ApeROO Research Project

Machine learning and Operations Research for scheduling problems

Group : Dynamic Systems and Optimization

Labelling: none

Duration: 3 years (2019 - 2022)

Funding:  RFI Atlanstic2020, CIRRELT and IVADO

Staff involved from LARIS: Christelle Jussien-Guéret

Project partners: Martin Cousineau (Professor at IVADO-HEC Montréal, CIRRELT), Vincent Barichard (LERIA) 

Abstract:

Scheduling problems have been of interest to many researchers in the field of Operations Research (OR) for several decades. The work done in this field consists of approximate methods as well as exact methods. Very often these methods are dedicated to specific problems, without concern for generality, which implies redeveloping a new method for each new problem. On the other hand, the results obtained are not always up to their implementation complexity. 

In recent years, researchers have started to look at machine learning to enrich their approaches. The idea is to try to "learn" before or during solving, to take advantage of the explorations already made and to accelerate the search for solutions.

The objective of this project is to develop the most generic exact solving methods enriched by machine learning techniques in order to efficiently solve several classes of scheduling problems (shop problems in particular).