Linear Programming: Chapter 3

Degeneracy Home Page

Title Page

Robert J. Vanderbei
Contents

February 18, 2003

JJ II

J I

Page 1 of 11

Go Back

Full Screen

Operations Research and Financial Engineering Close

Princeton University
Princeton, NJ 08544 Quit

http://www.princeton.edu/∼rvdb
. Degeneracy
Definitions.

A dictionary is degenerate if one or more “rhs”-value vanishes. Home Page

Example: Title Page

ζ = 6 + w3 + 5x2 + 4w1
Contents
x3 = 1 − 2w3 − 2x2 + 3w1
w2 = 4 + w3 + x2 − 3w1 JJ II
x1 = 3 − 2w3
w4 = 2 + w3 − w1 J I

w5 = 0 − x2 + w1
Page 2 of 11

A pivot is degenerate if the objective function value does not change. Go Back

Full Screen
Examples (based on above dictionary):
1. If x2 enters, then w5 must leave, pivot is degenerate. Close

2. If w1 enters, then w2 must leave, pivot is not degenerate.

Quit
. Cycling
Definition.
A cycle is a sequence of pivots that returns to the dictionary from which
the cycle began. Home Page

Note: Every pivot in a cycle must be degenerate. Why? Title Page

Pivot Rules. Contents

Definition. JJ II
Explicit statement for how one chooses entering and leaving variables
(when a choice exists). J I

Largest-Coefficient Rule. Page 3 of 11

A common pivot rule for entering variable:

Choose the variable with the largest coefficient in the objective Go Back

function. Full Screen

Hope. Close

Some pivot rule, such as the largest coefficient rule, will be proven never
to cycle. Quit
. Hope Fades

Home Page
An example that cycles using the following pivot rules:
• entering variable: largest-coefficient rule. Title Page

• leaving variable: smallest-index rule.

Contents

JJ II

ζ = 10x1 − 57x2 − 9x3 − 24x4 J I

w1 = − 0.5x1 + 5.5x2 + 2.5x3 − 9x4

w2 = − 0.5x1 + 1.5x2 + 0.5x3 − x4 Page 4 of 11

w3 = 1 − x1 .
Go Back

Full Screen

Here’s a demo of cycling... Close

Quit
Home Page

Title Page

Contents

JJ II

J I

Page 5 of 11

Go Back

Full Screen

Close

Quit
Home Page

Title Page

Contents

JJ II

J I

Page 6 of 11

Go Back

Full Screen

Close

Quit
. Perturbation Method
Whenever a vanishing “rhs” appears perturb it.
If there are lots of them, say k, perturb them all.
Make the perturbations at different scales:
Home Page

other data  1  2  · · ·  k > 0.

Title Page

An Example.
Contents

JJ II

J I

Page 7 of 11

Entering variable: x2
Leaving variable: w2 Go Back

Full Screen

Close

Quit
. Perturbation Method—Example Con’t.
Recall current dictionary:

Home Page

Title Page

Contents

Entering variable: x1 JJ II
Leaving variable: w3
J I

Page 8 of 11

Go Back

Full Screen

Close

Quit
. Other Pivot Rules
Smallest Index Rule.
Choose the variable with the smallest index (the x variables are
assumed to be “before” the w variables). Home Page

Note: Also known as Bland’s rule.

Title Page

Random Selection Rule.

Contents

Select at random from the set of possibilities.

JJ II
Greatest Increase Rule.
J I
Pick the entering/leaving pair so as to maximize the increase
of the objective function over all other possibilities. Page 9 of 11

Note: Too much computation.

Go Back

Full Screen

Close

Quit
. Theoretical Results

Cycling Theorem. If the simplex method fails to terminate, then it must Home Page

cycle.
Title Page

Why? Contents

JJ II

J I
Fundamental Theorem of Linear Programming. For an arbitrary linear
program in standard form, the following statements are true: Page 10 of 11

1. If there is no optimal solution, then the problem is either infeasible

Go Back
or unbounded.
2. If a feasible solution exists, then a basic feasible solution exists. Full Screen

3. If an optimal solution exists, then a basic optimal solution exists. Close

Quit
. Geometry

Home Page

maximize x1 + 2x2 + 3x3 maximize x1 + 2x2 + 3x3 Title Page

subject to x1 + 2x3 ≤ 3 subject to x1 + 2x3 ≤ 2
x2 + 2x3 ≤ 2 x2 + 2x3 ≤ 2 Contents
x1 , x2 , x3 ≥ 0. x1 , x2 , x3 ≥ 0.

JJ II
x3
x3

J I
x2+2x3=2

x2
x2=0 1 2
x2+2x3=2 x2 Page 11 of 11
1

x2=0

Go Back

x1+2x3=3

x1+2x3=2 Full Screen

x1
x1

Close

Quit