André Hottung

Reinforcement Learning Guided Neural Deconstruction Search for Flexible Job Scheduling

Davide Zago, André Hottung, Fynn Martin Gilbert, Rossella Cancelliere, Kevin Tierney

Abstract Learning-based approaches have made substantial progress on solving combinatorial optimization problems, increasingly rivaling classical operations research methods. In particular, improvement-based machine learning methods, which iteratively refine an existing solution,…Read more