Combinatorial Optimisation and Machine Learning

James Cussens

There is a growing interest in methods which combine solving combinatorial optimisation (CO) problems and machine learning. In some cases machine learning is used to learn the best strategies for to solving an NP-hard problem. In other cases, CO methods (e.g. integer linear programming (ILP)) are used to solve hard machine learning problems. In yet another combination CO solvers have been included as a layer in a deep learning architecture. The goal of this project is to explore this area further to allow fruitful cross-fertilisation.

Refs: “Bridging the Gap between Machine Learning and Optimization” https://sites.google.com/usc.edu/cpaior-2022/master_class

Bertsimas and Dunn. “Machine Learning under a Modern Optimization Lens”, Dynamic Ideas, 2019

Paulus et al. “CombOptNet: Fit the Right NP-Hard Problem by Learning Integer Programming Constraints” Proc. ICML 2021