Contributions à la théorie de la décision algorithmique : optimisation multi-objectifs et optimisation dans l'incertain
Olivier Spanjaard · HAL (Le Centre pour la Communication Scientifique Directe) · 2011
The works presented here concern the study of decision problems in terms of algorithms.Most works in decision theory until recently focusing on the characterization of preferred solutions,the computational aspect (i.e., the effective computability of solutions) is often relegatedto the background. It is precisely this second aspect that is central to the research topics discussedhere. They are therefore at the crossroads of several areas : algorithms, multicriteriadecision making, decision under uncertainty, combinatorial optimization, artificial intelligence.The presentation is organized into four sections. The first section is an introduction to theissues discussed in algorithmic decision theory. Various decision models for multiple criteriadecision making and decision making under uncertainty are presented in the second section.The next two sections deal respectively with multi-objective combinatorial optimization andoptimization under uncertainty, this latter denomination covering robust discrete optimizationas well as the optimization of non-consequentialist decision criteria in sequential decision problemsunder uncertainty. The brief is concluded by a presentation of future research directions.