lecture12

# lecture12 - Review Modeling PPP IE426 Optimization Models...

This preview shows pages 1–4. Sign up to view the full content.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document

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.

Unformatted text preview: Review Modeling PPP IE426: Optimization Models and Applications: Lecture 12 Jeff Linderoth Department of Industrial and Systems Engineering Lehigh University October 12, 2006 Jeff Linderoth IE426:Lecture 12 Review Modeling PPP MILP Relaxations Branch and Bound Got MILP? Mixed Integer Linear Program minimize c T x subject to Ax ≥ b x ≥ x j ∈ Z ∀ j ∈ I ⊆ N. Jeff Linderoth IE426:Lecture 12 Review Modeling PPP MILP Relaxations Branch and Bound It’s All About Relaxations Let z S = min f ( x ) : x ∈ S Let z T = min f ( x ) : x ∈ T S T Jeff Linderoth IE426:Lecture 12 Review Modeling PPP MILP Relaxations Branch and Bound Obvious, but Important Stuff z T ≤ z S !!! If x * T is an optimal solution to min f ( x ) : x ∈ T And x * T ∈ S , then x * T is an optimal solution to min f ( x ) : x ∈ S If we were to replace min by max, then z T ≥ z S . Jeff Linderoth IE426:Lecture 12 Review Modeling PPP MILP Relaxations Branch and Bound LP relaxation minimize z IP = c T x subject to Ax ≥ b x ≥ x j ∈ Z j ∈ P ⊆ N. minimize z LP = c T x subject to Ax ≥ b x ≥ z LP ≤ z IP If x * solves the LP relaxation and x * satisfies the integrality requirements, then x * solves IP Jeff Linderoth IE426:Lecture 12 Review Modeling PPP MILP Relaxations Branch and Bound x * ∈ X ? Then we branch! Partition the problem into smaller pieces. x * ∈ X ⇒ ∃ k ∈ P such that x * k is fractional. Create two new problems... 1 In one problem, add the constraint x k ≤ x * k 2 In the other problem, add the constraint x k ≥ x * k Jeff Linderoth IE426:Lecture 12 Review Modeling PPP MILP Relaxations Branch and Bound The Branch and Bound Algorithm All of the following assumes a maximization problem. It works equally well for minimization, but you need to replace “lower” by “upper” (and vice verse) everywhere Let z L be a lower bound on optimal objective value. Originally z L =-∞ . The initial list of LP relaxations to solve consists of just the LP relaxation to the entire problem. Jeff Linderoth IE426:Lecture 12 Review Modeling PPP MILP Relaxations Branch and Bound Branch and Bound (1 of 2) 1 Select an LP Relaxation 2 Solve the selected LP relaxation. If the LP relaxation is infeasible, Go to 1. 3 Otherwise, let x * be the optimal solution to the LP relaxation and let z * be its objective value. 4 If x * satsifies integer restrictions, then z * is the optimal value for the restricted IP problem. If z * > z L , then z L := z * . (Also keep track of x * as candidate solution). Go to 1. Jeff Linderoth IE426:Lecture 12 Review Modeling PPP MILP Relaxations Branch and Bound Branch and Bound (2 of 2) 5 Otherwise, x * does not satisfy the integer restrictions. z * is an upper bound on the optimal value of the restricted IP problem....
View Full Document

## This note was uploaded on 02/29/2008 for the course IE 426 taught by Professor Linderoth during the Spring '08 term at Lehigh University .

### Page1 / 9

lecture12 - Review Modeling PPP IE426 Optimization Models...

This preview shows document pages 1 - 4. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online