Scribd
Upload
329 views

Advanced Linear Programming

This document provides an overview of the simplex method for solving linear programming problems. It discusses key concepts like extreme points, basic feasible solutions, and the generalized simplex tableau. The simplex method works by starting at an initial basic feasible solution and iteratively moving to adjacent basic feasible solutions until an optimal solution is found. It describes how to determine leaving and entering variables using optimality and feasibility conditions. The revised simplex algorithm is presented in pseudocode form. Examples are provided to illustrate generating a simplex tableau and performing simplex iterations.

Uploaded by

Joshua Nathanael
Copyright
© © All Rights Reserved
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
329 views

Advanced Linear Programming

This document provides an overview of the simplex method for solving linear programming problems. It discusses key concepts like extreme points, basic feasible solutions, and the generalized simplex tableau. The simplex method works by starting at an initial basic feasible solution and iteratively moving to adjacent basic feasible solutions until an optimal solution is found. It describes how to determine leaving and entering variables using optimality and feasibility conditions. The revised simplex algorithm is presented in pseudocode form. Examples are provided to illustrate generating a simplex tableau and performing simplex iterations.

Uploaded by

Joshua Nathanael
Copyright
© © All Rights Reserved
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
You are on page 1/ 26

Advanced Linear Programming

(Part 1)
TUA 405
Operations Research 1
Simplex Method Fundamentals
 In linear programming, the feasible solution space is said to form a
convex set if the line segment joining any two distinct feasible points
also falls in the set.
 An extreme point of the convex set is a feasible point that cannot lie
on a line segment joining any two distinct feasible points in the set.
 A feasible point x can be expressed as a convex combination of its
extreme points Xl, X2, X3, X4, X5, and X6 using :
X = α1X1 + α2X2 + α3X3 + α4X4 + α5X5 + α6X6
where
α1 + α2 + α3 + α4 + α5 + α6 = 1
From Extreme Points to Basic Solutions
 It is convenient to express the general LP problem in equation
form using matrix notation.
 Define
 X as an n-vector representing the variables.
 A as an (m x n)-matrix representing the constraint coefficients.
 b as a column vector representing the right-hand side.
 C as an n-vector representing the objective-function coefficients.
 The LP is written as :
Maximize or Minimize z = CX
subject to
AX = b
X≥0
From Extreme Points to Basic Solutions
 In the simplex tableau format, the rightmost m columns of A al-
ways can be made to represent the identity matrix I.
 A basic solution of AX = b is determined by setting n – m variables
equal to zero.
 Solving the resulting m equations in the remaining m unknowns,
provided that the resulting solution is unique.
 Given this definition, the algebraic definition of basic solution can
be written as :
Extreme points of {X|AX=b}<> Basic solution of AX = b
From Extreme Points to Basic Solutions
 We conclude that the basic solutions of AX = b contain all the in-
formation we need to determine the optimum solution of the LP
problem.
 If we impose the nonnegativity restriction, X ≥ 0, the search for
the optimum solution is confined to the feasible basic solution only.
 To formalize the definition of a basic solution, the system AX = b
can be expressed in vector form as follow

 The vector Pj is the jth column of A.


From Extreme Points to Basic Solutions
 If XB is the set of m variables associated with the vectors of non-
singular B, then XB must be a basic solution, then we have :
BXB = b
 Given the inverse B-1 of B, we get the corresponding basic solution
as XB = B-1b
 If B-1b ≥ 0, then XB is feasible.
 The definitions assumes that the remaining n – m variables are
nonbasic at zero level.
 The maximum number of (feasible and infeasible) basic solutions is
given by
Example
 Determine and classify (as feasible and infeasible) all the basic so-
lutions of the following system of equations.

 The following table summarizes the results.


Generalized Simplex Tableau in
Matrix Form
 Consider the LP in equation form:
Maximize z = CX, subject to AX = b, X ≥ 0
 The problem can written equivalently as

 Suppose that B is a feasible basis of the system AX = b, X ≥ 0, and


let X be the corresponding vector of basic variables with CB as its
associated objective vector.
 The corresponding solution may then be computed as follows :
Generalized Simplex Tableau in
Matrix Form
 The general simplex tableau in matrix form can be derived from
the original standard equations as follow :

 Matrix manipulation yield the following equations

 Given Pj is the jth vector of A, the simplex tableau columns associ-


ated with variable xj can be represented as
Example

 Generate the simplex tableau associated with the basis B = (P1,P2)


Example
 Given B = (P1,P2), then XB = (x1,x2)T and CB = (1,4)

 Then we get,

 To compute constraint columns in the body of the tableau,


Example
 Next, we compute the objective row as follow

 Finally, we compute the value of the objective function as follow

 The entire tableau can be summarized as


Revised Simplex
 The general LP problem can be written as follows :

 For a given basic vector XB and its corresponding basis B and object-
ive vector CB, the general simplex tableau developed can be repre-
sented as :

 Which zj – cj the reduced cost is defined as :


Revised Simplex : Optimality
Condition
 An increase in nonbasic xj will improve the value of z relative to its
current value (= CBB-1b) only if its zj - cj is strictly negative in the
case of maximization and strictly positive in the case of minimi-
zation.
 The entering variable is selecting as the one with the most negati-
ve (most positive) zj - cj in case of maximization (minimization).
 In feasibility condition, the determination of the leaving variable is
based on examining the constraint equation associated with the ith
basic variable, which is :
Revised Simplex : Optimality
Condition
 When Pj is selected by the optimality condition to enter the basis,
its associated variable xj will increase above zero level.
 At the same time, all the remaining nonbasic variables remain at
zero level. Thus, the ith constraint equation reduces to

 The equation shows that if (B-1Pj)i > 0, an increase in xj can cause


(XB)i to become negative, which violates the nonnegativity
condition, (XB)i ≥ 0 for all i, and we have
Revised Simplex : Optimality
Condition
 The conditions yields the maximum value of the entering variable
xj as :
Revised Simplex : The Algorithm
 Step 0
Construct a starting basic feasible solution and let Band CB be its
associated basis and objective coefficients vector, respectively.
 Step 1
Compute the inverse B-1 by using an appropriate inversion method.
 Step 2
For each nonbasic variable xj, compute

If zj - cj ≥ 0 in maximization (≤0 in minimization) for all nonbasic


xj, stop; the optimal solution is given by
Revised Simplex : The Algorithm
Else, apply the optimality condition and determine the entering va-
riable xj as the nonbasic variable with the most negative (positive) zj
- cj in case of maximization (minimization).
 Step 3
Compute B-1Pj. If all the elements of B-1Pj are negative or zero,
stop; the problem has no bounded solution. Else, compute B-1b.
Then for all the strictly positive elements of B-1Pj, determine the ra-
tios defined by the feasibility condition. The basic variable xi asso-
ciated with the smallest ratio is the leaving variable.
 Step 4
From the current basis B, form a new basis by replacing the lea-
ving vector Pi with the entering vector Pj. Go to step 1 to start a
new iteration.
Example
Example
Example
Example
Example
Example
Example
End of Topic
TUA 405
Operations Research 1

You might also like