Scribd
Upload
28 views

Cutting Plane Algorithm

The cutting plane algorithm solves integer linear programs (ILPs) by starting with the linear programming (LP) relaxation and iteratively adding cutting planes (constraints) to separate the current optimal LP solution from integer solutions. At each iteration: 1. Solve the current LP relaxation to find the optimal solution. 2. If the solution is not integer, generate a cutting plane that separates it from the integer solutions. 3. Add the cutting plane to the LP relaxation and resolve. This process repeats until an integer optimal solution is found. The algorithm guarantees finding an optimal integer solution in a finite number of iterations, though the number of iterations is problem dependent.

Uploaded by

Yu Pin Fish
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
28 views

Cutting Plane Algorithm

The cutting plane algorithm solves integer linear programs (ILPs) by starting with the linear programming (LP) relaxation and iteratively adding cutting planes (constraints) to separate the current optimal LP solution from integer solutions. At each iteration: 1. Solve the current LP relaxation to find the optimal solution. 2. If the solution is not integer, generate a cutting plane that separates it from the integer solutions. 3. Add the cutting plane to the LP relaxation and resolve. This process repeats until an integer optimal solution is found. The algorithm guarantees finding an optimal integer solution in a finite number of iterations, though the number of iterations is problem dependent.

Uploaded by

Yu Pin Fish
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 17

CUTTING PLANE

ALGORITHM
BY:
DR. NUR FADHILAH IBRAHIM

Cutting plane algorithm also starts at the continuous


optimum LP solution.
Special constraints (called cuts) are added to the
solution space in a manner that makes an integer
optimum extreme point

Example 9.2-2 (Taha p. 379)


Consider

the following ILP:


Max
subject to

How cuts are used


Continuous
LP:

Max
subject to

The optimum/solution to this LP is

Next,

we add Cut 1, which produces the (continuous) LP


optimum solution .

Then,

we add Cut 2, which, together with Cut 1 and the


original constraints, produces the LP optimum . This
solution is all integer.

The added cuts do not eliminate any of the original


feasible integer points, but must pass through at least
one feasible integer point. These are basic requirements
of any cut.
The number of cuts, though finite, is independent of the
size of the problem.

CUTTING PLANE ALGORITHM


1.
Relax the ILP (ignore the integer condition), so we have an LP problem. Solve the
LP using standard simplex method. You have reach the optimum if all the
variables are integers. Stop the iteration. Else, go to Step 2.
2. To generate a cut, choose row-which is non integer.
If the equation at row- is generate a new cut:
where
is the fraction part of .
is the fraction part of .
is the positive integer slack.
Tips: choose the row where the value is the largest.
*Note : [] = the largest integer (eg: )
3. Add the new generated constraint (cut) in Step 2 into the tableau in Step 1. If the
tableau is infeasible, apply dual simplex method.

Example
Max

subject to

Step 1: Relax the ILP, solve using simplex method.


Basic

Solution
-21

-11

13

The final tableau:


Basic

Solution
0

The optimal solution is non integer.

39

row which is non integer; the only option is row


Choose

Generate new constraints/cut:

The new cut:

Rewrite

the new cut: . Add this new cut to the tableau:


Basic

Solution
0
0

1
1

3
3

0
0

1
1

0
0

0
0

1
1

39
39

Optimal tableau:
Basic

Solution
0
0
1
0
0

Row
is non integer, so choose this row.
0

Generate new constraints/cut:

The new cut:

Rewrite

the new cut: . Add this new cut to the tableau:


Basic

Solutio
n
0
0

0
0

0
0
0

0
0
0

The
optimal tableau:
Basic

Solutio
n
1

11

-1

-4

All the solutions are integers so we can stop.

Exercise:
Min

subject to

By ignoring the integer condition, we have the following


optimal tableau:
Basic

Solution
0
0

0
0

1
1

0
0

The solution is not integer, so generate a new cut.

You might also like