Download presentation
Presentation is loading. Please wait.
Published byLeslie Stallman Modified over 9 years ago
1
Branch and Bound See Beale paper
2
Example: Maximize z=x1+x2 x2 x1
3
Solve First LP problem: Solution is [1.5 2.5] x2 x1
4
[1.5 2.5] X1 <= 1 [1 1.5] x2 x1 X1 >= 2 [2 1.5], z=3.5, z=2.5
5
[1.5 2.5] X1 <= 1 [1 1.5] x2 x1 X1 >= 2 [2 1.5], z=3.5, z=2.5 X2 = 2 No solution [2.25, 1], z=3.25
6
[1 1.5] x2 x1 [2 1.5], z=3.5, z=2.5 X2 = 2 No solution [2.25, 1], z=3.25 X1 = 3 [2,1], z=3 No solution
8
Example: Maximize x1+x2 x2 x1
9
S Sbar Sums edges out of S >= 2
10
In TSP, we solve LP problem with constraint {each vertex has 2 edges incident to it} and we add just relevant ‘subtour inequalities’ to cut off any subtour solutions. So each time we solve LP and if we get a subtour solution, we add the specific subtour inequality to cut off that solution and resolve LP. This continues until we get a final tour solution.
11
Objective Function to be minimized Unbounded
12
Infeasible solution
13
LP feasible, but integer infeasible
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.