Operations Research

Operations Research by Engr. Shahid Ismail 2-1

 The Essence of the Simplex Method
 Algebraic procedure
 Underlying concepts are geometric
 Example
 Figure shows constraint boundary lines
 Points of intersection are corner-point solutions
 Five points on corners of feasible region are CPF
solutions
 Adjacent CPF solutions
 Share a constraint boundary

Operations Research by Engr. Shahid Ismail 2-2

 Example

Operations Research by Engr. Shahid Ismail 2-3

 The Essence of the Simplex Method

Operations Research by Engr. Shahid Ismail 2-4

 The Essence of the Simplex Method
 Optimality test
 If a CPF solution has no adjacent CPF solution that
is better (as measured by Z):
 It must be an optimal solution
 Solving the example with the simplex method
 Choose an initial CPF solution (0,0) and decide if it
is optimal
 Move to a better adjacent CPF solution

 Iterate until an optimal solution is found

Operations Research by Engr. Shahid Ismail 2-5

 Setting Up the Simplex Method
  
First step: convert functional inequality
constraints into equality constraints
 Done by introducing slack variables
 Resulting form known as augmented form

 Example: constraint is equivalent to

Operations Research by Engr. Shahid Ismail 2-6

 Setting Up the Simplex Method

Operations Research by Engr. Shahid Ismail 2-7

 Setting Up the Simplex Method
 Augmented solution
 Solution for the original decision variables
augmented by the slack variables
 Basic solution
 Augmented corner-point solution
 Basic feasible (BF) solution
 Augmented CPF solution

Operations Research by Engr. Shahid Ismail 2-8

 Setting Up the Simplex Method
 Properties of a basic solution
 Each variable designated basic or nonbasic
 Number of basic variables equals number of
functional constraints
 The nonbasic variables are set equal to zero

 Values of basic variables obtained as

simultaneous solution of system of
equations
 If basic variables satisfy nonnegativity
Operations Research by Engr. Shahid Ismail 2-9
 The Algebra of the Simplex Method

Operations Research by Engr. Shahid Ismail 2-10

 The Simplex Method in Tabular Form

 Tabular form
 Records only the essential information:
 Coefficientsof the variables
 Constants on the right-hand sides of the
equations
 The basic variable appearing in each equation

 Example shown in Table on next slide

Operations Research by Engr. Shahid Ismail 2-11

 The Simplex Method in Tabular
Form

Operations Research by Engr. Shahid Ismail 2-12

 The Simplex Method in Tabular Form

 Summary of the simplex method

 Initialization
 Introduce slack variables
 Optimality test
 Optimal if and only if every coefficient in row 0 is
nonnegative
 Iterate (if necessary) to obtain the next BF solution
 Determine entering and leaving basic variables
 Minimum ratio test

Operations Research by Engr. Shahid Ismail 2-13

 Tie Breaking in the Simplex Method
 Tie for the entering basic variable
 Decision may be made arbitrarily
 Tie for the leaving basic variable
 Matterstheoretically but rarely in practice
 Choose arbitrarily

 Condition of no leaving basic variable

Z is unbounded
 Indicates a mistake has been made

Operations Research by Engr. Shahid Ismail 2-14

 Tie Breaking in the Simplex
Method
 Multiple optimal solutions
 Simplex method stops after one optimal BF
solution is found
 Often other optimal solutions exist and
would be meaningful choices
 Method exists to detect and find other
optimal BF solutions

Operations Research by Engr. Shahid Ismail 2-15

 Adapting to Other Model Forms
 Simplex method adjustments
 Needed when problem is not in standard
form
 Made during initialization step

 Artificial-variable technique
 Dummy variable introduced into each
constraint that needs one
 Becomes initial basic variable for that
equation
Operations Research by Engr. Shahid Ismail 2-16
 Adapting to Other Model Forms
 Types of nonstandard forms
 Equality constraints
 Negative right-hand sides

 Functional constraints in greater-than-or-equal-to

form
 Minimizing Z

 Solving the radiation therapy problem

 Text reviews two methods: Big M and two-phase

Operations Research by Engr. Shahid Ismail 2-17

 Adapting to Other Model Forms
 No feasible solutions
 Constructing an artificial feasible solution may
lead to a false optimal solution
 Artificial-variable technique provides a way to
indicate whether this is the case
 Variables are allowed to be negative
 Example: negative value indicates a decrease in
production rate
 Negative values may have a bound or no bound

Operations Research by Engr. Shahid Ismail 2-18

 Postoptimality Analysis
 Simplex method role

Operations Research by Engr. Shahid Ismail 2-19

 Postoptimality Analysis
 Reoptimization
 Alternative to solving the problem again
with small changes
 Involves deducing how changes in the
model get carried along to the final simplex
tableau
 Optimal solution for the revised model:
 Will be much closer to the prior optimal
solution than to an initial BF solution
constructed the usual way
Operations Research by Engr. Shahid Ismail 2-20
 Postoptimality Analysis
 Shadow price
 Measures the marginal value of resource i
 The rate at which Z would increase if more
of the resource could be made available
 Given by the coefficient of the ith slack
variable in row 0 of the final simplex
tableau

Operations Research by Engr. Shahid Ismail 2-21

 Postoptimality Analysis
 Sensitivity analysis
 Purpose: to identify the sensitive parameters
 These must be estimated with special care
 Can be done graphically if there are just two
variables
 Can be performed in Microsoft Excel

Operations Research by Engr. Shahid Ismail 2-22

 Postoptimality Analysis

Operations Research by Engr. Shahid Ismail 2-23

 Postoptimality Analysis
 Parametric linear programming
 Study of how the optimal solution changes
as many of the parameters change
simultaneously over some range
 Used for investigation of trade-offs in
parameter values
 Technique presented in Section 8.2

Operations Research by Engr. Shahid Ismail 2-24

 Computer Implementation
 Simplex method ideally suited for execution
on a computer
 Computer code for the simplex method
 Widely available for all modern systems
 Follows the revised simplex method

 Main factor determining time to solution

 Number of functional constraints
 Rule of thumb: number of iterations required equals
twice the number of functional constraints

Operations Research by Engr. Shahid Ismail 2-25

 The Interior-Point Approach to Solving
Linear Programming Problems
 Alternative to the simplex method developed
in the 1980s
 Far more complicated
 Uses an iterative approach starting with a
feasible trial solution
 Trial solutions are interior points
 Inside the boundary of the feasible region
 Advantage: large problems do not require
many more iterations than small problems
Operations Research by Engr. Shahid Ismail 2-26
 The Interior-Point Approach to Solving
Linear Programming Problems

Operations Research by Engr. Shahid Ismail 2-27

 The Interior-Point Approach to Solving
Linear Programming Problems
 Disadvantage
 Limited capability for performing a
postoptimality analysis
 Approach: switch over to simplex method

Operations Research by Engr. Shahid Ismail 2-28

 Conclusions
 Simplex method
 Efficient and reliable approach for solving linear
programming problems
 Algebraic procedure

 Efficiently performs postoptimality analysis

 Moves from current BF solution to a better BF

solution
 Best performed by computer except for the very
simplest problems

Operations Research by Engr. Shahid Ismail 2-29