Add functions to find optimal partitions using ILP solvers. #3
Conversation
| such as Karmarkar-Karp and Greedy. | ||
| """ | ||
|
|
||
| from numberpartitioning import karmarkar_karp, greedy, complete_greedy |
There was a problem hiding this comment.
Could be useful if the examples here were included as docstrings on the relevant functions or modules. This way, they would be easy to include in auto-generated documentation, and it would be possible to run them as tests locally and in CI.
E.g. something like the example sections in the methods in https://github.com/scipy/scipy/blob/master/scipy/sparse/csgraph/_flow.pyx works well.
|
|
||
| import cvxpy | ||
|
|
||
| DEFAULT_SOLVERS=[cvxpy.XPRESS, cvxpy.OSQP, cvxpy.SCS] |
There was a problem hiding this comment.
Won't this require that the relevant binaries are available on the user's machine? And at least one of them requires a license for large problem instances.
Allowing to use these solvers is fine, but it would then be nice if all relevant dependencies were available through pypi or conda, so it's possible to use the library in a self-contained fashion.
Those would then be specifies as optional dependencies for the Python package.
| except cvxpy.SolverError as err: | ||
| logger.info("Solver %s fails: %s", solver, err) | ||
| if not is_solved: | ||
| problem.solve(solver=solvers[-1]) # If the first n-1 fail, try the last one. |
There was a problem hiding this comment.
Looks like it should be possible to treat the final solver alongside the other ones. Any reason it's handled specifically?
|
|
||
| Returns | ||
| ------- | ||
| A partition representing by a ``PartitioningResult``. |
There was a problem hiding this comment.
So this always runs to optimality right? Given the NP-hard nature of the problem, that might be a bit of an ask. Would it be a lot of work to use callbacks to generate solutions along the way, or maybe make it easy to allow it to return non-optimak results (through time constraints, allowed gap, etc.)?
There was a problem hiding this comment.
The code uses an existing ILP solver, so - as far as I know - it cannot return solutions along the way. On the positive side, ILP solvers use many tricks that make the solution very efficient in practice.
Also added a "demo" folder with some simple demo programs.