I want to use column generation technique to solve cutting stock problem. Is there any other python modeling library that supports column generation?
$\begingroup$
$\endgroup$
3
-
2$\begingroup$ remember that when you would like to solve an integer program, you need branch-and-price, not (only) column generation; most frameworks/codes support CG but not the branching -- when you want to solve a cutting stock problem, you can directly use a Python API of any solver and hope for a good integer solution, or you can trry GCG, of course ;) works like a charm on cutting stock $\endgroup$– Marco LübbeckeCommented Aug 26, 2020 at 11:27
-
3$\begingroup$ @MarcoLübbecke you need branch-and-price if you want the optimal solution. If you are willing to sacrifice it (ouch), you can do price-and-branch (column generation followed my a MIP with the columns in the pool). This is what is done in Pulp's examples. $\endgroup$– KuifjeCommented Aug 26, 2020 at 12:31
-
$\begingroup$ yes, this is what I meant by "directly use a Python API", which probably works well in many cases; thanks for the comment $\endgroup$– Marco LübbeckeCommented Aug 27, 2020 at 9:35
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
You can do this with PuLP. A column wise formulation for the cutting stock problem is given in the examples. So you don't have much to do...
However, it would be interesting to see how PuLp compares with Pyomo.
