Incremental PW linear to MIP transformation #3287
Conversation
…ify variables mode does not work, gives infeasible models
…hoice, but not J1 implementation yet.
… transformation. However, gurobi is not magical enough to see the structure here so this basically amounts to asking for a Hamiltonian path in a big graph, and then some. I think I will have to do the 70s stuff after all. Some optimizations would reduce this to "only" asking for a Hamiltonian path on a sparser graph but that is still terrible.
… mild assumptions
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #3287 +/- ##
==========================================
+ Coverage 88.59% 88.64% +0.04%
==========================================
Files 874 877 +3
Lines 99124 99579 +455
==========================================
+ Hits 87822 88267 +445
- Misses 11302 11312 +10
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. |
…n its file, and wrapping it in a function so that it doesn't get built when pyomo.environ is imported
There was a problem hiding this comment.
@sadavis1, I put in a test of the 3D Hamiltonian paths, but I'm not quite sure how to check that they are the right paths given the keys, so would appreciate input if you get a chance. Otherwise this looks good to go. (We've agreed to refactor the simplices and triangulation arguments for PiecewiseLinearFunctions in a follow-on PR (not targeting the upcoming release)).
|
@jsiirola this is ready for you to take a look! :) |
| # starting permutation, such that a fixed target symbol is either the image | ||
| # rho(1), or it is rho(n), depending on whether first or last is requested, | ||
| # where rho is the final permutation. | ||
| def _get_Gn_hamiltonian(n, start_permutation, target_symbol, last, _cache={}): |
There was a problem hiding this comment.
This is dangerous (although maybe you intend it?): all calls to _get_Gn_hamiltonian that do not provide _cache share the same dict object.
There was a problem hiding this comment.
I'm pretty sure this is intended. But maybe @sadavis1 can confirm at some point... (And perhaps we should add a comment).
Summary/Motivation:
Add another transformation, the incremental MIP formulation, to our slowly growing library of things to do with piecewise-linear functions. This one requires a specially ordered triangulation, so some infrastructure around triangulations is also needed.
Changes proposed in this PR:
Legal Acknowledgement
By contributing to this software project, I have read the contribution guide and agree to the following terms and conditions for my contribution: