Résumé:
Machine learning approach for single machine scheduling problems This theme aim to design a
machine-learning algorithm to tackle a NP-hard single machine scheduling problems with big size.
This approach consists on two parts: the firs par “learning” aims to learn the system by supplying
the system with a significant number of small size instances solved by an exact method as dynamic
programming. The second step consist to design a machine learning approach to tackle some big
size instances, as we go along, even these instances supply our system to improve its efficiently. A
comparison with chosen metaheuristic is needed to justify the contribution.