ESI 6448 Discrete Optimization Theory - PowerPoint PPT Presentation

1 / 21
About This Presentation
Title:

ESI 6448 Discrete Optimization Theory

Description:

anticycling rules for simplex. Lexicographical pivoting rule. lexicography ordering ... (b) Use the lexicographic version of the dual simplex. algorithm. ... – PowerPoint PPT presentation

Number of Views:110
Avg rating:3.0/5.0
Slides: 22
Provided by: Min26
Category:

less

Transcript and Presenter's Notes

Title: ESI 6448 Discrete Optimization Theory


1
ESI 6448Discrete Optimization Theory
  • Lecture 20

2
Last class
  • Fractional dual algorithm for IP
  • Using Gomory cuts
  • Finite of steps
  • Lexicography
  • More inequalities
  • Disjunctive inequalities

3
Fractional dual algorithm for IP
  • Initialization.
  • Solve LP and get optimal sol x
  • Iteration. (reoptimization)?
  • Choose a source row i (ai0 lt 0).
  • Add the Gomory cut using the row i and a
    corresponding basic variable s.
  • Apply the dual simplex algorithm.
  • If dual is unbounded, stop w/ empty sol.
  • Let x be the new optimal sol from the
    reoptimization.
  • If x is integral, stop w/ optimal integer sol
    x.

4
Example

x2
3
2
1
x1
3
2
1
0
5
Finiteness of fractional dual algorithm
  • To prove the finiteness,
  • no cycle should be generated
  • anticycling rules for simplex
  • Lexicographical pivoting rule
  • lexicography ordering
  • Blands rule
  • subscript ordering

6
Lexicography
  • lex-min choice will be unique

7
Anticycling
  • Lexicographic pivoting rule
  • Entering column any j s.t. cj is negative
  • Exiting column for each i s.t. ui gt 0, divide
    i-th row by ui and choose the lex-min row l.
  • Blands rule (smallest subscript pivoting rule)?
  • Entering column smallest j s.t. cj is negative.
  • Exiting column out of all ties, choose the
    smallest l.
  • first row is selected.

8
Lexicography for primal simplex

9
Blands rule for primal simplex

10
Finiteness of fractional dual algorithm
  • Make the following choices in the fractional dual
    algorithm for IP (a) Choose the source row to
    be the first row with a noninteger ai0.
    (b) Use the lexicographic version of the dual
    simplex algorithm.Then, assuming that
    the original problem has an upper bound on the
    cost z of a feasible solution, the algorithm
    terminates with an integer solution in a finite
    number of steps or finds that there is no
    feasible integer solution to the original problem.

11
Disjunctive constraints

12
Disjunctive inequalities

-x13x2 ? 7
(5, 4)?
(10, 4)?
P 2
(4, 2)?
P 1
(0, 1)?
(5, 0)?
13
Disjunctive inequalities

14
Disjunctive procedure
  • Disjunctive procedure (D-inequalities)?

15
D-inequalities

16
Basic Mixed Integer Inequalities

17
Mixed integer rounding inequalities

18
MIR procedure
  • MIR procedure (MIR inequalities)?

19
Gomory Mixed Integer Cut

20
Example

(20/7, 3)?
(2, 1)?
(2, 1/2)?
21
Today
  • Finiteness of fractional dual algorithm for IP
  • Finite of steps using lexicography
  • More inequalities
  • D-inequalities
  • MIR inequalities
Write a Comment
User Comments (0)
About PowerShow.com