Monday 31 Oct 2016: Dynamics seminar: Various Applications of Multi-Objective Optimization: Multi-modal transportation, Real Time systems scheduling and Yacht weather routing
Laurent Lemarchand - Brest
Multi objective optimization (MOO) aims at finding, for an optimization problem, a set of solutions that represent trade-offs for antagonist objectives. The designer has then to choose the solution that fits at best his needs. All solutions of an MOO problem can be found with, e. g. an adapted linear programming framework, for small instances, but many meta-heuristics exist for solving approximately large testcases. We detail the use of one of them, Pareto Archive Evolution Strategy (PAES) and show applications we have developed with it for solving problems in various fields.