Models and algorithms for combinatorial optimization problems of resource sharing between agents.
Résumé
In this habilitation report, I present models and algorithms for combinatorial optimization problems of resource sharing. The resources can be time and machines (scheduling problems), machines in datacenters (assignment of tasks), or more generic resources (generalized assignment problem). Therefore, the problems studied are mainly scheduling and assignment problems.
Our aim is to design and study algorithms for resource sharing between independent agents, in order to optimize a social-choice function (social welfare). When the agents can have the choice between different strategies, this falls into the field of algorithmic game theory, and we try to design systems which induce good equilibria. We also see how a trusted third entity can help in obtaining good solutions for the social choice function. When agents have private information, this falls into the mechanism design field, and we are interested in the quality (in term of approximation ratio) of truthful algorithms. When the aim is to compute a solution which are common to all the agents (the agents having preferences on the resources), we use tools from the multi-objective combinatorial optimization field and the computational social choice field.
Origine : Fichiers produits par l'(les) auteur(s)
Licence : CC BY NC SA - Paternité - Pas d'utilisation commerciale - Partage selon les Conditions Initiales
Licence : CC BY NC SA - Paternité - Pas d'utilisation commerciale - Partage selon les Conditions Initiales