mlopt
is a package to learn how to solve numerical optimization problems from data. It relies on cvxpy for modeling and gurobi for solving the problem offline.
mlopt
learns how to solve programs using pytorch (pytorch-lightning), xgboost or optimaltrees. The machine learning hyperparameter optimization is performed using optuna.
Online, mlopt
only requires to predict the strategy and solve a linear system using scikit-umfpack.
To see mlopt
in action, have a look at the notebooks in the examples/ folder.
Coming soon at mlopt.org!
If you use mlopt
for research, please cite the following papers:
-
@Article{bertsimas2021, author = {{Bertsimas}, D. and {Stellato}, B.}, title = {The Voice of Optimization}, journal = {Machine Learning}, year = {2021}, month = {2}, volume = {110}, issue = {2}, pages = {249--277}, }
-
Online Mixed-Integer Optimization in Milliseconds
@article{stellato2019a, author = {{Bertsimas}, D. and {Stellato}, B.}, title = {Online Mixed-Integer Optimization in Milliseconds}, journal = {arXiv e-prints}, year = {2019}, month = jul, adsnote = {Provided by the SAO/NASA Astrophysics Data System}, adsurl = {https://ui.adsabs.harvard.edu/abs/2019arXiv190702206B}, archiveprefix = {arXiv}, eprint = {1907.02206}, keywords = {Mathematics - Optimization and Control}, pdf = {https://arxiv.org/pdf/1907.02206.pdf}, primaryclass = {math.OC}, }
The code to reproduce the results in the papers is available at bstellato/mlopt_benchmarks.