This action might not be possible to undo. Are you sure you want to continue?

Welcome to Scribd! Start your free trial and access books, documents and more.Find out more

N. D. Vohra

Quantitative Techniques in Management

© 2010

Chapter 3

**1. Formulation Conditions for Application of Simplex Method 2. Assumptions Simplex Method Maximization Problem
**

a) b) c) d) Standardization of the Problem Obtaining Initial Solution Testing the Optimality Revised, Improved Solution

**3. Justification and Significance of Elements in the Simplex Tableau 4. Simplex Method Minimization Problem
**

a) Big M Method b) Two-phase Method

( continued)

5. Special Topics

a) b) c) d) Multiple Optimal Solutions Infeasibility Unbounded Solution Degeneracy

**1. Non-negative bi values
**

Multiply a constraint on both sides by -1 if it involves a negative bi value

**2. Non-negative variables
**

Replace every unrestricted variable by difference of two nonnegative variables

( continued)

Z = 8x1 + 5x2 +12x3 2x1 + 5x2 + 4x3 44 3x1 7x2 6x3 10 7x1 + 2x2 +4x3 = 54 x1, x2 0 x3: unrestricted in sign Should change as follows: (i) Let x3 = x4 x5, where x4, x5 0 (ii) 3x1 7x2 6x3 10 be replaced as 3x1 + 7x2 + 6x3 10 Revised LPP: Maximize Z = 8x1 + 5x2 +12x4 12x5 Subject to 2x1 + 5x2 + 4x4 4x5 44 3x1 + 7x2 + 6x4 6x5 10 7x1 + 2x2 + 4x4 4x5 = 54 x1, x2, x4, x5 0 Maximize Subject to

**(all constraints of type)
**

1. Standardise the problem

Introduce slack variables

**2. Obtain initial solution (with slack variables) 3. Test if solution is optimal
**

To be optimal, variables

ij

0 for all

4. If optimal, stop and exit; else go to step 5 5. Obtain a revised solution and go back to step 3

Example 3.2 data

Max Z = 5x1 + 10x2 + 8x3 St 3x1 + 10x2 + 2x3 60 4x1 + 4x2 + 4x3 72 2x1 + 4x2 + 4x3 100 x1, x2, x3 0 Contribution Fabrication Hrs Finishing Hrs Packaging Hrs

Conditions for application of simplex method are both satisfied here We need 3 slack variables to convert the inequalities in to equations The problem becomes:

Max St Z = 5x1 + 10x2 + 8x3 + 0S1 + 0S2 + 0S3 3x1 + 10x2 + 2x3 + S1 = 60 4x1 + 4x2 + 4x3 + S2 = 72 2x1 + 4x2 + 4x3 + S3 = 100 x1, x2, x3, S1, S2, S3 0

Simplex Tableau 1

Basis S1 0 S2 0 S3 0 Cj Sol j x1 3 4 2 5 0 5 x2 5* 4 4 10 0 10 x3 2 4 5 8 0 8 S1 1 0 0 0 60 0 S2 0 1 0 0 72 0 S3 0 0 1 0 100 0 Z=0 bi 60 72 100 bi/aij 12 18 25

Key column Key element Includes variables that yield identity matrix

Key row

**To derive values in the next tableau,
**

Divide each element of key row by key element: 3/5, 5/5, 2/5, 1/5 etc. For other rows (for example row 2):

4 4 × 3/5 = 8/5 4 4×1=0 4 4 × 2/5 = 12/5 0 4 × 1/5 = -4/5 etc.

Simplex Tableau 2

Basis

x2 10

x1 3/5 8/5 -2/5 5 0 -1

x2 1 0 0 10 12 0

x3 2/5 12/5 17/5 8 0 4

S1 1/5 -4/5 -4/5 0 0 -2

S2 0 1 0 0 24 0

S3 0 0 1 0 52 0

bi 12 24 52

bi/aij 30 10

260/17

S2 0 S3 0 Cj Sol j

120

**Simplex Tableau 3: Optimal Solution
**

Basis

x2 10

x1 1/3 2/3 -8/3 5 0

-11/3

x2 1 0 0 10 8 0

x3 0 1 0 8 10 4

S1 1/3 -1/3 1/3 0 0 -2/3

S2 -1/6 5/12

-17/12

S3 0 0 1 0 18 0

bi 8 10 18

x3 8 S3 0 Cj Sol j

0 0 -5/3

160

Solution optimal as all j 0 Optimal Product mix: x1 = 0, x2 = 8 and x3 = 10 Maximum contribution = Rs 160 Capacity utilization: Fabrication and Finishing Full; Packaging 18 hours unutilized No production of x1 since every unit of it produced would result in a loss of Rs 11/3 Production of a unit of x1 will result in losing 1/3 unit of x2 and 2/3 unit of x3 together with a release of 8/3 hours of Packaging time

**1. Standardise the problem
**

Introduce surplus and artificial variables for all constraints Introduce artificial variables for all = constraints In the objective function assign coefficients to each of the artificial variables [For max problems: assign M and for min problems: assign M]

2. Obtain initial solution (with slack/artificial variables) 3. Test if solution is optimal (to be optimal, ¨ij 0 for all variables) 4. If optimal, stop and exit; else go to step 5 5. If not, obtain a revised solution and go back to step 3

Needed to be used to obtain initial solution Introduced where a constraint is of ³ ´ or ³ = ´ type May also be required in maximisation problems Expected to leave the basis one-by-one in successive simplex iterations Solution infeasible as long as one or more artificial variables present in the basis Not expected to be present in the basis of the final solution In case artificial variable present in basis of final solution, there is infeasibility

**Phase I 1. Standardise the problem
**

For each artificial variable, assign co-efficients in the objective function For max problem: assign 1 For min problem: assign 1 For every other variable, assign coefficient of zero

2. Solve by Simplex Method. If the objective function value for optimal solution is zero, move to Phase II

Phase II 1. Begin with Simplex Tableau of Phase I final solution, eliminate all entries for artificial variables and replace the zero co-efficients of decision variables by original coefficient values 2. Solve the modified problem by Simplex Method

**No artificial variable present in the basis
**

It is a feasible solution

It is a feasible solution and the ¨ij values of all non-basic variables are 0 for a max problem (or 0 for a min problem)

The solution is optimal

It is an optimal solution and the ¨ij values of all non-basic variables are < 0 for a max problem (or > 0 for a min problem)

The solution is optimal and unique

It is an optimal solution and the ¨ij value of one or more non-basic variables is equal to zero

The solution is not unique: the problem has multiple optimal solutions

**It is final in terms of the ¨ij values and has an artificial variable in the basis
**

It indicates infeasibility

**It is non-optimal and the elements of the key column are all zero/negative
**

The problem solution has unbounded

A solution in which a basic variable has solution value equal to zero is degenerate solution The basic variable which has solution value of zero is called degenerate variable Whenever there is a tie in the replacement ratios (bi/aij) to be selected, the next solution will be a degenerate solution A degenerate solution may or may not be optimal Revision of a non-optimal solution leads to no improvement (in terms of the Zvalue) if the degenerate variable is the outgoing variable

Max Stt

**Z = 28x1 + 30x2 6x1 + 3x2 18 3x1 + x2 8 4x1 + 5x2 30 x1, x2 0 With slack variables S1,S2 and S3, solution follows:
**

Tie

Simplex Tableau 1

Basis

S1 0

x1 6 3 4 28 0 28

x2 3 1 5 30 0 30

S1 1 0 0 0 18 0

S2 0 1 0 0 8 0

S3 0 0 1 0 30 0

bi 18 8 30

bi/aij 6 8 6

S2 0 S3 0 Cj Sol j

Key column

Next solution would be degenerate

( continued)

Simplex Tableau 2: Non-optimal Solution Basis

S1 0

x1 18/5 11/5 4/5 28 0 4

x2 0 0 1 30 6 0

S1 1 0 0 0 0 0

S2 0 1 0 0 2 0

S3 -3/5 -1/5 1/5 0 0

bi 0 2 6

bi/aij 0

10/11

S2 0

X2 30

15/2

Cj Sol j

180

**-6 Degenerate outgoing variable Simplex Tableau 3: Optimal Solution x1 1 0 0 28 0 0 x2 0 0 1 30 6 0 S1 5/18
**

-11/18

Basis

x1 28

S2 0 1 0 0 2 0

S3 -1/6 1/6 1/3 0 180 -16/3

bi 0 2 6

S2 0 x2 30 Cj Sol j

-2/9 0 0 -10/9

( continued)

A tie in the minimum replacement ratios in Tableau 1 results in the solution in next tableau to be degenerate (even if the outgoing variable was chosen to be S1 instead of S3) Since all values in the key column are positive, the least ratio to be considered is 0 and the outgoing variable (S1) is degenerate As a result, there is no change in the objective function value: it remains at 180 The solution in Tableau 3 is also degenerate While the solution in Tableau 2 is nonoptimal, the solution in Tableau 3 is optimal

**Mark the wrong statement:
**

1. An LPP with n variables and m constraints gives nCm basic solutions. 2. A basic solution with all nonnegative variables is called basic feasible solution. 3. For any LPP, there is a unique extreme point of the feasible region corresponding to every basic solution. 4. The basic feasible solution that maximises or minimises, as the case may be, is called optimal solution.

**Mark the correct alternative. A feasible solution is the one which
**

1. satisfies all problem. constraints of the

2. is necessarily an optimal solution. 3. makes use of all available resources. 4. yields more than one way to achieve the objective.

**Which of the following statements is not true about application of simplex method?
**

1. The RHS of each constraint should be greater-than-or-equal-to zero. 2. All decision variables of the problem should be non-negative. 3. All constraints of the given problem should be either e or u type. 4. All constraints should be converted into = type, with slack/surplus variables which, like decision variables, are also non-negative.

**Choose the incorrect statement:
**

1. Simplex method is an iterative process wherein variables are substituted until optimal solution is reached. 2. Successive simplex tableaus always yield better solution in terms of the objective function. 3. For a minimisation problem, the optimal solution is reached when all (j u 0, with no artificial variable in the basis. 4. The key column indicates the incoming variable and the key row represents outgoing variable.

If (j = cj zj is equal to 24 7M for x1; 28 7M for x2; 45 4M for x3 and 28 2M in respect of x4 in a minimisation problem, then the entering variable would be:

1. x1 2. x2 3. x3 4. x4

**Mark the wrong statement:
**

1. First Key element can never be negative. 2. Key element cannot be zero. 3. Key element has to be positive. 4. Key element can be negative, zero or positive.

The bi values and the elements in key column are (12, 0, 18, 6) and (-2, 8, 1, 3) respectively. Which of the rows will be selected as the key row?

1. First 2. Second 3. Third 4. Fourth

**Which of these is not true?
**

1. All basic variables in an LPP have (j = 0. 2. A slack variable cannot be there in the basis of optimal solution. 3. If a problem has an optimal solution, the artificial variables, if any, introduced must be driven out of the basis one-by-one. 4. An outgoing variable in a simplex tableau cannot re-enter the basis in follow-up iterations.

**Which of the following statements is not true about artificial variables?
**

1. They are introduced when initial solution to a problem cannot be obtained for want of identity submatrix in the simplex tableau. 2. Each of such variables has to be assigned a very large negative coefficient in the objective function when it is of the maximisation type. 3. They help to obtain an initial feasible solution to the problem. 4. They bear no tangible relationship with the decision problem.

**Mark the wrong statement:
**

1. If a non-basic variable in the optimal solution to a problem has (j = 0, it indicates multiple optimal solutions. 2. If an artificial variable is found in the basis of the final solution to an LPP, it implies feasibility. 3. A constraint with e sign does not require artificial variable to be introduced. 4. A slack variable is used to convert u type and a surplus variable is used to convert e type of constraint into equality.

**Mark the wrong statement:
**

1. If there are no non-negative replacement ratios in a simplex tableau, then unboundness is indicated. 2. In a tableau, the key column elements are 5, 0, 0, 7 while the corresponding bi values are 20, 6, 8 and 0. This indicates unbounded solution. 3. A tie in the minimum replacement ratio implies that the next solution would be degenerate. 4. It is possible for the initial solution to an LP to be degenerate.

**Which of these is not correct?
**

1. The solution to a maximisation LPP is unique if (j values for all nonbasic variables are less than zero. 2. If infeasibility is present, it is detected in phase I of the two-phase method. 3. In two-phase method, the artificial variables are each assigned a coefficient of 1 in case of minimisation, and 1 in case of maximisation problems. 4. If the given problem has an optimal solution, the artificial variables are all removed one by one in phase I of the two-phase method.

If there is no non-negative replacement ratio in a solution which is sought to be improved, then the solution is indicated to be

1. infeasible 2. unbounded 3. degenerate 4. unique optimal

**Which of the following is not correct?
**

1. As and when a degenerate variable is outgoing, no change will take place in the objective function value. 2. Degeneracy may be a temporary phenomenon. 3. It is possible for the repetition of same tableaus in the course of successive iterations when degeneracy is encountered. 4. A degenerate solution cannot be optimal.

Degeneracy in LPP (1) renders the solution infeasible, (2) leads to multiple optimal solutions, (3) increases computations without affecting the solution as long as the outgoing variable happens to be degenerate. Which of these statements is/are correct?

1. 2. 3. 4. 1, 2 and 3 1 and 3 2 and 3 3 only

- 1 Successive Approx AUTO
- Chapter 6 Assignment Problems
- Quadratic inequalities.docx
- Chapter_5_Transportation and Transshipment Problems
- Solving Systems of Equations
- Chapter 2 LP
- Scan Doc0002
- N.D Vohra
- numerical methods ch 06/10
- Chapter 9 Inventory Management
- Non-Homogeneous Equations_Undetermined Coefficients
- Chapter_4_LP Duality and Sensitivity Analysis
- ot
- Quantitative Techniques for Management
- Método de Frobenius
- B Simplex Method
- ADVANCE OF SMARANDACHE APPROACH TO SOLVING SYSTEMS OF DIOPHANTlNE EQUATIONS
- Chapter 1
- Application of Direct Methods (IROS 2004)
- chp1
- Lp Notes
- Chapter 17
- Solving Linear Equations
- Simplex Method - 4th Semester - Numerical Programming
- soluções de exercicios
- Quantitative Technique
- Solving Polynomial Equations
- Ignited Minds
- Maranas_and_Floudas,_Computers_and_Chem._Eng.,_1997.pdf
- Metrology Lab -Manual

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

We've moved you to where you read on your other device.

Get the full title to continue

Get the full title to continue listening from where you left off, or restart the preview.

scribd