Math 482 (Lecture 9): Duality III: The dual simplex
method
This lecture concerns Sections 3.5, 3.6 and 3.7 of the textbook.
What we have shown before Test 1 is the following:
1. Given any LP we can put it in standard form
2. Using phase 1 of the 2phase simplex method we can find a basic feasible solution
(column 0 is nonnegative)
3. Row 0 has negative entries, we attempt to drive towards nonnegativity by selecting
columns that are negative and pivoting using the (row) ratio test.
4. When we terminate at a finite optimum both row 0 and column 0 are nonnegative.
Complaint:
Sometimes 2phase is tedious.
Response:
For certain LP's it is better to use the dual simplex method:
**** This is often the case when the cost vector c is nonnegative.
Warning about possible confusion because of the name "dual simplex":
The dual
simplex method secretly uses duality theory to solve an LP, and is often helpful to solve
LP's of the form
min y'b, s.t., y'A≥ c and y≥ 0
which is the usual form of our duals. HOWEVER, it actually solves the LP it is given, not
the dual of that LP.
Example handed out in class.
Summary of dual simplex method:
The rules are "mirror images" of the rules for usual
(primal) simplex:
A.
When to use dual simplex:
The standard form of the problem has nonnegative row 0
The terminology for this is "primal optimal" or "dual feasible" (more later).
However, column 0 is not nonnegative (primal infeasible)
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
This is the end of the preview.
Sign up
to
access the rest of the document.
 Spring '08
 Staff
 Math, Linear Programming, Optimization, Duality, Stratch, dual simplex method

Click to edit the document details