Local Search with SubProblem Exact Resolution: Application to a real Resource-Constrained Project Scheduling Problem

Nowadays, agility in project management is an essential need for R&D departments. The well-known in literature "Resource-constrained project scheduling problem" is a very tough problem, and it's actually one of the most difficult classical problem. Its resolution with the use of an exact method is an issue for big problems (60 activities and 4 resources). The work of this thesis aims applying an existing heuristic procedure to a real problem, faced during the internship. The algorithm in question alternates intensification and diversification search, combining heuristics to an exact (and external) procedure. The algorithm, written in Python, is widely inspired by the work of Artigues, Palpant and Michelon, that designed "LSSPER" algorithm in 2004. With respect to their work, in this thesis, some procedures have been introduced, changed, or removed, in order to make the algorithm better suited to the problem in question.