École Nationale Supérieure de
Techniques Avancées

Abstract

Real-world problems are often about optimizing simultaneously several criteria. You can solve this problem as single-objective problems, but you will have to make choices before-hand about which kind of compromises you want. Multi-objective optimization algorithms aim at obtaining information on the whole range of best- compromises, the Pareto set and the Pareto front. SOFOMORE framework is a recent framework allowing to create multi-objective algorithms from single-objective algo- rithms, allowing to exploit the full diversity of single-objective algorithms. In this master thesis, we had two goals. The first was to obtain convergence results to under- stand better its behavior theoretically . The other was to design an extension of this framework.