe fashion industry. It manufactures different types of shits indifferent sizes, ‘as wellas ties and other fashion accessories for upwardly mobile urban professionals. Recently, it decided to make use of some latest techniques in order to help itselfin the cloth eutting operatien The cutting operation for shits is extremely time-consuming and gives rise to high set-up costs. The problem that the company was mainly facing was that a large amount of cloth were being wasted in the cutting ‘operation. The process of cutting involved putting several layers of cloth of standard width on a table and. putting stencils, i. the templates in order to cut the cloth, However, the choice of the templates was based on judgement, which was leadingto a lot of wastage. To tackle the problem it was given tothe R&D department of the company. The objective facing them was to minimize the setup cost and excess, subject to certain con- straints, The following data was available regarding the cutting operation of a shirt whose demand for different sizes is given as follows Sie: 3B. 40 a 44 46 Demand : 54 a 91 60 29 ‘The numbers of ayers that can be cu is limited by the length ofthe knives and the thickness ofthe fabric. A second restriction of the cutting process is introduced by the length of the cutting table, which limits the ‘number of stencils that can be cut in one operation. As the length ofthe stencils for the different sizes is almost ‘equal, the maximum number of stencils on the cutting table is actually independent of the combination of the stencils used. Thus the R&D Department assumed that all stencils have equal length without loss of generality. ‘The spreading of the fabric on the cutting table, the fixing of the ayers and stencils, and the cutting itself are extremely delicate and time-consuming operations. Therefore, the aumber of these operations should be ‘minimized. The problem then consists of finéing the optimal combination ofthe number of layers of cloth on the cutting table andthe associated set of stencils that wll result in the minimum numter of setups, whilesatisfying the demand, with litte or no exe ‘There was an upper bound on the number of layers at 35 and the cutting table length could hold at most four stencils. Three cutting patterns were to be used. Questions for discussion Develop an appropriate mathematical model and solve it so that it helps the management achieve its objective.chapter Goal Programming “In the end. all business operations cam bo reduced to three words: people, product and profs. Unless you've got a good team, you can? do mach with the other m0.” ~ Lee lacoosa Preview Goal programming is an approach used for solving any ‘multi-objective optimization problem that balances trade-off in ‘multiple and often conflicting incommensurable goals at different priority levels. Learning Objectives After studying this chapter, you should be able te ‘+ appreciate the nced of a goal programming approach for solving multi-objective decision problems. + distinguish between LP and GP approaches for solving a business decision problem. ee ‘+ formulate GP model of the given multi-objective decision Lae Aaa a , 83. Concept of Goal Programming + understand the method of assigning different rinks and 24 Go\prrymmng eel Exmnaon eee eons! oils He 85. Graphical Souton Method for Goal Programming ‘use both graphical and simplex method for tolving 2 GP 9.6. modified Simplex Method of Ge! Programming model 87. Aternative Simplex Method for Goal Programming + Conceptual Questions + Sait Practice Problems Chapter Summery Chapter Concepts Quiz D Case study‘Operations Research: Theory and Applications GP approach establishes a ‘specitic numeric ‘goal for each of the ‘bectve and then attempts 10 achieve ‘each ‘geal ‘sequentially upto a ‘salistaciory level rather than an ‘optimal level 8.4 INTRODUCTION In previous chepters, linear programming models were formulated and solved in order to optimize an objective function value a single measure of effectiveness) under a set of constraints. However, the optimization of such a single objective function is often not representative of the reality due to divergent and conflicting interests and objectives (economic as well as non-economic) of any business, service or ‘commercial organization, Consequently there arises a need to attain a ‘seistactory” level of achievement amongst multiple and conflicting interests oF goals of an organization or a decision-maker Goal programming (GP) is an approach used for solving a multi-objective optimization problem that balances trade-off in conflicting objectives. In other words, it is an approach of deriving a best possible “satisfactory” level of goal attainment. A problem is modelled into a GP model in « manner similar to that of an LP model. However, the GP mod:! accommodates multiple, and often conflicting, incommen- surable (dimension of goals and unit of measurement may not be same) gcals, in a particular priority order (hicrarchy). A particular priority structure is established by ranking ond weighing various goals and their ssubgoals, in accordance with heir importance. The priority structure helps to deal with all goals (cbjectives) that cannot be completely and/or simultancously achieved, in such a manner that more important goals are achieved first, at the expense of the less important ones. 8.2 DIFFERENCE BETWEEN LP AND GP APPROACH Linear programming has two major limitations from its application point of view: i) single objective function, and (ii) same unit of measurement of various resources. ‘The LP model has a single objective function to be optimized such as profit maximization, cost ‘minimization, etc. However, in actual practice, the decision-maker may not be satisfied with a single objective. That is, he may desire to get simultaneous solution to a complex system of competing objectives. The solution of any LP problem is based on the cardinal value (the number that expresses exact amount such as, 1, 2, 3, ...) such as profit or cost, whereas that of a GP allows ordinal ranking of goals in terms of their contribution or importance to the organization. Since it may not be possible to obtain information about the value or cost of a goal (the specific numerical target value desired to achieve) or a sub-goal, therefore their upper and lower limits are determined. Usually, desired goals are assigned priorities and then these priorities are ranked in an ordinal sequence. ‘Whenever there are multiple incommensurable (different units of measurement) goals, an LP incorpo- rates only one of these goals in the objective function and treats the remaining goels as constraints. Since the optimal solution must satisfy all the constraints, this implies that (a) the several goals within the constraining equations are of equal importance, and (b) these goals have absolute priority over the goal incorporated into the objective function. 8.3 CONCEPT OF GOAL PROGRAMMING ‘The concept of GP was introduced by Chennes and Cooper (1961). They suggested a method for solving an infeasible LP problem arising from various conflicting resource constraints (goals). A few examples of ‘multiple conflicting goals are: (i) maximize profit and increase wages paid to employees, (ii) upgrade product ‘quality and reduce product cost, (ii) reduce credit losses and increase sales. Ijiri (1965) developed the concept of pre-emptive priority factors, assigning different priority levels to incommensurable goals and different weights to the goals atthe same priority level. Lee (1972) and Ignizio (1976) have written textbooks on the subject of goal programming. Goal programming has been applied to a wide range of planning, resource allocation, policy analysis and functional management problems. An important feature of GP is that the goals (a specific numerical target values that the decision- ‘maker would ideally like 10 achieve) are satisfied in ordinal sequence. That is, the solution of the GP problem involves achieving some higher order (or priority) goals first, before the lower order goals are considered. Since itis not possible to achieve every goal (objective), to the extent desired by the decision- ‘maker, attempts are made to achieve each goal sequentially rather than simultaneously, up t0 a satisfactory level rather than an optimal level In GP, instead of trying to minimize or maximize the objective function directly, as in the case of an LLP, the deviations from established goals within the given set of constrainis are minimized. In the simplex algorithm of linear programming such deviational variables are called slack variables and they are used onlyGoal Programming a dummy variables. In GP, these slack variables take on a new significance. The deviational variables are represented in two dimensions — both positive and negative deviations from each goal and subgoal. These 0, at acertain priority level, and there is no negative entry at higher tnachieved priority levels, “in the same column, the current solution is not optimal. if ha ngat van ofeach geal tre gan Xx tha aro i op A, ibs eto i ew tn a mt cag (i aeaghik Piiy (educa inner acpaive value Otherwise cibvto he sxt ghar ty (Pp aad arias topes sags oe 5. Apply te unl procedure forcatelatng the “ninimum rt! to choose a vrshle that need to ene ‘the current solution mix (basis). Ce Ti lgoee Tha i beorss A viata fre te Rigas ploy goal would be ncosnd wi he entry of this variable in the solution mix. Example 8.6 Use modified simplex method to solve the following GP problem. Minimize Z = Rdv + (2d; +d5)+ Bay subject to the constraints @ xj+x, +47 -df=400, (il) x, +d;=240, ii) x, +d; =300 and Nyt d diddy 20 Solution The initial simplex table for this problem is presented in Table 8.1. The basic assumption in formulating the initial table of the GP problem is the same as that of the LP problem. In goalGoal Programming Be programming, the pre-emptive priority factors and differential weights correspond to the c, values in linear programming, 9° 0 Oo PF P,P Py PB 400 Bibrcreden (Se es In Tabie 8.1, the optimality criterion (c,—z) is 3 * 6 matrix because we have three priority levels and six vatiables (2 decision, 4 deviational) in the model. By using the standard simplex method for the calculation of Z-value, we would obtain Z-value in GP as: Z= P, x 400 +2P, x 240 +P, 300 = 400P, + 780P, 780 and P, = 0 in the xy-column below the line represent the unachieved portion of cach goal Now let ws calculate oz, ves in Table 8.1. We have already sid that values represent the priority factors assigned to deviationel variables and that the z, values represent the sum of the product of entries in eycolumn with columns of coefficient matrix. Thus, the ¢; ~z; value for each column is calculated as follows: ey-2,=0 -(P, «1 +27) x1 +P, x0) e—2 20 —(P, 1+ 2, x0+P, x 1)= G— HPP DA PHP, Iteration 1: The selection of key column is based on the per unit contribution rate of exch variable in achieving the most important goal (P,). Tae pre-emptive priority factors ae listed from the lowest to the highest so that the key column can be essily identified at the bottom of the table. ‘The column with the laigest negative ¢) ~ z; value at the F, level is selected as the key column. In Table 8.1, there are negative values (.c. ~ {) in'the x, and x, columns. Remove this tie, as always, and choose variable x, to enter into the new solution mix. The key row is the row with the minimum non-negative value, which is obtained by dividing the xy-values by the corresponding positive coefficients in the key column. The coefficient 1 is circled in ‘able 8.1 to indicate the fact that itis the key clement at the intersection of the key column and key row. P,-2P, A-P, By using the standard simplex method, the solution in Table 8.1 is revised to obtain the second improved solution, as shown in Table 8.2 > 0 0 PR 2 n a 160 Of yp ove ee oie ats ReneS ° rs 240 ep oils ens P, a 300 ON Gane nae ad 10 | 3001 > ° = St Ree errr 300 eee Sue aaa a A 160 snes espe Table 8.1 Initial Selution‘Operations Research: Theory and Applications Asper Table 8.2, the value of objective function: 160«P, ~ 300P. indicates thatthe unachieved portion of the first and second goals has decreased. The revised solution is shown in Table 8.3. Since the solution in Table 8.3 indicates that c,~ z values in P,-row are either positive or zer0 as well as the value of Z in terms of P, is completely minimized to zero, we tur our attention to the second priority level (P,). An additional rule must be followed at this stage: 4 column cannot be chosen as the Key column wih & poatioe Walia ae Ea be: priein: lee 400 240 300 : 2 ° 2 o9s eae ee | sage 40 Sie iia a oe ; 0 is les 1 Oye | Table 8.3 at ‘The largest negative value in P,-row is selected in order to determine the key column, The revised solution is shown in Table 8.4. In Table 8.4, all c, ~z; values in the P,-row are either positive or zero. Thus, the second goal (P;) is {ally achieved. It may be noted in Table 8.4 that there are two negative values in the P,-row. However, we could not choose dz or ds 2s the key column because there is already a positive value at a higher priority level (P,). Hence, the solution shown in Table 8.4 cannot be improved further. The optimal solution therefore 4) = 240, x5 = 300, dy “> o o P, Ww, bP ° ae 300 1 1 0 ADAMO eae 240 0 1 oo eae 140 ° ° =i 1 {Bese Py 140 = = 1 =i anaes aac toes o 2 a ° ens Table 8.4 0 a E 1 o Ofc 8.7 ALTERNATIVE SIMPLEX METHOD FOR GOAL PROGRAMMING’ This proposed procedure is based on Baumol’s simplex method for solving GP problems but is with minor modifications. It sppears to be more efficient than the one suggested by Lee. The steps of the procedure aresummarized below: Step 1: Formulation of initial solution table The generalized formulation of a GP problem into an initial solution table begins in a similar fashion to the LP problem table. First, the goal constraints are reformulated in terms of their d;* variables (i.e. basic variables): f= +S ayy edz; i= 1,2 If-2 goal constraint does not possess a d;* variable, itis artificially assigned with a zero priority for the sake of the initial table formulation, In Table 8.5 an initial table is presented. Row 1 in this table labelsGoal Programming Pagal the decision variables x, and negetive deviational variable d>. These variables are zeto (noa-basic) variables, The right-hand-values,b, are placed in column 3; the decision-variable coefficients a, are placed in column 4; and an identity matrix is placed in column 5, representing the inclusion of negative deviational variebles d; .In problems where a goal constraint does not have a d; variable, the value of ‘zero’ would be placed in that column instead of ‘one’. Mathematically, this requireds that all deviational variables, not included in the problem formulation, are equal to zero. Column 1 in the table lists the appropriate priority factors P, and weights w, for each positive deviational variable (ie. basic veriabes), including artificial deviational variables as shown in column 2. Row 2, column 3, in the table contains a value called “iotal absolute deviation’. It represents the amount of total deviations from all goals for each table asthe iterative process proceeds. Row 2, column 4, is simply a row vector of zero, representing the inclusion of all decision variebles in the computational process. Row 2, column 5, lists the appropriate weights w, for each negative deviational varable included in the objective function. «. Q wo é 2 Table 8) Inital Table for a Generalized GP Model Step 2: Modify the initial solution (@) Determine which variable is to exit the solution basis. This is accomplished by selecting the variable with the highest-ranked priority. When two or more variables have the same priority ranking, the variable with the greatest mathematical weight determines which variable isto be selected first. When ‘wo or more variables have the same weighted priority level, the selection of the variable that has the greatest negative right-hand side value is used as the selection criterion. Its selection will eliminate the ‘most negative values ffom the solution basis, thereby completing, quickly, the required computations {or an optimal solution. This variable labels the pivot rov. (0) Determine which variable i to enter the solution basis. This is accomplished by selecting the column that has the smallest resulting ratio when the positive coefficients in the pivot row are divided into their respective positive elements in row 2. This variable labels the pivot column. The element at the intersection of the pivot row and pivot column is called the pivot element If there isa tie in the column ratio, select the column that has the smaller ratio in the next priority row. Ifthe tie cannot be broken by examining the next priority, then the selection of the pivot column would be based on the largest coefficient in the pivot row for the tied columns. (Note that artificial variables are not allowed to re enter the solution basis once they have been removed.) (©) Seta framework for the new table by exchanging the variables inthe pivot row and pivot columa. Also, the priority attached to the variable brought into the solution basis should be placed in column 1 of the new table. All other variables resume their places in the new table. (@ Im the new table, the new element that corresponds to the pivot element is found by taking the reciprocal of the pivot element. All other elements in the row are found by dividing the pivot-row elements by the pivot cloment and changing the resulting sign. (€) Determine the new elements thet corespond to the elements in the pivot column. These elements are found by dividing the pivot-column elements by the pivot element. (Except for the clement representing total absolute deviation (j.. column 3, row 2), all other elements are found by the following formula: Product of Two Comer Elements New element = Old Element — me * Based on, Schniederjans, Mare J. and N.K. Kwak, An Alternative Solution Method for Goal Programming Problems, J. Opns. Res., Vol. 33, 1982,EEA Operations Research: Theory and Applications The product of two corer elements is found by selecting elements out of the pivot row and pivot column, () Determine the new total absolute deviation by the following formula: Z Ew, -b)| (b) Check to see if the solution is optimal. The solution is optimal if all the basic variables are positive (i. positive b,) and if the pre-emptive priority rule is satisfied. If one or more of the basic variables are negative, repeat Steps 1 to 8. Ifall the basic variables are positive but the pre-emptive priority rule is not satisfied, the solution is not optimal. Continue to Step (i). (@ Determine which variable is to exit the solution basis. This is accomplished by selecting the largest positive element in column 3, with the highest priority level. The pivot element comes from this row. (@ Determine which variable is to enter the sclution basis. This is accomplished by selecting the column that has the stnallest resulting ratio wiven the negative coefficients in the pivot row are divided into their respective positive elements in row 2, changing the resulting sign. This variable labels the pivot ‘column. Repeat Steps (c) to (h). (The solution is optimal ifthe basic variables are all positive and one or more of the objective function rows (ie. row 2) have a negative sign. Example 8.7 Solve the following GP problem Minimize 2= Pydy + Pydf + SPydy +3Pyd3 + Pydit subject to the constraints Ox +H+ 4 = dj = 80, @ x,+ 4; =70 (ii) x + ds = 45, (i) x) ty t dz and Using the steps ofthe method that have already been discussed, the wiven GP problem data is displayed in the initial Table 86. SER eg Ye Ra aah IC Ie OORT 0% leg Ob, 0 ae Table 8.6 0% ay SAA tan BO gh Ren Intial Solution PB Heol ea wih 02 Sib aaa In the simplex Table 8.6, zero elements in the Z row, (¢, and x, columns) are treated as very small numbers for a pivot-column selection. . Py de a0, 1 1 1 2 o 0 Pe TBA take opel shape Daccis delle NG, OP Bp es Ore die 0 Ores ew aebeeee Py cap a0 1 Gobel 1Goal Programming Table 8.9 Be ey 1 Optimal Solution {t can be seen thatthe Table 8.9 presents an optimal solution, es columa 3 contains all positive values (ie. positive basie variables) and row 2 has a negative value in column d/'. The solution can be read from the solution bas in a total absolute deviation from desired goals of Z x, = 70, x) =20, d;* = 10 and ds = 25. This results, 5. Examining the weighted priority column, it can be seen thatthe higher level priorities P, and P, are now fully sstisfied as they do not eppeer in this column, CONCEPTUAL QUESTIONS 11. What is goal programming? Cleary state its assumptions, 2. (a) Compare the silferencee/simiartics betwaen linear pro- gramming and goal programming. (b) Wry are al! goa tinear programming problems minimization problems? Can a goal pregramming problem be irfeasble? Discuss. [Delhi Univ., MBA, 1999, 2003} 3. Explain the following terme (0) Doviatonal variables (i) Pre-enptve prenty ‘actors 4. Explsn the dlfference between cardinal value and ordinal value. 5. Under what cicumstances can cardinal weights be used in the abjecive function of a goal progranming model? What happens 1 the cardinal weighs are attached io all riots in the objective function of a goal programming moder? 6. (a) Explein the diterences between solving a linear proaram- ming as agains! a goal programming problem by the simplex ‘method (b) What conditions require that a GP model rather than an LP model be used to solve a decision preblam SELF PI 1. Salve the following goal programing problems using both the 1raptical method and the simplex method to oblain the solution (a) Minimize 2= Pid + Pads yds X+m%t dy-dj =40 Nyt dpe = 20 Ny ker GF F520, forall subject to (b) Minimize 2= Pd + PylBd,+ 6d) + Py sujet to +x + df= = 16 “ ae deg and Mite Fs 20, forall 7. ‘State some problem areas in management where goal program- ‘ming might be applicable. 8. (a) Why a the derivational variables associated with a partcu- lar goa! complementary? (&) What ‘is the diference between a postive and negaive ‘deviauonal variable? 8. ‘Goal programming appears to be the most appropriate, flexble and poweral technique for complex decision problems vohing mutiple confictng cbjectves.’ Discuss. [Dehi Univ, MBA, 1999, 2000) 10. What Is goal programming? Why are all goal progamming Froblems minimization problems? Why does altering the goal Portes result na diferent solulon to a problem? Expl. ‘11. What is meant by the terms ‘satisfying’ ard wiv is the term otten Used in conjunction with goal programming? 42, What are daviational variables? How do fey difer from decision Variabies in tradtional linear programming problems? 13, What does to renk goals’ mean in goal programming? How does this affect the problem's solution? \CTICE PROBLEMS (0) Minimize 2 Pd; + Pes + Py subject to 2x, + 3% 530 6x, 44,560 mrad a8 stds a5 =7 and Me dps 20, forall 2. An ofice equipment manufacturer produces wo kinds of prod- ucts, chairs and lamps. The production of ether, a chair or a lamp, requires 1 hour of production capacity in the plant. The Pant has a maximum production capacity of 10 hours per wet Because of the limited sales capacty, the maximum number of ‘aire and lamps that can bo cod aro 6 and 8 por wook, respectively. The gross margin from the sale of a chair is Rs 0 and from the sale of a lamp ts Re 40. ‘The plant manager has set the following goals, arranged in cerder of importance:Operations Research: Theory and Applications () He wants to avoid any undorutitzation of production pacity. (i) He wants o sel as many chairs and lamps 2s possible. Since the gross margin trom the sale of a chair is set al twibe tre amount of prot from a lamp, he has twice as much desire to achieve the sales goal for cheirs as tor lamps. (W)_ He wants 'o minimiza the overime operation of the plant as mich as possibe. Formulate and solve this problem as a GP problem, so thet the plant manager makes a decision that will help hm achieve his goals as closely as possible. [Delhi Univ, MBA, 1999) ‘A production manager faces the frobiem of ob allocation emong three of his teams. The processing rales ofthe three teams are 5, 6 and 8 units per hour, respectively. The nonral working hhours for each team is § hours per day. The production ‘manager has the following goals forthe next day 1 onder of pron (0) The manager wants to avoid any underachievement of production level, which is set at 180 units of product (Any overtime operaton of team 2 beyond two hours and ‘eam 3 beyond three hours should be avoided. (w) Minimize te sum ct overtime Formulate and solve this protem as a coal proaremmina preblem. 1. XYZ company produces two products ~ record players and tape-rocerdere. Both the products aro produced in two eoparato machine centes within the plant. Each record player requires {wo hours in mactine centre A and one hour in machine centre BB, Each taperecoer, on the other hand, requres ene hour in the machine contre A’and three hours in'machine centre 8. in dation, each product requires some inprocess inventory. The runt! The in-process. inventory required ts Fis 30. for ecard. layer and Rs 30 for the tape-recorde. Tho fm has ‘rma monthiy operation hours of 120 for both machne contres, ‘A and B. The estimated prot per unit is. Rs 100 forthe record player and Rs 75 for the tape-recorder. According to the ‘matketing deoartment, the forecast of sales for the record player and the tape-ecorder are 50 and 80, respectively forthe oming: mont The presicent ofthe firm has established the flloning goals for prodtction in the next month, in ordinal rank of importance: ()Liit the emount ied up in inprocess inventory forthe month to Ms 4.000, (i), Achieve the sales goa of 80 tape-recorders forthe month. (Gi) Umit the evertine eperation ef machine centre A 10 20, hours. (iv) Achieve the sales goal ot 50 record players forthe month. (¥) Limit the sum of overtime operation for both machine centres. Avoid any underutiization of reguler operation hours of both mactine centre. Formulate anc solve this protiem as a goal proaramming preblem ABC Fumitures produce three products ~ tables, desks and chaire. The luriture ie produced In the certral plant. The production of the desk requires & hours in the plant the table 2 hours and the chair only 1 hour. The regular plant capacity ‘is 40 hours a week. According tothe marketing department, the ‘maximum number ot desks, tables and chairs that can be so! \weekty are 10, 10 and 12, respectively. The president of tho fm has establishes the folowing goals, according to heir importance (i) Avoid any underutiiation of production capacity (i) Meet the order of XYZ Store for saven desks ard five chairs (i) Avoid overtime operation of the plant beyond 10 hours. {v) Achiove alos goal of 10 docks 10 ables, and 12 chairs (v) Minimize evertine operation 2s mach as possible. Formulate and solve this protiem as a goal programmirg preblem [Bern Univ, MEA, 1998] wi) | 6. ABC Computor Compsny produces three ifleront types of comouters — Epic, Galaxie and Utopia. The production of all ‘computers is conducted in a conplee and medem assembly line. The production of an Epic requires § hours inthe assembly ling, a Galaxse requires 8 nours and a Utopia requires 12 nours. ‘The normal operation hours of assemby line are 170 per month, “The marketing and accounting departments have estimated thal the profis, per unt for the three types of computers are Rs 41,00,000 for the Epic, Rs. 1,44,000 for the Galaxie and Rs 2,82,000 for the Utopia. The marketing department futher reper that the demand is such that the finn can expect to sel all the cemputers it procuces inthe month, The chairman of the ‘company has established the follwing goels. These are listed blow, according to their importance: (@) Avoid underutiizatio of capacity in tems of regular houre of oporaton cf the assembly line. (i) Meet the comand of the north-eastem sales district for fve Epics, five Galases, and sight Utopias (ferentia weights should be assigned according tothe profit ratios ‘among the three types ot computes). i). Limit overtine operation ofthe assembly ine to 20 hours. (iv) Meet the sales goal for each ‘ype of computer: Epic — 10; Galaxie — 12; and Utopia ~ 10 (again assign weights eccording to the relative profit function for each computer. (v) Minimize the total overtime operation of the assembly tno, Formulate and solve this problem as a goal progamming model through two leratone (hree tables) by the eimplox method. 7. The manager of the only record shop in a town has a decision protiem that Invols multiple goals. The recors shop employs: five full-time and four partine salesmen. The normal working hours per month fora fulltime salesman are 160 hours and for parttime salesman 80 hours, According to pedormance ecord of the salesmen, the average sales has been five records per tour for fulkime salesmen and two records per hour for part-time salesmen. The average hourly wage rates are fs 3 for ful-ime salesmen and Rs 2 for partsime salesmen. ‘The average profit Fom the sales of a record is Fs 1.0. In view cr past record of sales, the manager lees that the sales goal {or the next month should be 5,500 records. Since the shop is open six days a week, salesmen are often raquired to work extra (not necessarly overtime but extra hours forthe parttime salesmen). The manager delioves that a good employer: employee relaionshp is an essential factor of business success. “Therefore, he feels fata stabie employment evel with sccasionel ‘overtime requirement is a better practice than an unstable fempioyment level with no overtime. However, he feels that foverime of more than 100 hours among the fultime salesmen, ‘should be avoided because of the decining sales effectiveness, caused by fatigue. ‘The manager has set the folowing goals: (i) The frst goal is to achieve a sales target of 5,500 records for the next month |W) The second goal is to mit the vertine cf fulltime salesmen to 10) hours. (i) The third goal is to provide job securty to salesmen, ‘The manager feels that ful ullzation of employees’ regular working hours (no layos) is an Imporant factor for a good employe-employee relationship. Howaver, ha Is twico as concerned with tho fll utization of ful time salesmen as wih the ful utization of parttime salesmen. () The last goal is to minimize the sumber of houws of ‘overtime f both fultime and parvime salesnen. The manager desires to assign diferetial weights to the ‘minimization of overtime accoraing to the net marginal Profit ratio between the fulHtime and partsime salesmen. Formulate and solve the given problem as a goal programming probiem.8. Anhospital admiristraton is reviewing departmental requests prior 1 the design of a new emeraency room. At issue is the number ‘of beds for each department. The current plans call for a 15,000 ‘square feet facity. The hospital board has establiched the {allowing goals in order of importance: Deparment No, of Cost per Area per Peak Requ- ‘Beds Bed bed trement Requested (nel. (ax. no. fequimpment) of Patients (Ps) at one me A 5 12600 «474 3 a 20 400542 18 c 20 600438 15 (i) Avcid overspending of the budget Ré 8,00,000. (i) Avoid plan requiring more than 15,000 sq, f. i) Meot the peak requirement (iv). Meet the departmental requirements. Formulate end selve (up te three iterations) the given problem as 2. goal programming problem. (Delhi Univ., MBA, 2003} 9, MEX has inherited Re 10,00,000 and seeks your advice concerning Investing his money. You have determined that 10 per cent can be eamed on a bank acoount and 14 per cent by investing In cetiicates of depost. In real estate, you estimate an annual Fetun of 15 per cent, while on the stock market you estimate that 20 per cent can be eamed arnualy. Me X has established the ‘ollowing coals in order of importance: () Minimize risk; therefore he wants to invest not more than 35 per cent in any one type of investment. (Must have Rs 1,00,000 in tho bank account to moot any emergency. (i) Maximum annual cash return. Formulate end solve (up to three iterations ony) tis preblem as 2 goal programing model that detemnines the amcunt of money Which X should invest in each investment option. 10. A company plans to schedule its annual advertising campaign. ‘The toll advertising budget is et at Fs 10,00,000. The firm can purchase local radio spots at Fs 2,000 per spot, lacal television pote at Re 12,000 per epot of magazine advoricing at Re 4,000 er insertion. The payoff from each advertiing medium is a function ofits audiorce size end audience characteristics. Let this payoff be defined as audience points. Audience poinis forthe twee adverising vehicies are: Radio 50 pints per spot Television 250 pints per spot Magazine ‘200 pints per insertion. ‘The Advertking Manager of the fm has establshed the following ‘goals forthe advertising campaign, listed inthe ard of importance, (i) The total budget should’ not excaed Rs 10,00,000. (The conract win he rasto and television station requres thatthe frm spend atleast Bs 9,00,000 for television and. radio ads. Goal Programming (ll) The company does not wish to spend more than Rs 2,00,000 lor magazine ads. (iv) Audience points from the adverising campaign should be maximized. Formulate and sole (4p to three iterations only this protiem as. 4 goal programming problem by the simplex method. [Deni Univ, MBA, 1995, 99) 11. A shoe manufacturer produces hiking boots ‘manufacturing procese of the boots consi siitching. The company has avallable 60 hours per week for the ‘sowing process and 80 hous pe week for the titching process ‘at normal capacity. The firm realizes profs of Fs 150 ger pair ‘on hiking boets and As 100 per pair on ski Boots. N requires 2 hours of sewing and 5 hours of sitching to produce one pair of hiking boots and 3 hours of sewing and 2 hours of siting to produce one pair of ski boots. The president of the company Wishes fo achieve the following goals, listed in the order of their importance: (i) Achieve the profit goal of Rs 5.250 per week: (0) Limit the overtme operation of the sowing center to 30 hours, (ii) Mest the sales goal for each type of boot ~ 26 hiking ‘boots an 20 ski boots. (@) vols any underutiizaion of reguar operation hours: ‘of the sewing center Fomulate and solve this problem as @ goal programming model, [Delhi Univ, MBA, 2000) 12, Dota Hospital is 2 medium-size many heathcare facies (HCF) hospital, located in a sina city of Bhar. HCF is specialzed in performing four types of surgeries: T, A, H, and ©. The Performance of these surgeres is constrained by three resources: ‘operating room hours, recevery room bed hours, and surgical sence bed days. The direcor of HCF woul lke to achieve the following objectives, in order of their irportence. Also given belew Is fhe information pertaining to the hospital ‘Resources ‘Operating. room Recovery room Surgical servee Average contribution to profit (Rs) 2,100 2,600 2,800 3.000 Pj: Achievement of at least Fs 5,00,000 in proft in a Sspedtied period of time given the avaliable resources, Pe: Minimizaton of idle capacity of available resources. Formulaie and soWve this problem as a GP problem, CHAPTER SUMMARY Goal programming (GP) is an approach used for solving a multi-objective optimization problem that balances conflicting objectives to reach a ‘satisfactory’ levelof goal attainment. A problem is modelled into a GP model in a mamner similar to that of an LP model, However, the GP model accommodstes multiple, and often conflicting, ineommensurable goals, in a particular priority order (hierarchy). A particular priority structure is established by ranking and weighing various goals and their subgoals, in accordance ‘with their importance. The priority structure helps to deal with all goals that cannot be completely and/or simultaneously achieved in such a manner that more important goals are achieved first, at the expense of the less important ones. ‘An important feature of a GP is that the goals (a specific numerical target values that the decision-maker would ideally like to achieve) are satisfied in ordizal sequence. That is, the solution of a GP problem involves schieving some higher orderOperations Research: Theory and Applications (or priority) goals first, before the lower order goals are considered. Since itis not possible to achieve every goal (objective), to the extent desired by the decision-maker, attempts are made 10 achieve each goal sequentially rather than Simuulcaneously, Up to-a satisfactory level rather than an optimal level. In GP, instead of trying to minimize or meximize the objective function directly, as in the case of an LP, the deviations from established goals within the given set of constraints are minimized. The deviational variables are represented in two dimensions — both positive and negative deviations from each goal and subgoal. These deviational variables represent the extent to which the target goals are not achieved, The objective function then becomes the minimization of a sum of these deviations, based! on the relative importance within the pre-emptive priority structure assigned to each deviation, CHAPTER CONCEPTS QUIZ True or False 1. The main advantage of the goal programming over linear programing ia ils ablity to sola problema compriing muliple Constraints 2. Goal programming appears to be the most appropriate, flexble ‘and powerful technique for complex decsion problem involving ‘multiple conficting. objective 4. Inga programming formulation each goal generates anew constraint and adds atleast one new variable tothe objective function. 4. ngoal progamming, a goal consaint having underachievement and ‘veractievement varibles 's exoressed a8 an equalty constraint. , who make use of goal programming, have to specty fe impertance of the goals as well as how much more Important one is expected to other. 16. The value of the objective function atthe optinal solution to the G> problem rellects the overachievement of the goal in i 7. In goal programming, the degree of goal achievement depends upon the relative managerial effort applied to an activity. 8. Unlike the simplex method of the linear programming, goal programming doze not use the least postive quotient rule’ when eciding upon the replaced variable 9. In order to apply goal prograrming, management must make the assumption that lineaity exists in the usage of resources te tho attainment of the goals 10. Azeume that in goal programming, tho variable Du and Do aro defined as the amount we eiher {all short of or exceed the target for some goal. Thus, these variabes can both be absent In the product mix ifthe goal is underachieved. Fil in the Blanks 1 {8 used for solvieg a mult-objective optimization problem that balances trade-cf! in conficting objectives. 12. Ina goal programming problem with poriized goal, the coaficlents (ol the under achievement varioles in the objective function called g 49, In contrast to goal progremming, the objective function of the linear programing is measured in only 1. Programming is. en extension of the ’rogramming in which multple goals ranked according to prioriies ‘may be Incuded in the formulation, 15, Goal programming used for__tachnology management. 16. If goal is stated in terms of variable only, ten Mey have 10 De restated In terms of variable belore foregoing with the graphical solution. 417. The sclution of the GP problem is ifthe targeted value of each gpal in xp column ie 18. In goal programming, if there are two or more 2 and cj ~ rows, then the problem tas goals. 419. The deviational variable with the identical ‘otherwise out level are 20. While solving a GP problem we enter the variable heading the column with the negatve « — z, value provided corresponding 6, ~ 3, value in the P, area is Multiple Choice 21. The use of GP model is prefered when (@) gods are satstoe in an ordinal sequence (©) goals are muliple incommensurable (©) more tnan one otjectve 1s set 10 acteve (6) allot the above 22. Deviational variables in GP model mus satsy folowing cortions (@ dpxdp=0 df dp =0 (©) dj+d;=0 (6) none of the above 23, In GP, at optimally, which othe ‘lloing conditions ndcate that 2 goal hav boon xacly satisfied (2) postive deviaional vaable 1s Ia the soliton mix with a Pegative value (0) both postive and negative devitona! varables are in the soision moe (6) both postive and negative devitonal varables are notin the eotuion mix (3) nore of the above 24. In GP problon, goals are assigned prortes such that (@) higher prorty goote must be aciavod before lowor priory goals (©) goals may not have equal priory () goals of greatest importance are’ given lowest prorty (6) allot the above 25. In a GP protiem, @ constraint that has an unechioved variable is expressed a: (2) an equally constant (b) @ léss than or equa to type constant (6) @ groater than oF equal 12 type constraint () allot the above 26. For appying a GP aporoach, the decision-maker must (a) set targets for each of the goale (b) assign pre-enplve priory 1 esch goal {c) assume that linearity exits in the use of resources to actiove goals (8) al of the above 27. Consider a gcal with constraint 9, (Xy. fe... %)+ dy —af = 4, land the term 3 (d) nore of the aboveGoal Programming Baad nr pa ih cae gnc) 2 | ff maul eah 8 aalayon (dj 2 0) with d,~ in the objective function. Then (0) prority goats {d) unattainable goals le toes ees we hn tae fie foc Coan seg ae (c) both (a) and (b) (c}_ beth (a) and (b) (qd) one of the above: 1, Gn goer |* joes CEE as ccmemremeumernga| edly {b) allows you to have multiple goals, with or without priorities | (c)_ beth (a) and (b) () is an approach to achieve goal of a solution to all integer LP | (d)_aificial variable problems valu jal in the ‘soluton-value, xg) column is 30. Ih simplex method o! goal programming, the variable to enter the (a) multiple solution (b) infeasible solution re oe ate may |= eee Poestenmmerssorce| Giscet 9 cee. 1 ee ee eee La if Naeseaa mata si itt _ Sf beds _ Se habinesln 2 a Oe cello at 26. (8), 27. (a), 28 (0), 29. (a), (0), 34. (C),_—82. (a), 93. (@), 94. (0), _35. (0) CASE STUDY Case 8.1: Blended Gasoline Company ‘The Blended Gasoline Company purchase blended gasoline from a network of three vendors. This brand of gasoline i composed of one or more of thre blending constituents, each with a cifferent ostane rating. There is a different vendor for each blending constituent. The blended gasoline is characterized by its overall octane biend, The vendors ‘ave the following characteristics Fe Service Ratio Ps Level (9) Limit (barrels) 1 680 4 5000 2 560 3 6000 3 720 36, 12000 Service is measured in terms of buyer stisfaction and ranges from zero to one, with zero being satisfactory and ‘one being the most satisfactory. The ration limit of a vendor represents the maximum purchase that a buyer ean make fom that vendor, The limit is imposed by the government and the vendor must abide by it ‘The company wishes to purchase 10,000 barrels of blended gasoline, while meeting the following goals, lied in onder of priority to the company: (@) Quali goat "The aggregate octane level of the blended gasoline purchased should be at least 100, (ii) Lead ime goal: All the gasoline purchased should arrive within 28 days. (ii) Service goal ‘The service level of each vendor supplying gasoline should be at Teast 0.05 Gv) Price goal ‘The aggregate cost of the gascline purchased should be no more than Rs 600 per barred on an average. Suggest an appropriate quantitative mode! to help the management ofthe company to trade-off conti {for the replenishment of gasoline, 8 cals Case 8.2: Paper Manufacturing Company Green Wood Peper Manufacturing Company is in the business of manufacturing different types of paper used in the ‘rinting of newspapers and magazines, Hardwood and bamboo chips are the inputs required to produce paper. Those‘Operations Research: Theory and Applications chips mixed in cxrtain proportion are cooked in a kamyr digester with the help of chemicals to produce pulp, the intermediary to paper. This pulp is further processed to finished paper. There are three quality characteristics of pulp which are to be maintained in order to got right quality of paper, These are: (@ K-number (i) Burst factor, and (ii) Breckinglengt K-aumber is an indirect index of digestion or cooking that has been done by the chemical called cooking liquor, at the kamyr digestor. If the chips fed into the digestor are overcooked, resulting in excess disintegration of chips, the K-aumter will accordingly be low. This disintegration of raw material is a loss to the company. On the other hand, if chips are undercooked, indicated by a high K-number, then ata later stage consumption of the bleaching agent will be high Thus, K-number is the most important characteristic that should be stictly maintained within limits. The ‘other two characteristics determine the srengih of finished paper. ‘The proportion of hardwood and bamboo can easily be adjusted. All the process variables are controllable and ‘can be measured directly. The details of the input, process variables, and output are given as follows: | 3 Input: (e,) Hardwood (%) 2- 40 Processvariebles + Upper cooking zone temperature (°C) 40 ~ 175, Ry: Lower cooking zane temperature (°C) 140-173 Ry: LP steam pressure (kg/em?) 10- 44 2k, HP steam pressure (kyon?) € 80-205 Rg: Acive alkali as NaOH (%) 2-35 Rg: Soiphidity of white liquor (2) 1B 25 RR, : Alkali index (n0) 125 ~ 187 Onuput characteristics Y,: Kemumber e- 18 1: Burst factor sloseto 33 Y, : Breaking length lose 10 5,000 mm ‘In some cases specification did not exist and hence permissible limits were considered instead. The company was unable 10 maintain the desired characteristics of pulp. The problem was to fx the levels of the input and the process varisbles so tat the specification was met. Pre-emptive priority factors The K-aumber is the most important characteristic to be fulfilled and gets the top priority. Pricrities for others were fixed by the management after giving due consideration to the quality aspect «s well as the ease of adjusting/medifying the levels ofthese variables. ‘The company hired the services of « consultant. A follow-up study was underteken by the consultants, which linked the input with output vis the process variable. Exactly 46 sets of such data were collected over a period of 13 days. Multiple linear regression analysis was undertaken and the following relationships were obtained. Y, = 21840 + 006, ~ 0.05R, + 0004R, ~ 0.678, + 0.242, ~ 0.13R, + 0.198, ~ 0.18%, (Mukiple correlation coefficient = 0.74) ¥,= 38.94 + 0.05x, — 0.02, + 002K, + LOTR, + O.21R, + 0.06R, + 0.02R, — 0.69%, (Mukiple corre Y= 327240 — 24.37, + 9.997R, + 8ABR, ~ 268.682, + 12092R, + 67.27R, + 27.89R, ~ 13846R, (Multiple correlation coefficient = 0.66) It should be noted that the relaionship of these output characteristics with the variables are conflicting in the sense that for increase (decrease) in the Jevel of some of the variables, the values of some of these characteristics, increase (decrease) whereas the values of other decrease (increase). ‘Suppose, your services are hired by the company, then suggest optimum values of input and process variables while satisfying conflicting goals. ion coefficient = 0.72)Transportation Problem “We want these assets 10 be productive. We buy them. We own ‘them. To say we care only abott the short term is wrong. What J eare about is seeing there assots in the best hand” ~ Carl Teahn Preview Thestructure of transportation problem involves. large ‘number of shipping rovtes from several supply origins to several demand destinations. The objective isto determine the number of units of an item (commodity or product) that should be shipped from an origin to a destination in order to satisy the required quantity of goods or services at each destination centre. Learning Objectives After studying this chapter, you should be able to recognize and formulate a transportation problem involving 2 large number of shipping routs. drive initial feasible solution using several methods. + drive optimal solution by using Modified Distribution Method. * handle the problem of degenerate and unbalanced transportation problem. * examine multiple optimal solutions, and prohibited routes in the transportation probiem. ‘+ construct the initial transportation table for a trans-shipment problem, + solve a profit maximization transportation problem using suitable changes in the transportation algorithm. Chapter Outline a4 as 96 97 98 ooog Introduction Mathematical Mode! of Transportation Problem ‘The Transportation Algorithm Methods of Finding initial Solution + Conceptual Questions A + Self Practice Problems A + Hints and Answers Test for Optimality + Conceptual Questions 8 + Self Practice Problems B + Hints and Answers ve jons in Transportation Problem Maximization Transportation Problem Trane:shipment Problom + Conceptual Questions C ‘+ Self Practice Problems C + Hinte and Anewore Chapter Summary Chapter Concepts Quiz case Stuty ‘Appendix: Theorem and Results Copyrighted mdeni optimal transportation routes ‘along with units of commodity to be shipped in order to ‘mnimize total transport (Operations Research: Theory and Appications 9.4 INTRODUCTION One important application of linear programming is in the area of physical distribution (transportation) of ‘goods and services from several supply centres to several demand centres. It is easy to mathematically ‘express a transportation problem in terms of an LP model, which can be solved by the simplex method, But because it involves a large number of variables and constraints, it takes a long time to solve it However, transportation algorithms, namely the Stepping Sione Method and the MODI (modified distribution) Method, have deen developed for this purpose ‘The structure of transportation problem involves a large number of shipping routes from several supply origins to several demand destinations. The objective is to determine the number of units of an item (commodity or product) that should be shipped from an origin to @ destination in order to satisfy the required quantity of goods or services at each destination centre. This should be done within the limited quantity of goods or services available at eack supply centre, at the minimum transportation cost and/or time. The transportation algorithm discussed in this chapter is applied to minimize the total cost of transporting a homogeneous commodity (product) from supply centres to demand centres. However, it can also be applied to the maximization of some total value or utility. For example, financial resources are distributed in such a way that the profitable return is maximized. ‘There are various types of transportation models and the simplest of them was first presented by FL Hitchcock (1941). Itwas further developed by T C Koopmans (1949) and G B Dantzig (1951). Several extensions of transportation models and methods have been subsequently developed. 92 MATHEMATICAL MODEL OF TRANSPORTATION PROBLEM Let us consider Example 9.1 to illustrate the mathematical model formulation of transportation problem of transporting a single commodity from three sources of supply to four demand destinations. The sources ‘of supply are production facilities, warehouses, or supply points, characterized by available capacities. The destinations are consumption facilities, warehouses or demand points, characterized by required levels of demand. Example 9.1 A company has three production facilities S,, S, and S, with production capacity of 7, 9 and 18 units (in 100s) per week of a product, respectively. These units are to be shipped to four warehouses D,, Dz, D, and D, with requirement of 5, 6, 7 and 14 units (in 100s) per week, respectively. ‘The transportation costs (in nipees) per unit hetween factories to warehouses are given in the table below: Dy Te Ta een] 0 | 7 Ss, 19 30 ms 70 30 40 60 9 Ss; 40 8 70 20 18 Demand 8 1 4 4 Formulate this transportation problem as an LP model to minimize the total transportation cost. Model formulation Let x, = number of units ofthe prociuct to be transported fom factory i (1 to warehouse j (j= 1, 2, 3, 4) ‘The transportation problem is stated as an LP model as follows: 2.3) Minimize (total transportation cost) Z = 19x, + 30x,2 + 50r,5 + 10r,4 + 70x, + 30x, + 40x55 + 60534 + 40 ry) + 8x52 + TOiy3 + 254 subject to the constraints aptagtaatae= 7 yy +¥yy tay +, = 9 [Capacity available) Hyp ty tg tpg = 18image not availableOperations Research: Theory and Applications When total demar equals total supply, the transportation problem is sald to be balanoad| In this problem, there are (m +7) constraints, one for each source of supply, and distinction and ‘mx nvariables. Since all (m+ 7) constraints are equations, and since the transportation model is always balanced (total supply = total demand), one of these equations is extra (redundant). The extra constraint ‘equation can be derived from the other constraint equations, without affecting the feasible solution. It follows that any feasible solution for a transportation problem must have exactly (m+n ~ 1) non-negative basic variables (or allocations) x, satisfying the rim conditions. Remarks 1. When the total supply’s equal to the total demand, the problem is called = balanced transportation problem, otherwise it is called an unbalanced transportation problem. The unbalanced transportation problem can be made balanced by adding a dummy supply centre (row) or a dummy demand centre (column) as the need arises. 2. When the number of positive allocations (values of decision variables) at any stage of the feasible solution is less than the required number (rows + columns ~ 1), i.e. number of independent constraint equations, the solution is said to be degenerate, otherwise non-degenerate. For proof, sec Appendix at the end of this chapter. 3. Cells in the transportation table that kave positive allocation are called occupied cells, otherwise they are known as empty or non-oceupied cells 9.3 THE TRANSPORTATION ALGORITHM The algorithm for solving to a transportation problem may be summarized into the following steps: Step 1: Formulate the problem and arrange the data in the matrix form The formulation Of the transporation problem is similar to the LP problem formulation. Here the objective function is the total transportation cost and the constraints are the supply and demand available at cach source and destination, respectively. Step 2: Obtain an initial basic feasible solution In this chapter, following three different methods are discussed to obtain an initial sok North-West Corner Method, # Least Cost Method, and * Vogel's Approximation (or Penalty) Method. ‘The initial solution obtained by any of the three methods must satisfy the following conditions: (The solution must be feasible, ie. it must satisfy all the supply and demand constraints (also called rim conditions). (i) The number of positive allocations must be equal to m + n— 1, when m is the number of rows and nis the number of columns, Any solution that satisfies the above conditions is called non-degenerate basic feasible solution, otherwise, degenerate solution. Step 3: Test the initial solution for optimality _In this chapter. the Modified Distribution (MODI) method is discussed to test the optimality of the solution obtained in Step 2. If the current solution is ‘optimal, then stop. Otherwise, determine a new improved solution, ‘Step 4: Updating the solution Repeat Step 3 until an optimal solution is reached. 94 METHODS OF FINDING INITIAL SOLUTION ‘There are several methods available to obiain an initial basic feasible solution. Here we shall discuss only three different methods: 9.4.1 North-West Corer Method (NWCM) It is a simple and efficient method to obiain an initial solution. This method does not take into account the cost of transportation on any route of transportation. The method can be summarized as follows: Step 1: Start with the cell at the upper left (north-west) comer of the transportation matrix and allocate ‘commodity equal to the minimum of the rim values for the first row and first column, ie, min (4,, 0).‘Transportanon Problem Fay Step 2: (a) If allocation made in Step 1 is equal to the supply available at first source (a, in fist row), then move vertically down to the cell (2, 1) in the second row and first column. Apply Step 1 again, for next allocation. () If allocation made in Step 1 is equal to the demand of the first destination (by in first colura), then move horizontally to the cell (1, 2) in the first row and second column. Apply Step 1 again for next allocation. © a, = 5, allocate x, ‘Step 3: Continve the procedure step by step till an allocation is made in the south-east comer cell of the transportation table. Remark If during the process of allocating commodity at a particular cell, the supply equals demand, then the next allocation of magnitude zero can be made in a cell either in the next row or column. This condition is known as degeneracy. Example 9.2 Use North-West Comer Method (NWCM) to Example 9.1 to find an initial basic feasible solution to the transportation problem. Solution The cell (S;, D,) is the north-west comer cell in the given transportation table. The rim values for row S, and column D, are compared. The smaller of the two, is. 5, is assigned as the first allocation; otherwise it will violate the feasibility condition. This means that 5 units of a commodity are to be transported from source 5; to destination D,.. However, this allocation leaves a supply of 7 - 5 = 2 units of commodity at S, Move horizontally and allocate as much as possible to cell (S,, D3). The rim value for row S, is 2 and for column D, is 8. The smaller of the two, ie. 2, is placed in the cell. Proceeding torow S,, since the demand of D, has been met, nothing further can be allocated to D;. The unfulfilled demand of D, is now 8 ~ 6 units, This can be fulfilled by S with capacity of 9 units. So 6 units are allocated to cell (Sy, D,). The demand of D, is now satisfied and a belance of 9 — 6 =3 units remains with 5. =a, or by and move diagonally to the cell (2, 2). Table 9.2 Initial Solution using NWCM We now move horizontally and vertically in the same manner. This should be done because successive demand and supply are met. Ensure that the solution is feasible, that is, the number of positive allocations (occupied cells) is equal tom +n—-1=3 + 4— The total transportation cost ofthe initial solution derived by the NWCM js obtsined by multiplying the quantity x, in the occupied cells with the corresponding unit cost c, and adding al! the values together. ‘Thus, the total transportation cost of this solution is ‘Total cost= 5x 19 +2 30 +6 30+3 x 40 +4 = 70+ 14% 20=Rs 1,015 9.4.2 Least Cost Method (LCM) Since the main objective is to minimize the total transportation cost, we must try to transport as much as possible through those routes (cells) where the unit transportation cost is lowest. This method takes into account the minimam unit cost of transportation for obtaining the initial solution and can be summarized 2 follows: ‘Step 1: Select the cell with the lowest unit cost in the entire transportation table and allocate as much as possible to this cell. Then eliminate (line out) that row or column in which either the supply or demand is exhausted. Ifa row and « column are both satisfied simultaneously, then only one may be crossed out.Operations Research: Theory and Applications Table 9.3 Table 9. In case the smallest unit cost cell is not unique, then select the cell where the maximum allocation can bbe made. Stop 2: After adjusting the supply and demand for all uncrossed-out rows and columns repeat the procedure withthe next lowest unit cost among the remaining rows and columns ofthe transportation table and allocate as much as possible to this cell. Then eliminate (Line out) that row and column in waich either supply or demand is exhausted, Step 3: Repeat the procedure until the entire avsilable supply at various sources and demand at various destinations is satisfied. The solution so obtained need not be non-degenerate, Example 9.3. Use Least Cost Method (LCM) to Example 9.1 in order to find initial basic feasible solution to the transporiation problem. Solution The cell with lowest unit cost (ic. 8) i (Sy, D3). The maximum units which we ean allocate 10 this cell is 8. This meets the complete demand of D, and leave 10 units with S,, as shown in Table 9.3. In the reduced table without column D,, the next smallest unit transportation cost, is 10 in cell (5, Dp). The maximum which can be allocaied to this cell is 7. This exhausts the capacity of 5, and leaves 7 units with D, as unsatisfied demand, This is shown in Table 93, 7 ee = 10 | s ® 75 s |? [0 40 oo : t . |* 8 70 20 3 18 | Be Demand 5 8 7 n 34 In Table 93, the next smallest cost is 20 in cell (S,, D,). The maximum that can be allocated to this cell is 7 units. This satisfies the entire demand of D, and leaves 3 units with S, es the remaining supply, shown in Table 9.4 In Table 94, the next smallest unit cost cell is not unique. That is, there are two cells ~ (Sy, D,) ard (S,, D,) ~ that have the same unit rransportation cost of 40, Allocate 7 units in cell (S,, D,) first because it can accommodate more units as compared to cell (S,, D,). Then allocate 3 units (only supply left with $2) to cell (Sy, Dy). The remaining demand of 2 units of D, is fulfilled from S,. Since supply and demand at each origin and destination is exhausted, the initial solution is arrived at, and is shown in Table 9.4, 5 ® ss Demand ‘The total transportation cost of the initial solution by LCM is calculated as given below: Total cost = 7 « 10+2% 70+7*40+3% 40+8<8+7% 20=Rs 814 ‘The toxal transportation cost obtained by LCM is less than the cost obtained by NWCM. The optimal solutioa can also be obtained faster by using LCM.‘Trangportation Problem 9.4.3 Vogel's Approximation Method (VAM) Vogel’s approximation (penalty or regret) method is a heuristic method and is preferred more than the other ‘two methods described above. In this method, each allocation is made on the basis of the opportunity (or penalty or extra) cost that would have been incurred if the allocations in certain cells with minimum unit transportation cost were missed. In this method allocations are made so that the penalty cost is minimized. The advantage of this metkod is that it gives an initial solution which is nearer to an optimal solution or js the optimal solution itself. The steps of VAM are as follows: Stop 1: Calculate the penalties for each row (column) by taking the difference between the smallest and next smallest unit transportation cost in the same row (column). This difference indicates the penalty or extra cost that has to be paid if one fails to allocate to the cell with the minimum unit transportation cost. Step 2: Select the row or column with the largest penalty and allocate as much as possible in the cell that has the least cost in the selected row or column and satisfies the rim conditions. If there is a tie in the values of penalties, it can be broken by selecting the cell where the maximum allocation can be made. Step 3: Adjust the supply and demand and cross out the satisfied row or column. Ifa row and a column are satisfied simultaneously, only one of them is crossed out and the remaining row (column) is assigned «zero supply (demand). Any row or column with zero supply or demand should not be used in computing, future penalties Step 4: Repeat Steps 1 to 3 until the entire available supply at various sources and demand at various, destinations are satisfied. Example 9.4 Use Vigel’s Approximation Method (VAM) to Example 9.1 in order to find the initial basic feasible solution to the transportation problem. Solution The differences (penalty costs) for each row and column have been calculated as shown in Table 9.5. In the first round, the maximum penelty, 22 occurs in column D,. Thus the cell (S,, D,) having the least transportation cost 8 is chosen for allocation. The maximum possible allocation in this cell is 8 and it satisfies demand in column D,. Adjust the supply of S, ftom 18 to 10 (18 ~ 8 = 10). ee oe ee Dy | Sen | Rowatterncs ; ED 0 7 7 39 aw @ 5 [pe ® wo 2 | w 2 2 2 |_ @ | 3 [a 8 20 18 12 20 30 | © @ [aewee | 8 3 saints 3 | conn | 24 2 10 0 ldigferences| 21] wo | wo | Segal ieee) tg 2 | The new row and column penalties are calculated except column D, because D,,s demand has been satisfied. The second round allocation is made in column D, with target penalty 21 in the same way as in the first round as shown in cell (S,, D,) of Table 9.5. In the third round, the maximum penalty 50 occurs at Sy, The maximum possible allocation of 10 units is made in cell (S,, D,) that has the least transportation cost of 20, as shown in Table 9.5, The process is continued with new allocations till a complete solution is obtained. The initial solution using VAM is shown in Table 9.5. The total transportation vost associated with this method is calculated as: Total cost =x 10-42 x10-+7 x 40.42 x 60-+Rx K+ 10x 20= Re 779 Table 9.5 Initial Solution Using VAMZ Operations Research: Theory and Applications Example 9.5 A dairy firm has three plants located in a state. The daily milk production at exch plant is as follows: Plant 1:6 million lites, Plant 2:1 million lites, and Plant 3 : 10 million litres Each day, the firm must fulfil the needs of its four distribution centres. The minimum requirement of each centre is as follows: Distribution centre 1: 7 million litres, Distribution centre 2: 5 millon litres, Distribution centre 3:3 milion lites, and Distribution centre 4 : 2 million lites ‘Cost (in hundreds of rupees) of shipping one million litre from each plant to cach distribution centre is given in the following table: Distribution Centre Dy Dy Dy De A 2 3 u 7 Plant P, 1 0 6 1 » 3 8 15 9 Find the initial basic feasible solution for given problem by using (@ North-west comer rule (b) Least cost method (©) Vogel’s approximation method ifthe objective is to minimize the total transportation cost. Solution (a) North-West Corner Rule 2 3 1 7 al © 6-4, 1 0 6 1 Pat |" @ =a ae 8 15 9 ee =a Table 96 ‘ ® @ ® Demnd T= SBS 22h (@ Comparing a, and 5, since a, < b,; allocate x,, = 6. This ethausts the supply at P, and leaves 1 unit as unsatisfied demand at 2. Gi) Move to cell (Py Dy). Compare a, and by (i.e. 1 and 1). Since a; ~ by, allocate x3, ~ 1 Gii) Move to cell (Py, D,). Since supply at Py, is equal to the demand at D,, Ds and Dy, therefore, allocate x33 = 5, x53 = 3 and x = 2. It may be noted that the number of allocated cells (also called basic cel) are 5 which is one less than the required number m +n - 1 (3 +4 — 1 = 6). Thus, this solution is the degenerate solution. ‘The transportation cost associated with this sohution is: Total cost =Rs (2 *6+1*1+8 «5+ 15*3+9 2) 100=Rs 11,600‘Transportation Problem (b) Least Cost Method nied? Ce 6 1 : ale 8 15 9 i © @ Ore Demand Z 5 3 2 @ The lowest unit cost in Table 9.7 is 0 incell (P,,D,), therefore the maximum possible allocation that which can be made here is 1. This exhausts the supply at plant P,, therefore, row 2 is crossed out. (The next lowest unit cost is 2 in cell (P,, D,). The maximum possible allocation that can be made hte is 6. This exhausts the supply at plant P,, therefore, row P, is crossed out. (i) Since the total supply at plant P, is now equal to the unsatisfied demand at all the four distribution centres, therefore, the maximum possible allocations satisfying the supply and demand conditions, are made in cells (PD), (Py, D,), (Py, D3) and (Py, D,). ‘The number of allocated cells in this case are six, which is equal to the required number m+n~1G +4~1=6), Thus, this solution is non-degenerate. The transportation cost associated with this solution is Total cost = Rs (2% 6 +5 x1+8 x4 +15 x3 +9 *2) x 100 =Rs 11,200 (©) Vogel's Approximation Method: The method is self-explanatory as shown in Table 9.8. A D ® 1 ® 6 1 roo = - Plant Py @ Real 8 5 9 10x 30 4 =e eo Demand 7 s 3 2 Column penalty 1 3 5 6 3 5 4 2 3 - 4 2 Distribution Cente ‘The number of allocated cells in Table 98 are six, which is equal to the required number m+n—1(3 + 4-1 = 6), therefore, this solution is non-degenerate. The transportation cost associated with this solution is Total cost = Rs (2 1+ 3%5+1*145* 6415 «3 +9 x 1) 100 =Rs 10,200 It can be seen that the total transportation cost found by VAM is lower than the costs of transportation ‘determined by the other two methods. Therefore, itis of advaniage to use this method in order to reduce ‘computational time required to obtain optimum sclution, Table 9:7 Table 98 PigOperations Research: Theory and Applications CONCEPTUAL QUESTIONS A ‘Show thet all the bases for @ transportation problem are triangular, With reference to a transportation protiem define the following () Feasble solution Basic feasible solution (i) Optimal soutien | (iv) Nondlegenerata basic feasible sclution | Given a mathematical fomultion of the tranapoation problem | and fhe simplex methods, what are the differences inthe nature of problems that can be solved by using these mettods? Prove that there are only m+ n~ 1 independert equations in a transporaton protlem, m and n beirg the nunber of ofgins and destination and only one equation can be drepped as being redundant. (For preot see Appendix) . Dosetibe the tranaperatin preblom with its genorel mathematical formulation ‘Show thet a transportation problem is a special type of LP problem. In what areas of management can the transportation model be effectively used? Discuss. What are the characteristics of transportation problem of linear programming? 8, What {s meant by the tangular form of 2 system of near equatons? When does a system of linear equations have tiangular basis? (See Appendix for preot) 9. What ie maant by ron-dageneeate basis feasible sokion of transportaton problem? Explain in briel three, methods of inital feasible solution for transportaton problem. 11. Explain the various steps involved In solving transportation problem using (i) Least cost method, and (li) Vogel's approximation method, 12. Explan the () North-West Commer method, (i) Least Coat method, {and (il) Vogel's Approximation method, for obtaining an intial basic feaable colution of a traneportation problem. 19. State the transportation protiem. Describe clearly the steps Involved in solving i. 14, Is the transportation model an example of decision-making Under certainty or under uncertainty? Why? 15. Why does Vogel's aoproximaton method provide a good intial feasible sclution? Can the North-West Comer method ever be able 10 provide an inal soluton with @ cost as iow a5 this? SELF PRACTICE PROBLEMS A Determine an initiel basle feasibie solution to the following transportaion problem by using (a) NWCR. (b) LM and (c) VAM. 3. Determine an Initial basic feasible solution to the following transportaton problem by using (a) NWCM, (6) LCM, and (c) VAM. Destination Destination Bm | | % | % | sumey Dy | 2% | 0% | & | supoy s [a|wlsfe] a a [aaa] a [as | 250 souce | 9, | 7 | 10] 4 | 2 | 19 source | 2 | 16 | 10 | 14 | 10 |) 200 sg wl a7| wa | wo co | a| | 3 | 0 | 400 Demand | 6 | 10 | 12 | 18 Demand | 200 | 225 | 275 | 250 Determine an initial basic feasitie ‘ransportaion problem by using (a) solution to the folowing the least cost method, and 4, Dotomine an initial basic transgortaton problem by feasible solution to the following Using he NorthWest comer rule, {0) Wopets approximation mea where 0) and_D, represent Ah gin and. fh destination, roapectve Sete ad Destination Lo [|] % | 4% | suey >|] % | | swey Be psteta lal # a cer icae souo | | 9 | 3 | 2] 1 | so Sourc:| “Gy 2 6 s | 4|2|s5 |e] 2 a elas conena| | | | wf ewe | ie | | we | HINTS AND ANSWERS yg = 19, My = 8, Mag = 8. Mg = 4s Map = 8. May = 4, Mag = 14) Total cost = 686. fa) and (Om (a) X= 200, x12 2D, Xa 50, Xp 10, Xz = 20. X49 = 20. tog = 10. 2 20; Tota cost 120, 178, Xp5 = 125, Ap = 150, May = 250; Toll cost = 12,200. 1b) x1, = 200, xp = 50, Xap = 175, Aq = 125, 249 = 150, Xyq = 250; Total cost = 12,200. Ox 200, 2 = 50, Keg = 978, Hey = 125, May = 275, Xy Mag = Bi Xap = 2 Nyy = 14 Xyy = As My = 4: Toda cost = 125; Total cost = 12,075. Rs 128,Transportation Problem 9.5 TEST FOR OPTIMALITY Once an initial solution is obtained, the next step is to check for its optimality. An optimal solution is one where there is no other set of transportation routes (allocations) that will further reduce the total transportation cast. Thus, we have to evaluate each unoccupied cel (epresents unused route) inthe transportation table in terms of an opportuity of reducing total transportation cost. ‘An unoccupied cell with the largest negative opportunity cost is selected to include in the new set of transportation routes (allocations). This value Indicates he per unit cost reduction that can be achieved by raising the shipment allocation in the unoccupied cell from its present level of zero. This is also known as — an incoming cell (or variable). The outgoing cell (or variable) in the current solution is the occupied cell The negative (@asic variable) in the unique closed path (loop) whose allocation will first become zero as more units are Opportunity oat allocted to the unoccupied cell withthe largest negative opportunity cost. Such an exchange reduces the MaCS!08 tho 96" ‘otal transportation cost. The process is continued until there is no negative opportunity cost. That is, the that ean be current solution cannot be improved further. This would be the optimal solution, ‘achieved by raising An efficient technique called the modified-distriburion (MODI) method (also called u-v method or the shipment ‘method of multipliers), which helps in comparing the relative advantage of altemative allocations for all SMe OF samy ‘the unoccupied cells simultaneously, is discussed below. The MODI method is based on the concept of fs present level of duality zero 9.5.1 Dual of Transportation Model For a given basic feasible solution if we associate numbers (also called dual variables or multipliers) u, and v, with row i (i= 1, 2,....m) and column j (j= |, 2,....n) of the transportation table, respectively, then i, and v, must satisfy the equetion u,+ v= cy, for each occupied cell (, /) ‘These equations yield m +n — 1 equations in m+ n unknown dual variables. The values of these variables can be determined by arbitrarily assigning a zero value to any one of these varisbles. The value of the remaining m + n ~ 2 variables can then be obtained algebraically by using the above relationship for occupied the cells. Once the values of v, and v, have been determined, evaluation in terms of opportunity cost of each unoccupied cell (called non-basic variable or unused route) is done by using the equation: ,,= Cy ~ (ut, + ¥,), for each unoccupied cell (r s) For proving these two results, at us consider the general transportation model using sigma notation, Mininze2= &, £6 subject to the constraints sam (Supply) Modl method helps mn (Demand) in eomparing the and 3y20° forall jand j alloatiors forall the since all of the constraints are equalities, write each equality constraint equivalent to two inequalities Unovcupied cells = folowe , ee salts taneous (Supply constraints) (Demand constraints) Let 1 and u; be the dual variables, one for each supply constraint i. Similarly vj, Vj be the dual variables one for each demand constraint j. Now the dual of the transportation model is given byOperations Research: Theory and Applications The dual variables us and ys represent the ‘shadow price (valve ‘of the commodiy) for the supply ‘centres and demand centres, respectvely Table 9.9 subject to the constraints (af -u7 + (yt Wo) Sey and uf sur syt sy 20, forall iandj. ‘The variables x} and xj” that appear in the objective function, may take positive, negative or zero values. Thus, cither of these will appear in the optimal basic feasible solution because one is the negative of the other. The same argument may be given for a and vj. Thus, let j= uf muy, i= 1,2,. ye yt -y JEL. The values of wand v, will then be unrestricted in sign | reac dar ir transportation model can now be written as Maximize 2" = $ ma) + 3 vy by subject to the constraints uty S¢, and uy v, unrestricted in sign for all # and j. ‘Tae variables x form an optimal solution tothe given transportation problem provided (solution x, is feasible for all (i, )) with respect to original transportation model. () solution 1, and vis feasible for alli, ) with respect to the dual of the criginal transportation model. Gi), = 4, =v) x, = 0 for all and j ‘The relationship (cy —u,— ¥))xy = 0 is also known ss complemeniary slackness for a transportation problem ‘and indicates that (@) if'x, > 0 and is feasible, then cy — u — v, = 0 or cy = 1, + v, for each occupied cell, (© ite) = 0 and c) >, +p then is tot desinble 18 have 4 > O in the solution mix because it would cost mote to trandpont on a route (3) (0 ifcy 0, then the current basic feasible solution is optimal (i Ifd,=0, then the current basic feasible solution will remain unaffected but an alternative solution exists, (ii) If one or more d, < 0, then an improved solution can be cbtained by entering unoccupied cell (if) inthe basis. An unoccupied cell having the largest negative value of d, is chosen for entering inio the solution mix (new transportation schedule), Step 4: Construct a closed-path (or loop) for the unoccupied cell with largest negative opportunity cost. Start the closed path with the selected unoccupied cell and mark a plus sign (+) in this cell, Trace # path along the rows (or columns) to an occupied cell, mark the comer with a minus sign (-) and continue down the column (or row) to an occupied cell. Then mark the comer with plus sign (+) and minus sign (-) alternatively. Close the path back to the selected unoccupied cell. Step 5: Select the smallest quantity amongst the cells marked with minus sign on the comers of closed loop. Allocate this value to the selected unoccupied cell and add it to other occupied cells marked with plus signs. Now subtract this from the occupied cells marked with minus signs. Step 6: Obtain a new improved solution by allocating. units to the unoccupied cell according to Step 5 and calevlate the new total transportation cost. Step 7: Further test the revised solution for optimality. The procedure terminates when all d,, = 0 for unoccupied cells 1, The closed-loop (path) starts and ends at the selected unoscupied cell. It consists of 1 (connected) lines whose end points must be oceupied cells, except for 12 end point associated with entering unoccupied cell. This means that every comer element of the loop must be an occupied cell Changing the shipping route involves adding to ‘alls on the closed path with plus signs ‘and. subtracting from alle with nogaive signsE324 Operations Research: Theory ana Applications It is immaterial whether the loop is traced in a clockwise or anti-clockwise direction end whether it starts up, down, right or left (but never diagonally). However, for a given solution only one loop can be constrvcted for cach unoccupied cell. 2. There can only be one plus (+) sign and only one minus (-) sign in any given row or column. 3. The closed path indicates changes involved in reallocating the shipments. ‘The steps of MODI method for solving a transporation problem can also be described by the flow chart shown in Fig. 9.1 Express data rable form 1s fo] = Add dumny row oF colum ittalnced SN” SS me italaced rable ‘Convert it nto a minimization problem by subtracting cach element ofthe table form ts highest lemeat Finan ital basi esibie solution sit \ Yes__[Remove degencatingby asigning Geesnera! Dt appropiate cells) Ne Revised the solution and + Calcuite wand vby using u + = ey for occupied cells and ‘check rin conditions| then + Calculate opportunity cost diy ey (uy +) for unoccupied cells + Select anunoceupid cell aving largest |, 8 egtve value of dy << 0 + Coratruct loopand ale sllocstions tw revise the current soutien 6 ‘Terminate the procedare 9.5.4 Close-loop in Transportation Table and its Properties Any basic feasible solution must contain m +n ~1 independent non-zero allocations. The independent ‘non-zero allocations imply that one cannot form a closed circuit (loop) only by joining the positive allocations made by horizontal and vertical lines. Mathematically, provided: @ any two adjacent cells of the ordered set lie either in the same row or in the same column, and {10 three or more adjacent cells in the ordered set lie in the same row or column. The first cell of the set must follow the last in the set, ie, each cell (except the last) must appear only once in the ordered set. Consider the following two cases represented in Tables 9.10(a) and 9.10(b). In Table 9.10(a), if we join the positive allocations by horizontal and vertical lines, then a closed loop is obiained. The ordered set of cells forming a loop is: L={(a,2), (a, 4), (6.4), (6, 1), (6,195 6, 2), (@2)} The loop in Table 9.10(b) is not allowed because it does not satisfy the conditions in the definition of a loop. That is, the cell (6, 2) appears twice.‘Transportation Problem Fay e 6 © € d| d| I e[ & é e | Table 9.10 @ © Romarks 1. Every loop has an even number of cells and has at least four ces. 2. The allocations are said to be in independent position if itis not possible to increase or decrease any independent individual allocation without changing the positions of these allocations, or if a closed loop cannet be formed through these allccations without violating the rim conditions 3. Each row and column in the transporiation table should have only one plus and minus sign. All cells that have 2 plus or a minus sign, except the starting unoccupied cell, must be occupied cells. 4, Closed loops may or may not be in the shape of a square Example 9.6 Use Example 9.1 datz, to obtain an optimal solution using MODI method. Dr D2 Ds Ds Supply Si 9 30 10 10 7 5; 0 30 4“ 0 9 Solution We apply Vogel's approximation method to obtain an initial basie feasible solution, as shown in Table 9.5. This solution is again reproduced in Table 9.11 for ready reference. 1. In Table 9.11, the number of occupied cells are m +n ~ 1 =3 +4~ 1 = 6, and initial solution is non- degenerate. Thus, an optimal solution can be obtained. The total transportation cost associated with this solution is 779. 2 Inorder to calculate the values of up (i= 1, 2, 3) and vis (j — 1, 2,3, 4) for each occupied cell, we aubitraily assign v, = 0 in order to simplify calculations. Given vy~ 0,1, a and u, can be immediately computed by using the relation cy =u, + v for occupied cells, as shown as follows: Table 9.11 Initial Solution, VANFEZ] Operations Research: Theory and Appications yy = ty +¥y oF 20=u+0 oF uy =20 atv, oF 60=u,+0 oF u,=60 atv, or 10=4,+0 oF m= 10 Given 4, ¥, and 14, value of v,, vp and v, can also be calculated as shown below: 9 -20 lo+y, ory, Cy, = ty tus oF 40 = 60 +5 oF V5 cy ty +¥y oF 8=204¥ or yy 3. The opportunity cost for each of the occupied cell is determined by using the relation ym y= (e+ op ld is sown bolo. ey2— (y+) =30-(10-12)= ei ~ (+9) oy ~ Gy +) 3p ~ (uy +) ey ~ (uy + v4) =40-(20+ 9) zy = yy ~ (uy + 9) =70- 20-20) = 70 4. According to the optimality criterion for cost minimizing transportation problem, the current solutien is not optimal, since the opportunity costs of the unoccupied cells are not all zero or positive. The value of d.; =~ 18 in cell (Sp, D,) is indicating that the total transportation cost can be reduced in the multiple of 18 by shifting an allocation to this cell. 5. A closed-loop (path) is traced along row S, to an occupied cell (S,, D,). A plus sign is placed in cell (Sy, D,) and minus sign in cell (S,, D;). Now take a right-angle tum and locate an ozcupied cell in column D,. An occupied cell (S;, D,) exists at row S, and a plus sign is placed in this cell. ‘Continue this process and complete the closed path. The occupied cell (S,, D;) must be bypassed otherwise they will violate the rules of constructing closed path 30 - (60~ 12) 6. Inorder to maintain feasibility, examine the occupied cells with minus sign at the comersof closed loop, and. select the one that bas the smallest allocation. This determines the maximum number of units that can be shifted along the closed path. The minus signs are in cells (Sy, D;) and (S,, D,). The cell (S,, Da) is selected because ithas the smaller allocation, ie. 2. The value ofthis allocation is then added to cell (S,, D3) and (53, D,), which carry plus signs. The same value is subtracted from cells (3, D,) and (Sy, D,) because they carry minus signs. 7. The revised solution is shown in Table 9.12. The total transportation cost assoviated with this solution is shown in Table 9.12. Total cost = 5% 19+ 210+ 2% 30+ 7 «40 +6 x 8+ 12 x20 = Rs 743 Demand 3 s 7 14 u Table 9.12 ‘Optima Solution 4D 8 Test the optimality of the revised solution once again in the same way es discussed in earlier steps The values of 1,8, vs and d,s are shown in Table 9.12. Since each of djs is positive, therefore, the current basic feasible solution is optimal with a minimum total transportation cost of Rs 743.‘Transportation Problem Fag Example 9.7 A company has factories at F,, F,, and F that supply products to warehouses at W7,, 1’, and 1. The weekly capacities ofthe factories are 200, 160 and 90 units, respectively. The weekly warehouse requirements are 180, 120 and 150 units, respectively. The unit shipping costs (in rupecs) are as follows: Warehouse my m, Ws Supply A 6 20 2 200 Factory [Fy “ a 8 16 A %6 28 16 30 Demand 10 120 150 350 Determine the optimal distribution for this company in order to minimize its total shipping cost Solution Initial basic feasible solution obtsined by North-West Comer Rule is given in Table 9.13. The inital solution has m + n—1=3+ 3-15 allocations. Therefore, its a non-degenerate solution. The ‘optimality tst ean, therefore, be performed. The total transportation cost associated with this solution is: Table 9.13 Initial Solution Total cost = 16 x 180 +20 x 20 +8 x 100 +18 x 60+ 16 x 90 =Rs 6800 ‘We now determine the values of us and v,s as usual, by arbitrarily assigning w, = 0. Given 2, the values of others variables obtained by using the equation cy = u, + v, for occupied cells, are shown in Table 9.14. eyem ty or 16=0 ty of y= 16 ca=m ty, or 20-0 +y, of v= 2D oy ty, or 8=m+20 or w=-22 ‘At each stop @ no ‘occupied cell with largest negative ‘oppertunity cost is selected t0 get ‘maximum reduction in teal ‘transportation cost Table 9.14F4G] Operations Research: Theory and Applications 915 Cy = tt vy oF 30 37 Ms +¥; OF -14 ‘The opportunity cost for each of the unoccupied cells is determined by using the equation, dy = ey ~ (u, + ¥) as follows y= %y- +9) dyy= Cy, ~ (uy +¥4) = 12-0 + 30) dy, = ey, = uy +4) = 14-12 + 16) = 10 4, ~ (uy +04) = 26 (14 + 16) = 24 Cy — (uy +¥y) = 24-14 +20) = 18 ‘The value of d,, = - 18 in the cell (F,, Ws) indicates that the total transportation cost can be reduced in a muliple of 18 by introducing this cell in the new transportation schedule. To see how many of the commodity could be allocated to this cell (route) we shall form a closed path, as shown in Table 9.14, ‘The largest number of units of the commodity that should be allocated to the cell (F,, 145) is 20 units because it does not violate the supply and demand restrictions (minimum allocation among the occupied cells bearing negative sign at the comers of the loop). The new transportation schedule (solution) so obtained is shown in Table 9.15. The total transportation cost associated with this solution is Total cost = 16 x 180+ 12 x 20 +8 x 120+ 18 x 40 + 16 x90 = Rs 6,240 To test the optimality of the new solution shown in Table 9.15, we need to again calculate the ‘opportunity cost of each unoccupied cell in the same manner as discussed earlier. The calculations for us, vyp and d,s are shown in Table 9.16. ey =m tyy or 2=44+0 or R y= mtv; or 18 or 18 Gy =wytyy or 16 or 16 etm ty oF 16 or 4 (2 =mty, or & or v,=-10 dyy = Cg (ty +4) oF 20-(12- 10) = 18 4, =cy,— (y+) or 14-(18 +4) =-8 dy ~ cy +) or 26- (1644) = 6 dyy = Cyp~ (tty +2) or 24- (16-10) = 18 The valve of d,, = 8 in the cell (F, 7,) indicates that the total cost of transportation can further be reduced in a multiple of 8 by introducing this cell in the new transportation schedule, The new solution ia obtained in the same manner by introducing 40 unis ofthe commodity inthe cell (F,, WV;), 08 indicsted in Table 9.16. The new solution is shown in Table 9.17.oe 18 Transportation Problem Bxag a2 8 © eo y= 18, 180 120 150 yy = C32 — (uy + ¥2) 24 (20-6) = 10 % a4 = 10 =o Table 9.16 The total transportation cest associated with this solution is Total cost = 16 140 + 12 * 60 + 14 x 40 +8 x 120 + 16 x 90 =Rs 5,920 Table 9.17 To test the optimality of the new solution shown in Table 9.17 we need to again calculate the opportunity cost of each unoccupied cell in the same manner as discussed earlier. The calculations are shown in Table 9.18 16 2 4 = 16 +10 20 mh 4 8 8 _o @ e | 1 Fy % a 20 6 +10 0 a eecntnlanemeil Demand 180 120 150 4-0 a6 Table 9.18 dy = C42 (+9) or 20- (16-6) = 10 dh, = C3~ (ty +5) OF 18-(14-4)= 8B 4, = 65-4 +0) oF 26-2O4+0)= 6B4£:1 Operations Research: Theory and Appitcations Table 9.19 Initial Solution Since none 6f the unoccupied cells in Table 9.18 has a negative opportunity cost value, therefore, the total transportation cost cannot be reduced further. Thus, the solution shown in Table 9.18 is the optimal solution, giving the optimal transportation schedule with a total cost of Rs 5,920. Example 9.8 The following table provides all the necessary infomation on the availability of supply to each warehouse, the requirement of each market, and the unit transportation cost (in Rs) from each ‘warehouse to each market. Market ; __ 4 6 2 Warehouse zB 3 5 c 5 t Demand | 7 2 7 4 ‘The shipping clerk of the shipping agency has worked out the following schedule, based oa his own experience: 12 units from A to Q, 1 unit from A to R, 8 units from A to S, 15 units ftom B toR, 7 units from Cto P and 1 unit from C to R, (@) Check ané see if the clerk has the optimal schedule. (©) Find the optimal schedule and minimum total transport cost. (©) Ifthe clerk is approached by a carrier of route C to Q, who offers to reduce his rate in the hope of, getting some business, by how much should the rate be reduced before the clerk would offer him the business. Solution (a) The shipping schedule determined by the cletk based on Table 9.19. The total transportation cost associated with this solution is Total cost =3 «12+ 5% 144% 9+2*154+5~748%1=Rs 150 Since the number of occupied cells (ie. 6) is equal to the required number of occupied cells (i. m+ n— 1)in a feasible solution, therefore the solution is non-generate feasible solution. Now, to test the ‘optimality of the Solution given in Table 9.19 we evaluate each unoccupied cell in terms of the opportunity ‘ecst associated with it. This is done in the usual manner and is shown in Table 9.20. experience is shown in In Table 9.20, cell (C, S) has a negative opportunity cost (.e.— 1). Thus, this solution is not the optimal solution and, therefore, the schedule prepared by the shipping clerk is not optimal. (©) By forming a closed-loop to introduce the cell (C, S) into the new transportation schedule as shown in Table 9.20, we get a new solution that is shown in Table 9.21.Transportation Problem Fae Table 920 While testing the optimality of the improved solution shown in Table 9.21, we found that the ‘opportunity costs in all the uncccupied cells are positive, Thus the current solution is optimal and the optimal schedule isto wansport 12 nits from A to Q; 2 units ffom A to R; 8 units from A to S; 15 units from B to R; 7 units from C to P and 1 unit from C to S. The total minimum transportation cost associated with this solution is Total cost = 3 12+5* 244x842" 15+5% 746% 1= Rs 149 Table 921 (©) The total transportation cost will increase by Rs 2 (opportunity cost) if one unit of commodity is ‘transported from C to Q. This means that the rate of the carrier on the route C to Q should be reduced by Rs 2, ie. flom Rs 7 to Rs 5 so as to get some business of one unit of commodity only. In case all the 8 units available at C-are shipped through the route (C, Q), then the solution presented in Table 921 may be read es shown in Table 9.22.Table 9.23, Table 9.24 ‘Operations Research: Theory and Applications ‘The total cost of transportation associated with this solution is Total cost = 6x 7+3x4+5x24+4x942 x18 +7 xR =Rs 186. ‘Thus, the additional cost of Rs 37 (= 186 ~ 149) should be reduced from the transportation cost of § units from C to Q. Hence transportation cost per unit from C to Q should be at the most 7 ~ (37'8) = Rs 2.38. Example 9.10 ABC Limited has three production shops that supply a product to five warehouses. The cost of production varies from shop to shop and cost of transportation from one shop 10 a waretiouse also varies. Each shop has a specific production capacity and each warehouse has certain emount of requirement. The costs of transportation are given below: Was Te] ml wl | Sen ape fs ~s )7 | s | Ww ‘Shop B 5 6 7 4 8 2s ee [omni | |» |» | m | » | ao Th cot of munufcting he product dierent produsion shops i “Sop Tile ot Find ot 1 7000 7 ‘ero ¢ is on Fie pn ry ep Roe a a I nos Wn cost. (Dethi Univ., MBA, 1997) Solution In this case. the fixed cost data is of n0 use. The transportation cost matrix will include the given transportation cost plus the variable cost, as shown in Table 9.23 Sees i a fee ianos de t4= 18 | 7 14=21 a [s+ i6=31 Tei6=23|4+16=20| s+ t0=04 | ws c [asis=u8 ge15=21| 3-15-18 | se = ed Ce ee ‘The optimal solution obtained by applying MODI method is shown in Table 9.24 tit DO SPA ico 1s 1k 21 0 a| © 4 a fai 2 3 2 2 o| @ «| @ @ [we » 21 i 0 @ | @ a un 0s enona | o ” 8s 10s AThe transportation cost associated with the solution is 1 ‘Total cost ‘Transportation Problem 18 x 15 + 18 x BS + 22 x 20+ 20 x 105 + 18 x 60 + 19 x 45 + 19 x 70= Rs 7,605 CONCEPTUAL QUESTIONS B CDoceribo the computational preceduro of the optimality toot in 2 transportation protiem. 2. Indlcale how you wil test for optimalty of inital feasible soluton of a transportation problem, ‘Symbotcally state the transportation problem. Establish that the cptimal olufon is not altered when the oy are replaced ty cfs, A.B and C. The warehouses have the following amounts in stock: Warehouse : a bed No.ofunts : 15 161213 ‘and the customers’ requiements. are Customer: A BOC No, of unis : 18 2018 ‘Tho table blow shows the costs of transportng one unit frm warehouse to the customer. Warehouse a é @ a a_[e ° e 2 Customer | a | 6 | it 5 _ [0 ce [3 a 7 ° Find the optimal transportation routes. ‘A firm manufacturing a single product has three plants |, I! end IL They have produced 60, 35 and 40 units, respectively durng this month. The firm had made a commitment to sell 2 units to eustomrar A, 45 unite to customer B, 20 unite to customer C, 18. Units 10 customer D and 90 units to customer E. Find the ‘minimum pessibletransporation costof shifting the manuiactured product to tre five customers. The net unit cost of transportng from the thee plants to te five customers given below: 3. Lat Sand D; (i= 1, 2... mt f= 1, 2... A) be the supply es ten forse! Seatabie. “Tadqeciay, er a’comadyy aft | Mawcf eo 4u+y UE 2.5 ai andy (fa, 2 godowns and n martts. Let c be the costo transpering ene . A) are arbitary real numbers. Lit of the commodity from godswn J to market j. Assurring that SELF PRACTICE PROBLEMS B Consicer feur bases of operation | and threo tamgots 7; Tho | Customers tons of bombs fer aicraf trom any base that can be dofvered | te any target ate given in the Tolowing table | alT*® le] |= Target (7) ytafs[sf[e|[a] 7 % t Ports |v | 2 | 3 | 2 | 2 | 3 a 8 5 mi>s[s[e[sal|e« Base (6) & s 8 s 4. The folowing table gives the cost of transportng material from 3 10 8 4 supply points A,B, C and D to demand paints E, FG, H and | a 8 8 4 To ‘The dally sortie capability of each of the four bases is 150 evrite Tey ‘otis per day. The daly requirrrent of soties spread over a [oer] ae] aes ae es ‘each individual target is 200. Find the allocation of sorties from ‘ach basa to each fargo! which maximizes tha total ‘onnage Fon { a | | 3 [we [| o ‘rer all three tergets. Explain each step in the process. Sal S| eae te ‘A company has four warehouses, 8, b, ¢ and d. I Is equited te deliver a product ftom these warehouses to three customers of»elwel7l/[s|/e Tho present allocation is 25 folows: Ato E00; Ato F 10; 8 10F 150; C10 F 10;C 106 50; Cto 1120; D to H 210; D to 1 70. (8) Gheck # this allocation Is optimum. tno, tne an optima schedule. (0) thn the above problem, the transportation cost trom A to Gis reduced 10, what wil be the new optimum sctedule? AS (ia), 1952) 5. A wholesling compaay has thee warehouses from which the supplies are drawn for four rtal customers. The company seals ha single produc. the supplies of which a! each warehouse ar Supply Customer Demand (unis) “Number units 1 20 1 15 2 28 2 19 3 7 3 13 4 18 ‘Total supply at the warchouses is equal to total demand from the customers. The table below gives the transporiation costs, pe" uni, shipped from each warehouse to each custome.Operations Research: Theory and Applications So & & Cy plan, mietele ie farriee ctreueeh cece of aoa hina an ‘Warehouse | My 8 1 ® s ‘Bombay, Kolkata and Delhi, require 6,000 tonnes, 4,000 tonnes, Eales Sie tate ast ta ater eee ee mone Sam a soe oe ae Se eT BE es Se ee ae room © as aay SS nacre aa maint oi Kolkata. These centres have availabiity of 40, 20 and 40 units of Allahabad Bombay Kokata Delt Con ky sca ae |- aa ee ee ee ee re a ‘The transport cost (in rupees) per unt between each cenire outlet Baghpat 7 5 2 3 ee Ee: Gwalior 2 5 4 5 aaa Foal Gas Seq an ora traps ded nn aaa z z c a z ‘cost of transporiation, [Gujarat Univ, MBA, 1990) cane” Ae GBF | ay tua ne eee amend renee Ia noes aos Sorgeea at saenane ws Nee, oem See oe ae ee ae Bee AG so a Se eerie omet_faay Fool o50 fee scr Vineea ee ere ve i sy eee Srpron Tooele a ‘Amanutactirer has dstribution centres lcatedat Agra, Alanabad als |wo |a @s5 10 Site "pass ates tae rear eo os rican ne reese ee | Mee a cd iu ate coiet mawoonr i | Pam ele @e |e |e Ga SL see oe pene ene “ - z. oe. = e ‘Answer the following questions, giving brief reasons for he Oe ees er Kolata «4060S Cetermine the optimal shipping cost [Allahabad MBA, 1990: Dalhi Univ. MBA 1986) AA steel company is concemed with the probiem of distibuiing imported ore fem tree porte to four sleel mille, The suppios of oFe arriving at the poris are Port Tomes per week a 20,000 > 38,000 ° 16,000 ‘The demand at the steel mils is as follows: Stool mils A 8 ie > Tomes per week : 10,000 16,000 22,000 24,000 ‘Tho tvansportation cost is Re 0.05 per tonne per km, The dstance between the ports and the stee! mils is as given below: a 3 c D a|_50 oo 100 0 o | 60 40, 70 50 e | 9% 70 20 ~» | 1 (6) Is this soluton optimum? (@) Does mis problem have more man one optimum solution? Ifo, show all cf then, (©) tthe cost tor the route Bul is reduced from Rs 7 to Rs 6 per unt, what will be the opimum solution? A baking tim can produce a specialty bread in eitrer of its two Plants, the detals of which are as follows: Plant Production Capactty Production Cost Loaves Asioat A 2.500 220 8 2,100 250 Fur restaurant chains are witlng to purchase this bread; that ‘demand and the prices they are wiling 10 pay are as follows: chain ‘Maximum Demand Price Ofte Loaves Asloat 1 1,800 390 2 2300 370 4 550 400 4 1,750 360image not availableimage not availableTransportation Problem In Table 9.26, all opportunity costs ds are not positive, the current solution is not optimal. Thus, the unoccupied cell (X, C), where da, =—8 must enter into the basis and cell (W, C) must leave the basis, as shown by the closed path, The new solution is shown in Table 9.27, 2 a Daa . ‘ 0 a) 8 “6 2 is 6 = 8 | @ @ @ 2 te 2 0 gece ® 16 «| 7 102 4 20 ae i wai | mao Table 0.27 Since all opportunity costs d,s are non-negative in Table 9.27, the current solution is optimal. The total minimum transportation cost associated with this solution is: Total cost = 8 * 76 + 24x 21-4 16 X41 +0 x 20+8 x72 + 16% 5=Rs 2,424. Example 9.11 A product is manufactured at four factories A, B, C and D. Their unit production costs are Rs2, Rs3, Re 1 and Rs 5, respectively. Their production capacities are 50, 70, 30 and 50 units, respectively. ‘These factories supply the product four stores, demands of which are 25, 35, 105 apd 20 units respectively. Unit transportation cost in rupees from each factory to each store is given in the table below. Stores: 1 an m Ww 4 2 4 6 W B 10 8 7 3 Factories e B 3 9 2 D 4 6 8 3 Determine the extent of deliveries from each of the factories to each of the stores, so that the total production and transportation cost is the minimum, Solution The new transportation costs that include both the production and the transportation costs is given in Table 9.28. i i i x [Se 4 2t2=4 142 =13 30 B loss 13 5+3=8 0 c +118 oei= | aera 0 op | 4t3=9| sree | sese8 0 Demand 25 35 los 20 200 185 Since the total supply of 200 units exceeds the total demand of 185 units by 15 units, a dummy destination (store) is added (or created) to absorb the excess capacity, The associated cost coefficients in dummy store are taken as zero. This may be due to the reason that the surplus quantity remains lying in the respective factories and is not shipped at all. The modified table is shown in Table 9.29.Operations Research: Theory and Applications Table 9.29 Initial Solution Table 9.30 Degeneracy arises whan the number of ‘cospied cells are less than the number of rows + ‘columna — 4 Demand 25 Using the VAM method the initial solution ig shown in Table 9.29. It can he seen that 15 units are allocated to dummy store from factory D. This means that the company may cut down the production by 1 S.units atthe factory that is proving to be uneconomical. Now to test the optimality of the solution shown in Table 9.29 we evaluate each unoccupied cell in terms of opportunity cost associated with it in the usual ‘manner as shown in Table 9.30. 7 at B ° so | m= 8 +10 +8 a |B 8 0 0 | w= 4 8 4 4 10 B 0 30 +12 “4 #12 #1 Demand 25 3 105 20 15 % 4-9 eet | 4=8 y=8 y= 0 Since the opportunity cost in all the unoccupied cells is positive, the initial solution is an optimal solution, The toval cost of transportation associated with this solution is Total cost =4 «25 +6 x5 +8 x20 410% 7044 x 30+ 13 x15 +8 x 20 +0 x15 =Rs L465, 9.6.2 Degeneracy and Its Resolution . A basic feasible solution for the general transportation problem must consist of exactly m +m — 1 (number of rows + number of columns ~ 1) positive allocations in independent positions in the transportation table. A solution will only be called degenerate if the number of occupied cells is less than the required number, m + n— 1. In such cases, the current solution cannot be improved upon because it is not possible to draw a closed path for every occupied cell. Also, the values of dual variables u; and v; that arc used to test the optimality cannot be computed. Thus, we need to remove the degeneracy in order to improve the given solution. The degeneracy in the transportation problems may occur at two stages: (a) When obtaining an initial basic feasible solution we may have less than m +n ~ 1 allocations, (b) At any stage while moving towards optimal solution. This heppens when two or more occupied cells with the same minimum allocation are simultaneously unoccupied. Case 1: Degeneracy at the initial solution To resolve degeneracy at the initial solution, we proceed by allocating a very small quantity close to zero to one or more (if needed) unoccupied cells 50 as to get m + n~ 1 number of occupied cells. This amount is denoted by a Greck letter € (epsilon) or A (delta). This quantity would neither affect the total cost nor the supply and demand values. In a minimizationTransportation Probiem Paar transportation problem itis better to allocate A to unoccupied cells that have lowest transportation costs, ‘whereas in maximization problems itshould be allocated to a cel! that has a high payoff value. In some cases, ‘A must be added in one of those unoccupied cells that uniquely makes possible the determination of wu, and ‘the quantity 4 is considered to be so small that if itis transfered to an occupied cell it does not change the quantity of allocation. That is, ay thnx, ~ Any A-A Ata=a OFA=A: kx AHA It is also obvious then that A does not affect the total transportation cost of the allocation. Hence, the ‘quantity 4 is used to evaluate unoccupied cells and to reduce the number of improvement cycles necessary 1 reach an optimal solution. Once the purpose is over, A can be removed from the transportation table, Example 9.12 A manufacturer wants to ship 22 loads of his product as shown below. The matrix gives the kilometres from sources of supply to the destinations, Destination Dy Dy Ds Ds Ds | Swply § 5 8 6 6 3 8 Soure | 5, 4 7 7 6 5 3 Ss a 4 6 6 3 9 Demand | 4 3 3 a a Pe ‘The shipping cost is Rs 10 per load per km. What shipping schedule should be used in order to minimize the total transportation cost? [Dethi Univ., MBA, 2001] Solution Since the total destination requirement of 25 units exceeds the total resource capacity of 22 by 3 units, the problem is unbalanced. The excess requirement is hancled by adding a dummy plant, Svcs with a capacity equal to 3 units. We use zero unit transportation cost to the dummy plant. The modified transportation table is shown in Table 931. sity ® : 7 7 6 5 5 4 6 6 4 | @ @ 2 fe |a0 0 0 Ligier Demand a 4 5 4 ‘The initial solution is obtained by using Vogel's approximation method as shown in Table 9.31. Since the solution includes 7 occupied cells, therefore, the initial solution is degenerate, In order to remove degeneracy we assign A to unoccupied cell (S,, D,), which has the minimum cost amongst the unoccupied cells, as shown in Table 9.32.Operations Research: Theory and Applications Table 932 Determine w, and v, for occupied cells as shown in Table 9.32. Since the opportunity cost in the cell (Soxess: D3) '8 largest negative, it must enter the basis and the cell (S,, Ds) must leave the basis. The new solution is shown in Table 9.33 Table 933 ¥y yn yes ye 6 ye x Repeat the procedure of testing optimality of the solution given in Table 9.33. The optimal solution is shown in Table 9.34, Table 9.34image not availableOperations Research: Theory and Applications Table 9.37 Demand | 150 80 50 280 To remove degencracy quantity A is assigned to one of the cells that has become unoccupied so that there are m +n ~ 1 occupied cells. Assign A to either (S., D,) ot (Sy, D3) and proceed with the usual solution procedure. The optimal solution is given in Table 9.38, 9.6.3 Alternative Optimal Solutions ‘The existence of alternative optimal solutions can be determined by an inspection of the opportunity costs, dj for the unoccupied cells. Ifan unoccupied cell in an optimal solution has an opportunity cost of zero, ‘an alternative optimal solution ean be formed with another sct of allocations, without increasing the total transportation cst. Illustration Consider the optimal solution of Example 9.7, given in Table 9.27, For ready reference Table 9.27 is reproduced as Table 9.39. Table 9.39Transportation Problem ‘The opportunity costs in all unoccupied cells are positive except for the cell (X, A) which has a zero ‘opportunity cost, This means if (X, A) is entered into the basis, no change in the transportation cost ‘would occur, To determine this alternative solution, form a closed path for cell (X, A) as shown in Table 9.40 Table 9.40 Optimal Solution ‘The maximum quantity that can be allocated to cell (X, 4) is 21. After this change, the new solution is shown in Table 9.41, Since alld values are positive or zero, the solution given in Table 9.41 is optimal with a minimum total transporiation cost of Rs 2,424, which is same as in the previous solution. Table 241 Initial Solution, a 42 [= [w= [wed] vant ‘warehouses located in the Example 9.14 XYZ tobacco company purchases tobacco and stores following four cities: ‘Warehouse Location (City) : A B c D Capacity (Tonnes) : 90 50. 80 60, ‘The warehouses supply tobacco to cigarette companies in three citics that have the following demand: Cigarette Company Demand (Tomes) Bharat 120 Janata 100 Red Lamp noOperations Research: Theory and Applications ‘The following railroad shipping costs per tonne (in hundred rupees) have been determined: Warehouse Location Bharat anata Red Lamp 4 7 10 5 B 2 9 4 c 7 3 5 1 D 9 Because of railroad construction, shipments are temporarily prohibited trom warehouse at city A to Bharat Cigarette company. (@) Find the optimum distribution for XYZ tobacco company, (b) Are there multiple optimum solutions? If yes, identify them. (©) Write the dual of the given transportation problem and use it for checking the optimum solution. [Delhi Unix, MBA, 97, 1999, 2002] Solution Since the total demand of 330 units exceeds the total capacity of 280 units by 50 units of the product, a dummy company is created to handle the excess demand. The associated cost coefficients for the dummy warehouse location are taken as zero, Further, the cost element (.e. 7) on the route city A~ Bharat company is replaced by M, since the route is prohibited. The modified table is shown in Table 9.42. SS eats Red Lamp esas M 0 3 [2S 2 9 4 (pase 7 3 u I DB 9 5 7 [Damn ° ° ° srabeieie2: Demand 120 100 110, 330, Using the VAM method, the initial solution i shown in Teble 9.43. Now to test the optimality af the solution as shown in Table 9.43, we evaluate each unoccupied cell in terms of opportunity cost associated with it in the usual manner. This is shown in Table 9.43. | : Red Lamp 4 |M 0 s M8 al) @ | 4-8 a | a 9 a @ a Br oeee ernie: 3 it | | or wn] 427] D 9 So 7 6 @)}—_}—40)" +o} 0 [imams [0 ° 7 | @ im A een ae [Siemens 120 700 iio 30 Optimal Sohtion = yi 7 ee‘Transportation Problem Since the opportunity cost in all the unoccupied cells is positive, the intial solution shown in Table 9.43 is also an optimal solution. The total transport cost associated with this solution is Toial cost ~ 5 * 90 + 12% 30 +4 * 20 +3 * 8049 * 40 +5 * 20~ Rs 1,59,000 (0) Since opportunity cost in cell (C, Bharat), ds, = 0, there exists an altemnative optimal solution: %y3= 90, x) = 30, X33 = 20, 35) 40, X4y = 60 and x5, = 50 and total cost = Rs 1,59,000 (© The dual of the given problem is Maximize Z = (90 u + 50 iy +80 1g + 60 14 + 50 us) + (120 vy +100 ¥ + 110¥5) subjest to the constraints Xs u, + ¥)SM uy + vy S12 ut ys7 uty s9 4, +¥ S10 wytys 9 uy > v5 3 my ty ss uy tS 5 tv, 5 4 uy + vyS 11 uy tvyS7 and uy vj unrestricted in sign, for alli and j. Now by substituting the values of 1s and vs from the optimal transportation in Table 9.43, we get MaximizeZ=90 x 13 + 50* 12 + 80 x7 + 60 x9 +50 x0 +120 0+ 100 x—4 +110 x-8=Rs 1,59,000 which is the same value as obtained earier. 9.6.4 Prohibited Transportation Routes Situations like road hazards (snow, flood, ete), traffic regulations, etc., may arise because of which it is usually not possible to transport goods from certain sources to certain destinations. Such type of problems can be handled but by assigning a very large cost, say M (or 2) to that route (or cell) Example 9.15 Consider the problem of scheduling the weekly production of certain items for the next four weeks. The production cost of the item is Rs 10 for the first two weeks and Rs 15 for the last two ‘weeks. The weekly demands are 300, 700, 900 and 800, which must be met. The plant can produce a maximum ‘of 700 units per week. In addition, the company can use overtime during the second and third week. This increases the weekly production by an additional 200 units, but the production cost also increases by Rs 5. Excess production can be stored at a unit cost of Rs 3 per week. How should the production be scheduled so as to minimize the total cost? Solution The given information is presented as a transportation problem in Table 9.44. The cost elements in each cell are determined by adding the production cost, the overtime cost of Rs 5, and the storage cost of Rs 3. Thus, in the first row, the cost of Rs 3 is added during second week onward. Since the output of any period cannot be used in a period preceding it, the cost element is writen in the appropriate cells ‘A dummy column has been added because the supply exceeds demand. Table 9.44 Nore: R: Regular, O: Overtime ‘The problem can be solved by using MODI method for the usual transportation problem. The solution is Toft as an exercise for the reader. Degencracy occurs at the initial stage if initial basic feasible solution is ‘btained by Vogel’s method. Degeneracy may be removed by adding A in the cell (R, Dummy), Table 9.45 provides the optimal solution,Operations Research: Theory and Applications Table 9.45 Optimal Solution Table 9.48 Production Schedule z ma 7S © Danny SP 0 sap |? aw |* aw 700 R map |S @ 700 2. 5 8 ii 7 20 % |- = 5 6 c 700 eeseathee - 20 2B an 200, R [= = = 5 o 700 Demand | 300 700 300 800 500 3200 ‘The production schedule is given in Table 9.46 Demand 30700 M800 ‘The toial minimum cost for the optimal production schedule given in Table 9.47 is Total cost = 10 * 300 + 16 * 200+ 19 = 100 + 10 700 + 15 = 700-+ 15 * 700 = Rs 36,100 Example 9.16 ABC company wishes to develop a monthly production schedule for the next three months, Depending upon the sales commitments, the company can either keep the production constant, allowing fluctuations in inventory, or inventories can be maintained at a constant level, with fluctuating production. Fluctuating production necessitates working overtime, the cost of which is estimated to be double the normal production cost of Rs 12 per unit. Fluctuating inventories result in inventory carrying cost of Rs 2 per unit. Ifthe company fails 0 fulfil its sales commitment, it incurs a shortage cost of RS 4 per unit per month, The production capacities for the next three months are shown in the following table: Production Capacity Month Keguiar Overtime Sates My 30 30 60 My 50 0 120 My 60 30 40 Determine the optimal production schedule. [Delhi Unix, Mba, 1998, 2001, AMIE 2005) Solution The given information is presented as a transportation problem in Table 9.47. The cost elements in each cell are determined as follows (@ If tems are produced in a month forsales during the same month, there will be no inventory carrying cost. Thus, the total cost will either be the normal production cost or the overtime production cost. ‘Thus the cost elemenis Por cells (Ry, 1), (Ry, 2) atid (Ry, 3) ate RS 12 each and for cells (Q,, 1), (Oy, 3) are Rs 24 each,(i) If items are produced in a particular month for sales during the subsequent month, in adcition to the production costs (normal or overtime) inventory carrying cost at the rate of Rs 2 per month will bbe incurred. Cel Production Inventory Tal Cost Cos: (Rs) Carrying Cost (Rs) (RS) 2) 2 2 4 3) 2 4 16 (0,2) ” 2 26 (2,3) 4 4 28 3) 2 a “4 i Ifthe company fails to fulfil its sales commitment, it incurs a shortage cost of Rs 4 per unit per ‘month, in eddition to the preduction costs (normal or overtime), carrying (storage) cost at the rate of Rs 2 per month will be incurred, Cel Production Cost ‘Shortage Cost. ‘Total Cost @ @e) Rs) @ 2 4 16 2d 2 4 16 Ry) 2 8 20 (n> 4 4 28 Os.) 4 8 2 ‘The solution is left as an exercise for the reader. The initial basic feasible solution obtained by Vogel's ‘method (shown in Table 9.47) is also the optimal solution. ‘Table 9.47 Copyrighted materialOperations Research: Theory and Applications Table 9.48 Production Schedule Table 9.49 Initial Solution Table 9.50 Optimal Sctution ‘The production schedule is given in Table 9.48, 50 e 50. ie 0 = 20 40 30 = 30 = Demand 0 120 40 ‘The total minimum cost for the optimal production schedule given in Table 9.49 is Toval cost = 12 « 50+ 24 x 10 +16 x 20+ 12 50+ 16 x 20 + 12x 40+ 28 x 30 = Rs 3,400. Example 9.17 The following is the information that concerns the operations of the XYZ manufacturing ‘company. The production cost of the company is estimated to be Rs 5 per unit Month 1 Month 2 Units on order 800 1400 Production Capacity Reguartime 520 920 Overtime 250 250 Excess cost (overtime) 125 12s Storage cost 050 Formulate and solve the ebove problem as transportation problem. [Delhi Uni, MBA, 2002; AMIE 2004] Solution Thestorage cost of Re0.50 per unit per month is charged only ifproduction during the month 1 is used for supplies during month 2. Costs for supplies against first month's order from the previous month have been assumed at infinity as this is treated not only as prohibitive but also as undesirable. This is because the order quantity of first month is even less than a regular time production capacity for the same month, This problem is unbalanced as net supply is of 2,340 units while the demand is only of 2,200 units. A dummy demand centre of 140 units with supply cost zero is added to solve the problem. The dsta are summa- rized in Table 9.49 along with initial solution. ‘The initial basic solution obtained by Vogel's method (shown in Table 9.49) is updated in order to obtain optimal solution, which isshown in Table 9.50. ‘The least cost production schedule to meet the sale demand is sho bel MM pin See) M, om Mm mon = as aon a). |_| 0 rs 625 0 ° 625 oO gon | igticn| faGn eo 250 ee che cae pons 008 aRTransportation Problem Fag Production Supp Mm, My Mi, wy Regulartime 920 10 300 230 Overtime 920 20 = 1,170 30 1,400 9.7 MAXIMIZATION TRANSPORTATION PROBLEM In general, the transportation model is used for cost minimization problems. However, iis slso used to solve problems in which the objective is to maximize total value or benefit. That is, instead of unit cost c,, the anit profit or payoff p, associated with cach route, (i, is given, The objective function in terms of total profit or payoff is then stated as follows: Maximize Z= 3%, py xy The algorithm for solving this problem is same as that for the minimizstion problem. However, since we are given profits instead of costs, therefore, a few adjustments in Vogel's approximation method (VAM) {or finding initial solution and in the MODI optimality test are required. For finding the intial solution by VAM, the penalties are computed as difference between the largest and next largest payoff in each row or column. In tis case, row and column differences represent payotTs. Allocations are made in those cells where the payoff is largest, corresponding to the highest row or column difference. Since itis a maximization problem, the criterion of optimality is the converse of the rule for minimi- zation. The rule is: A solution is optimal if all opportunity costs d, for the unoccupied cells are zero or negative. Example 9.18 A company has four manufacturing plants and five warehouses, Each plant manufactures the same product, which is sold at different prices in each warehouse area. The cost of manufacturing and cost of raw materials are different in each plant due to various factors. The capacities of the plants are also different. The relevant data is given in the following tabl Plant tem 2 3 4 Manufacturing cost (Rs) per unit 2 10 8 8 Raw material cost (Rs) per unit 8 7 7 5 Capacity per wait tine 100 200 0 %0 ‘The company has five warehouses. The sale prices, transportation costs and demands are given in the following table: Warehouse Transportation Cost (Rs) per Unit Sale Price Demand “if & @ per Unit (Rs) 4 4 7 4 3 30 80 B 8 9 1 8 32 120 e 2 7 6 10 28 150 D 10 1 3 8 34 70 E 2 3 8 9 30 90 © Formulate this problem as a transportation problem in order to maximize profit (©) Find the solution using VAM method. (©) Test for optimality and find the optimal solution. CHA, June 1990) Solution Based on the given data, the profit matrix can be derived by using following equation, Profit = Sales price ~ Production cost ~ Raw material cost — Transportation cost ‘The matrix, so obtained, is shown in Table 9.49.porations Research: Theory and Applications Table 9.49 Profit Matrix ‘Table 9.49 representing profit can be converted to an equivalent minimization of loss by subtracting all the profit values in the table from the highest profit value. AS the highest profit value is 15, by subtracting all cell values including itself from it, the new values that we obtain are shown in Table 9.50. ‘The problem now becomes a usual cost minimizing transportation problem. | 6 ul 13 ° a 4 io 2 ° 120 € 6 4 7 6 ° 150 D 4 10 i 14 0 70 E 8 8 7 9 0 90 Supply 100 120 80 10 510 Solution Apply Vogel’s method to find the initial basic feasible solution, as shown in Table 9.50. Since initially the number of occupied cells was 8, which is one less than the required number, m+ n-1 = 9, therefore, the solution is degenerate. However, after making an allocation of A to the cell (D, 4), the initial solution has now become eligible for optimality test. Apply MODI method to evaluate each unoccupied cell in terms of opportunity cost associated with it in the usual manner. This is shown in Table 9.50. ‘The cell (B, 4) has a negative opportunity cost (ic. ~ 2) as shown in Table 9.50. Introduce it into the new solution by constructing a loop shown in Table 9.50. The new solution is given in Table 9.51, where ‘A has shifted from cell (D, 4) to cell (B, 4).Transportation Problem | oS eGo es ‘Since there is ro negative opportunity cost in the unoccupied cells in Table 9.51, therefore, this solution is the optimal solution. However, the zero opportunity cost in cell (D, 2) indicates the existence of an alternative solution, The total maximization profit associated with the solution is Total profit = 9 x 70 +5 x 50 +9 x 100+ 11 x 4+ 15 x 10+ 1 x 70+7 x 90= Rs 4,580, 9.8 TRANS-SHIPMENT PROBLEM Ina transportation problem, the shipment of a commodity takes place among sources and destinations. But instead of direct shipments to destinations, the commodity can be transported to a particular destination through one or more intermediate or trans-shipment points. Each of these points, in turn supply to other points. Thus, when the shipments pass from destination to destination and from source to source, we have f2 trans-shipment problem. A problem dealing with four sources and three destinations is shown diagrammatically in Figs. 9.(a) and (6). Source Destination Source Destination Shipping Routes ‘Shipping Rowtes st st Dt ” > 5a * Py D3 84 (2) Transportation (6) Trans shipment Since the flow of commodity can be in both directions, arrows are not shown in Fig. 9.2(b). The solution to this problem can be obtained by using the transportation model. The solution procedure is as fellows: If there are mm sources end 7 destinations, we shall have a transportation table of size (m + n) x (m + n) instead of m 7 as in the usual case. If the total number of units transported from all sources to all destinations is N, then the given supply at each source and demand at each destination are added to N. The demand at source and the supply at each destination are set to be equal to N. The problem can then 951 Fig. 9.2‘Operations Research: Theory and Appitcations be solved by the usual MODI method for transportation problems. In the final solution, ignore the units transported from a point to itself, ie. diagonal cell, because they do not have any physical meaning (no transportation). Example 9.19 Consider a firm having two factories. The firm is o ship its produets from the factories tothree retail stores. The number of units available at factories X and Y are 200 and 300, respectively, while those demanded at retail stores A, B and C are 100, 150 and 250, respectively. Rather then shipping the ‘products directly from factories to retail stores, it is asked to investigate the possiblity of trans-shipment. ‘The transportation cost (in rupees) per unit is given in the table: Factory {z 6 0 * 4 3 Find the optimal shipping schedule. Solution The number of units available at X and Y are 200 and 300, respectively and the demand at A, Band C is 100, 150 and 250, respectively. The maximum amount which can be transported through a factory or retail store is the total supply and demand, i.e. N= 500 units If all these 500 units are not transported through a factory or retail store then the remaining units will pay the role of dummy. The trans-shipment table is shown in Table 9.52 where 500 units have been added to the supply and demand at factory and toa retail store. ‘The initial solution to trans-shipment problem given in Table 9.53 can be obtained by putting 500 units to each route on the diagonal and making allocation in the matrix, by using Vogel's approximation method. Applying the MODI method to test the optimality of the solution given in Table 9.53. For this first determine 1, and v, and then opportunity cost d, for each unoccupied cell as shown in Table 9.53. Since the opportunity cost corresponding to cach unoccupied cell is positive, therefore, the solution given in ‘Table 9.53 is also optimal. In order to interpret the optimal solution, allocations in the diagonal cells are ignored as these values show the extra dummies have been added in order to allow as much as flow possible.Transporation Problem er Bes 8 c [sem % | x 8 3 a0 | 4, =4 #2 Yy 6 0 5 4 3 (9, Rea 2 ee > 2 ° 5 1 @ » 4 45 4 4 Pali 5 1 0 4 6a) 500 4 +9 9 2 45 ebm 9 7 3 ° 415 #2 7 4 500 750 3,000 Table 9.53, H=0 3 Initial Solution CONCEPTUAL QUESTIONS C ‘What is meant by unbalanced transportation problem? Explain the method for sdiving such a problem. Expain how 2 proft maximization transportation problem can be converted to. an equivalent cast minimization transportation problem, What Is degeneracy ih trensportaticn problems? How is & transportation problem solved when the demand and supply are ‘ot equa’? (@) Explain how 10 resolve degeneracy in a transportation probiem, (b) How does the problem of degeneracy arise in 2 transportation problem? Explain how one can overcome it 5. State a transportation problem in general tems and explain the problem of degeneracy. How does one overcome it? UAS(Matns), 1990) 3. Explain trane-shipment problem, What are the main characteristics of a transshipment protien? . Explain tow a trans-stipment problem can be solved as a ‘waneportation probiom. What is a trans-shipment problem? Explain how t can be formulated and solved as a transportction problem. [Delhi Uni, MBA, 1995, 99, 2000) 10. Explain the method for solving trans-shipment problem. SELF PRACTICE PROBLEMS C 1. A soe! compeny has twee open hearth furnaces and five roling Te mils. The transpotation cosis (pees per quit) for shoping ‘sted from furraces to rang mils are gven in the following table: | 0 am Supply mM Mm MM Suny | ai 3 4 x eGR | lve ee ~ pe | « c efsleilsileitt 2 £ ei A & yA fells |s|i? [7 4 ome ¥ = = venana | 4 | 4 |e fe |e Since there isnot enough surly, some ofthe demands at hese [What is the cptimal shipping schodule? Consider the folloving unbalanced transportation problem. Gestnations may rot be salted Suppose there are penally costs for every unsatisfied demand unt which are given by 5, 3 and 2 for destinations |, and Il, respectively. Find the optimal soluton.‘Operations Research: Theory and Applications 3. A company produces a small component for an industial | product and distibutes i to five wholosalrs at a fixad daliverad | pice of Rs 2.50 per unit. The sales forecasts indicate that the ‘ronthly delveriee wil be 3,000, 3,000, 10,000, 5,000 ard 4,000 Units to wholesalers I, I, lll IV and V, respectively. The company’s monthly production capacities are 5,000, 10,000 end 12,500 at plants 1, 2 and 3, respectively. Respective dict costs of preducton each unt are Fie 1.00, Re 0.90, and Fle (080 at plants W, X and Y. Transportation costs of shipping Unit from a plant to a wholesaler ae as follows. Wholassler ' olmwftwiyv wl oos | 007 | a1 | o2s | 015 pint x [oe [ 006 | oo8 | o12 | ore y[ or | oo | os | or | ors nd how many components each plant shoud supply to each wholesaler in ordor t> maximize is proft 4, ABC Tool Company has 2 sales fore of 25 men, who operate from three regional offices. The compary produces four basic produc lines of hand tools. Mr Jain, the sales manager, fools tat 6 salesmen are needed to distibute product line 10 to distribute product line I, 4 for product tne Ill and § salesmen fer product line IV. The cost (in Rs) per day of assignng 6. A departmental store wishes to purchase the folowing quantities of sarees! Types of sarees A B ¢ Dd E Quantiy 150 100 75 250 200 Tenders are submitted by four diferent menufacturess who undertake to supply not more than the quantities. mentioned ‘below (all ypes of sarees combined); Monutecturer wox y 2 Total quantiy 300 250 150 200 Sarees ayes ]c] ole w | 275 | 350 | 425 | 225 | 150 x | s00 | s25 | 450 | 175 | 100 Manufacture Z = y | 250 | s50 | 475 | 200 | 125 z | 308 | 275 | 400 | 250 | 175 How should the orders be placed? 7. Avompany has four fectories F, Fz, Fy and F, that manviacture the same product Production and faw material costs differ frem factory to faciory and are given in the following table in we fist ‘wo rows. The tansportaton costs from the factories to the ‘sales depots, S,,S, and S, are also given. The last two columns salesmen from each of the offices for seling each of the prodict in the table give the sales price end the total requirement at each fines wo-as follows cee ‘depot. The production capacity of each factory is given in the Product Lin last row. 7 o fm fw FL Fa Fa Fy Sales Requrement Price a{ 2 | a [a | we per Unit foglnil Gina 28: |, Production 1% 18 4 13 Regional Off ze [| | te aie e [2 | 2 [1 | 20 Raw material 0 9 2 9 costunit, [A the present ime, 10 salesmen ara allocated to office A, 9 10 S83 9 5 4 w 80 ‘office B and 7 salesmen to offs C. How many salesmen shoul | Transportaion S11 7 4 5 38 120 be assigned trom each offeo tb soll each preduct lin in odor | costunit sls 8 3 6 a 150 te minmize costs? Icentity altemate optimum soluions, if any. Supply 50 150 50 100 [Delhi Univ, MBA, 1996) 5. The Purchase Manager, Nr Shah, of the Stato Road Transport Corporation must decide cn the amount of fuel that should be ‘bought from three possible verdors. The corporation refuels its ‘buses regularly athe four depots within the area of hs operations. “The oll companies have said that they can fumish uo to the {allowing amounis of fuel during the coming month: 2,75,000 lives ty olf company 1; 550,000 lies by oll corrpany 2; and {660,000 ites by of company 3. The required amount ofthe fuel fe 1,10,000 litres by dopot 1: 2,20,000 litrea at dopot 2; 3,30,000 lites at depot 3 and 4,40,000 litres at depot 4. ‘When the transportation cos’s are adied to the bid price per lite supplied, the combined cost per lire for fuel from each vendor, Servicing a specite depot, Is. as under: Determine the most profitable production and distibution ‘schedule and the comesponding profit. The deficit production ‘should be taken 10 yield zero proft. Wite the dual of this ‘tenspertation problem and use it for checking the optiral solution, [Delhi Univ, MBA, 1996) 8. Link Manufacturing Company has several plants, three of which manufacture two principal producis ~ a standard card table and deluxe card table. A new deluxe card table will be Inoduced, which must be considered in terms of soling price and costs. The solling pres are: Standard, Rs 14:95, delure Rs 1885 and Naw Deluxo, As 21.95. Variable Costs (in Rs) Comeany't Cane? ORY? | aed Qty. “Plant A Plant B Plant C Plant Capacty Depot 1 2500 2075 2425 Gandara 400-000 ~785.~~810 A 600 Deluxe 1,050 850-860 845 BOD Depot 2 | 2500 2550 2075 New Deluxe "600 925 920 980 G70 Dopet 9 | 2450 26.00 2500 ‘Salve this problem by the Wansportaion technique for the ‘maximum contribution. Depet 4 2550 a bei a 8. Ciy Super Market keops ive diferent pattems of a particular Determine the opts! echodule. (Delhi Univ, MBA, 1088, 2000 size of adymade garment for sel. There are four manufacturers40. " available to the manager of the market to whom be cen order the required quantity. The derrand for the coming season and the maximum quantity that can be produced ty a manufacturer is. as ‘follows: Patten 1 2 3 4 6 UCut Fem — Maxi Mini ih Quanity requied 200 150 20-150 900 Manuiacturer A 8 cD Capacity 200 30 280250 ‘The following quotations have been submitted by the rranufacturers for diferent paterns: 1 2 3 4 5 u.cu Prom Max) Min) Moh [ps 100 | As 130 | fe 160 | As 70 | ne 140 a1 120 160 75 120 e[__ 15 140 155 80 120 oft 145 10 60 125 ‘The super market is seling diferent patterns at the folowing pices: Pattern 1 2 39 4 5 Salling Price (Rs) 120 180 170-80 140. How should the ouper market maragor place the ordor? Wite the dual of this transportation problem and give ts economic Interpretation, ‘A company wishes to determine an fwesment stratogy for ech Othe next four years. Five investment types have been selected. ‘Tho invectmont capital has boon allocated for cach ef the coming fur years, and rraximum investment ievels have been established fer each imestment type, There is en assumption that emounts invested in any year wil remain invested unt the end ofthe planning horizon of four year. The folowing table sunrarizes the data of this problem. The values inthe body ofthe table represent the net fetum on investment of one rupee up to fie end of the planning horizon. For example, a rupee invested in investmenttype B at he beginning of year wil crow to Fs 1.90 by the endof the fourth year, Yielding a net retum of Re 0.90. Investment Invesiment Type Fupees Made at the ‘Avaliable Bogmning A 8 CG DE (in '0009) of Year 1 080 099 060 075 1.00 500 2 085 0.65 0.40 060 0.60 600 3 030 025 030 050 020 750 4 G45 042 025 035 010 800 ‘Maximum rupees 750 600 500800 1,000 investment (in 0008) “The objective inthis problem is to determine the amount to be invested at the toginning of each year in an Irvectmont typo co 1s to maximize the net rupee retum for the four year period. ‘Solve the above tiansporation problem and get an optimal solution. Also calculate the net return on investment for the planning herizon of four-year perios. (CA, Mey 1933} XYZ Company tas provided tho folowing dala ard cook your advice on the optimum investinent strategy: ‘Trenspontaton Problem FUR) Investment ‘Not Rotum Data “Amount Mede at the (n pais) of Avaiabe Boginning Solectod Invactment eco) re po A Ss 1 95 80 70 +6) 70 2 nr , ) 3 7m 4 8 4 9 4 o 4 © 2 % Max. Invesiment (As lakh) 4050 ‘The folowing addtional information is also provide: (a) P, Q, Rand S represent the selected investments (©) The company has dacdad to have a four years investment pia, (@) The potey cf the company is that the amount invested in any year will remain so until the end of fourth year. (@) The values (paice) in the table roprosent net return on investment of one rupee til the end ofthe planning horizon (lor example, a rupee in investment P at the begiming of year wil grow to is 1.95 by the end of the fourth year, Yylelaing a return of 25 paise). Using the above, determine the optimum investment statoay. ICA Nov. 1996} 12, A manutacturor must produce a certain product in cuticient quantity in order to meet contracted sales of the next four ‘months, The production facies available for this product ae limited, and vary in diferent months. The unit cost of production ‘also changes accordingly tothe facies and personnel available. ‘The preduct may be produced in one month and then held for sale ina later month, at an estinated storage cost of Re 1 per Unt per month. No storage cost is incurred fer goods that ae sold in the same morth in which they are preduced. Preseriy there is no iwven‘ory of this product and none is desired atthe fend of four months. Given the folowing table, show how much te produce in each of ur months in arderto minimize total cost. ‘Month Contracted Maximum Unt Cost of Unt Storage ‘Sales Production Production Cost per Moth (in unit) (in enitey (Fe) (Fe) 1 20 0 4 1 2 30 50 16 1 3 50 20 5 4 4 40 50 7 1 Formulate the problem as @ twaneporation protlem and solve it 18 A company has factories at A, B and C, which supply te producis to warehouses at D, E, F and G. The factory apactes ere 290, 260 and 160, respectively for regular producton. # overtime production is utiized, the capacities. can De Increased to 300, 360 and 180, respectively. Increment unit overtime costs are As §, Rs 4 and As 6, respectively. The Current warehouse requirements are 165, 175, 205 and 165, ‘ospectvely. Unit shipping costs In rupees betwoon th facto fad the warehouses ar: To D el] Fe a fe 7 | 8 10 From B | 4 wo | 7 6 c [a 2 | 2 " Determine the oplimum dletibutlon fer the company to minimize costs,Operations Research: Theory and Applications 14. The products of three plan's X, Y and Z are to be transported 10 four warehouses 1, Ill and IV. The cost of transportation of ‘each unit from plants to the warehouses along with the normal capacities of plants and warehouses are indicated below: Warehouse | ‘company standards and evaluated. The applicants have been ‘grouped into four categories by their abillis; the grouping and vvalues of each catagory to the company are shown in the table below: ‘Applicant Category Value (Rs per monti) Number of 4 ym Avaiabie | cawegoy A 6 6 DEF Applicants x [es [aw [2] 300 1 1900 1,190 1,800 1,400 1,400 1,450 5 M1200 1,250 1,200 1,950 1,400 1,400 54 panty | 16 | 10 [ 16 | 20 500 lM 1000 1,100 1200 1/800 1,500 160045 a Faecleol as [as coo | _¥ 1800 1,500 1600 1480 1400150074 Required [300 | 300 | 00 | 500 How many applicants of each category should the personnel anager hire and for which jobs? (a) Soive the problem for minimum cost of transportation. Aro there any alternative solutions? if any, explain the methodology. ‘Overtime can be used in each plant to raise the capacity bby £0 per cent of the normal but the coresponding cost of trans-shipment will also increase by 10, 15 and 20 to the unit costs of production at each plant ‘Asem manutctuing industil chemicals bas got 3 pants Fy, Pe and P, each having capacties to preduce 300 kp, 200 ig, and 500K Tespecively of @ paricular cromical per dey. The production costs per ig in plants P,P, and P, respectively re Fre 0.70, Re 0.00 and He O.6.'Fotr buk ‘consumers have Placed orders for the products on the folowing basis ©) 15, Consumer ig Required Price Otlered ‘per Day Asiig 1 400 1.00 " 250 1.00 mM 350 1.92 wv 150 11.03 SShpping costs (paise per kg) from plants to consumers re ‘oven in tho table below Consumer ec) a [3 5 + e || pane, [e | 1 |e | 2 || Py 4 6 2 2 |i Work out an optimal schedule for tne above situation. Under \wrat concitions would you change schedule? ‘The personnel manager of a manufacturing company is in the process of fling 175 jobe in eit affront entry love okllo due fo the establishment ofa thd stit by the comeany. Union wage scale and requirement for the skis are shown in tre fllowing tate: 16 Pay Scales and Skils Requirements: Entry Lovel Skils | AB CD EF 1/000 1,100 1200 1,300 1,400 1,500 ‘Wage scale Relmorth No, required 2% 2 9 40 O17 £230 applicants fo the jobs have been tested and their aptitudes: ‘and skils forthe jobs in question have been matched against 17. Detonalt Private Lid. 1s In the business of menufaeturing and ‘seling office shits for men. It has four factories located in ‘iferent parts ot me country ana the monnly capacties of te factories in thousand aie as given below. The shirts are made in a few standard designs and colours, and each factory can, ‘make all types of shits in ary size subject to the overall capacity of the factory. Factories ' " uw uv Monthly capacity 3 45 Bs 5 From the factories, the shits are transported to five warehouses, located in five cfferant regions in India. The warehouses fh turn ‘supply tothe distrbutors and the retaiers. The monthly demand fof shins (in thoueand) from the warehouses fe ae follows: Warehouses. A B c DE (Monthly capacity 3 5 15 2 26 ‘The cost of tansporting a shirt fom a factory to a warehouse opancs on the distance between them and the cost of transporting © shit from each factoy to each warehouse is ivan in the table below: Warehouse a 8 ¢G bd € {ss f« [es o {ua [7 [s | fo Facioy m [10 | 7 | 1 | 2 | 8 w [e [fo fs [s fs How many shits are to be produced, in which factory, and how ‘are these 10 be despatched to tre warenouse so that the total cost invelved in transportation is minimizes (@) Use the NorthWest Comer Methad to get an inital feasible solution, (©) Check ifthe soluton obtained in (a) above is an optimal allocation and if not, then find the optimal solution. 18, A manufacturer of jeansis interested in developing an advertising ‘carrpaign that wil reach four diferent age groups. Advertising ‘campaigns can be conducted through TV, radio and magazines. ‘The following table gives the estimated cost in palse per exposure for each age group according to the medium employed. In adtion, ‘maximum exposure levels possible in each of the media, namely ‘TV, radio and magazines are 40, 30 and 20 milions, respectively ‘Ase the minimum desired exposure within each age group, namely 48-18, 19-25, 26-35, 36 and oldersar0:30, 25, 15 and 10 milone, ‘The objective is to minimize the cost of ataining the minimum ‘exposure level in each ege gioup.Meda | ‘Age Groups 13:18 19-25 26-85 96 and older Ww 12 7 0 10 Radio 10 9 2 10 Magazine! 14 12 9 2 (2) Formulate the above as a transportation problem, and find ‘te optimal solution (0) Solve ths problem if te policy is to provide at least 4 million exposures through TV inthe 18-18 age group and least 8 milion exposures through ‘TW in the age group 19-25. ICA, May 1993), 19. Two drug companies rave invertories of 1.1 and 0.9 milion doses of a particular fis vaccine, and an epidemic of the flu ‘seems imminent in three stales. Since the flu could be fatal to Senior citzens, itis imperaive that they be vaccinated fist ‘others wil be vaccinated on a firs-come-fist-served basis while the vaccine supply lasts. The amounts of vaccine (in milons of doses) each state estimates it could administer are as folows: Elders Others 0.335 0.760 “The shipping costs (pabe per dese) between drug companies land states are as follows: 0.260 800 0.195 0.650 Transportation Probiem FAW To A © Suppy x [7 3 4 2 From y [2 1 a 2 z [3s a [e 5 Demand 4 1 @ 22. A manufacture: wants to ship 8 loads of his product as stown In the table. The matix gves the mileage trom origin to destination. Shipping costs are is 10 per load per mile, What Shipping scheaule should be uses? Saray o [ % | % | 2 1 o, [0 | « | 1 o, [#0 | 20 | = Demand 4 2 2 Trans-shipment 23. Consider the folowing trans-shipment problem wih two sources land three destinations, the cost for shipments (in rupees) is, siven below. State 1 State 2 State 3 Source Destination ‘Sup Company 1 30 20 60 s ay 4 Company 2 10 40 10 % 0 wo 20 30 1002800 Souse | § 10 2 © 5) 40. «200300 Determine a minimum-oost shipping schedule which wil provide each state win at least enough vacche to care for s6n10 |p oenagoy (4 29 99 2 $2 wD Chizane. Wet the cial of th tranpotaten problem and Use Soi te At gt ieee SER ee he, Demand 800 300 100+900 100 +900 100 + 300 [Delhi Unt, MEA (HCA), 1990, 95) 20. A leading firm has three audtore. Each auditor can work Up to 160 hours during the next morth, during which tme thre projects. must be completed. Project 1 wil take 130 our Project 2 will take 140 hours, and project 3 wil take 160 hours. ‘The amcunt, per hour that can be billed for assigning each audtor to each project is given in the table Project Formulate this as a trarsporaton protlem and tind te optimal soluion. Also find out the maximum total biling during the next ‘mont. ICA, May 1999, Degeneracy 21. Obtain an optimum basic leasible solution to the following ‘degenerate transportation problem, HINTS ANI ‘Total requirement 30) < Toul capacity (34), add a dummy mill with requirement (34 — 30) = 4. Degeneracy occur atthe initial solution (VAM) Determine the optimal shipping schecue. 24, A firm having two sources, S; and &, wishes to hip is preduct to two destinations, D, and D,. The funber of units avallabie at SS, and S, are 10 and 30 and the product demanded at O, and , are 25 and 15 urits, respectively. The firm instead of shipping {om sourees te destinations, decides to investigate the possbilty, of transshipment. The unit transportation cost (in As) is given In the following table: ‘Source ‘Destinaion ‘Supply § 8 Oy Be ‘fouve Ss 0 3 4 5 10440 S$ 3 0 3 5 30240 Destination [2 4 3 ° 2 «0 came (2 $3 2) 3) 8 40 4025440 15440 Determine the optimal shipping schecue ID ANSWERS Ase 539 = 4) 4 44 94 = 25 — 85 4G) ~ 4, yy ~ Cam 5g = 4 Total cost = Rs 80.COperatons Research: Theory and Applications Demand (145) > Supply (105), add dummy source with supply (145 — 105) = 40 and transportation costs 5, 3 and 2 for estimation 1, 2 and 3, respectively. Xap = 10, x3, = 10,45, = 5 and X43 = 40. RS SIS, Penalty for transportation of 40, units t0 destination 3 at the cost of Rs 2 per unit = Rs 40. Thus, Xz = 168, x5, = 20, 55 38s. ‘Total cost = Rs 51,700. 6. Profit maximization problem. First add a dummy column with demand of 125 sarees and then subtract all clements ofthe profit matrix from the highest element 475. 110, 2g) = 440, 6p total cost = 515 + 80 = Rs $95, Mannfcturer Variety Quantity 3. Maximization és well as unbalanced problem. Total capacity (27510) > donand 23000), Add» danny ole wih ADE 25,50, 300 cemand 2,500 nis. Sine direst costs of production of each unit ae age ae Re 1, Re 0.0 and Re 0.80 plants I, 2 and 3, ad this cost Ke is 1b each Figure 1ow wise o get exect data ofthe problem. Ans.) = 2500, Sy, = 500, X22 = 3,000, 233 = 2500, x5 = | 7. Profit maximization problem, First ald dammy row with supph 4800, x4) = 7.300 and x54 = 5,000. Tol cost Rs 23.730. |” OC $o units and thew bene all ements ofthe prof maain Since in total 25,000 units ‘are Supplied to the wholesslers Bon Aigice emet b at a fixed price of Rs 2.50 per unit, therefore, the tot cot is J ese Profit = Sales price ~ (Production cast + Raw material cost Te net maximum profit tothe manufacturer = Rs 62,500 — Teasapactaton’ 904) Rs 23,730 = Rs 38,770 73 Fact ‘Sales Depot Unit Profit 4, Supply (26) > Demand (25), add dummy column with demand 1. get YS poe a Ty Ans > Aanys = 1s ig = Soma ~ 6 95 ~ 3) = 2 5 < ayy = I Total cost = Rs 472 A SS bad Alternative solution exists because opportunity cost is zero in 5 %: 2 cell (B, 4) and (C, 4). Fa SS as 5. Supply (145,000) > Demand (11,00,000), add dummy column with demand of 3,85,000 lites of oil; xy ~ 110, xy) ~ 55, a ret Rene 12, Transportaton cost ible and optimal solution Fronffo | it | Ni esaegeS: Dummy | Supply 1 14 0 40. = 8 0 20) 50 5 ae 6 ° 30 PRE = = 7 o Demand 20 30 0 0 30 170 13, Transport ae ae 10 ° 230 6 0 aay 1 v 180 5 v n a 8 1% Tr jo 0 ® Gq 19 2 x 7 0 : te Demand 16s 135 208 168 wo | 880Transportation Problem Fg 22. Using NWCM, initial solution is degenerate. a Ans, 5 2 ay hy hy 2, ky = 4 and X= 1, ‘Total cost = Rs 33. HAPTER SUMMARY ‘Transportation problems (and their variants) are special types of linear programming problems that have a variety of important applications. A transportation problem is concerned with distributing a commodity from given sources to their respective destinations. Each source has a fixed supply of the commodity and each destination has a fixed demand. A basic assamption is that the cost of distribution from each source to each destination is directly proportional to the amount distributed. Formulating « transportation problem requires constructing a special purpose table that gives the unit costs of distribution, the supplies and the demands, The objective of this chapter is to enable readers to recognize a problem that can be formulated and analyzed as a transportation problem or as a variant of one of these problem types. CHAPTER CONCEPTS QUIZ True or Fal 1. The test for degenecay of a solution is to check if there ere lurused cells wity an improvernent index equal to zero. It $0, degenecey exists 2, The advantage of the mest method over the stepping stone ‘method of computing improvement Indices for unused cells las in the grealer computational eficiency. 3. Ifyou have to salve a transporiation problem with m rows and 1 columns using the simplex method, you would have to formulate the problem with m variables and n constrains, |4. A aoltion 10 the cost minimization trareportation problem ie ‘optimal when all unused cells have an improvement index which Is. nor-negatve. 5. A dummy row of cclumn is introduced in the transportation ‘mathod is order io handle an unbalanced problem. The dummy serve, the same purpose as @ slack veriable in the simplex ethos. 6. Each tration of the tranepertation mathod involves the elimination ‘of one occupied cell and the Intioduction of one unoccupied call hich similar © a pivot in the simplex method. 7. In the transportaion problem, extra constaint equation can rot be derved rom any other constraint equations as f affects the feasible solution of the problem. 8. In the north:west corer method, the cost of transportation on ary route of transportation ic taken into account 8. The unbalanced transportation protiem can be balanced by acing a durmy supply row of a demand column as per te need. 410. In the transportation problem, the rim requirement for a row is the capacity of a supplier. white a rim requirement for a columa is the demand of a user. Fill in the Blanks 11, The northwest comer rule provides @ for obtaining an soluton to the transportation probe. 12, The degnecacy may occurs when there are two of more cals wih the same smallest value in a closed fer an incoming cal 18, Wo have alternative optimal solutions to a minimization lrensportation problem whenever we find a soluvon where the improvementindices are all with at east 14. An improvement index in the trarsporation method is analogous toa value in the quantity column in the ‘mathod Th rvee the sare purpose forthe tranepertaton method a8 all slack variables in the simplex methed. 16. An unused cell in the transporation table is analogous to a vatiable notin the solution column in the simplex method. ‘The end result of an iteration in the transportation method is the same as the end resut in the simplex method in that value of the solution is 18. In Voge's approxinatien method, each allocation fs made on the basis of the cost that would have been incurred i allecation in ‘arain cots wit Unit transpertaton ‘cost were missed. ‘method Is based cn the concept of dualty ifthe taal evpply total demand, thon an addtional ‘column known a ‘added to the transportation table to absorb the same 46 Multiple Choice 21. The inital sclution of 2 trarsporation protiom can be obtained by applying any known method. However, the only condtion is that (a) ) the solution be optimal the rim condtions are sated {(c) the soltion not be degenerate {(@) all of the above ‘The dummy source oF destination in a transportation problem i= added (a) satsty tim conciions: (©) prevent soluton fom becoming degenerate (©) ensure that total cost does not exceed a limit (2), none of the above ‘The occurrence of degeneracy while solving a. transportation problem means that (@) ‘total supply equals total demané (©) the soliton so obtained is not feasible (©) the few allocations become negative (@) none of the above {An altemative optimal solution to a minimization trenspertation preblem exists whenever opportuni cost corresponding, 19 unused route of ansportaton is 22 2