Josu Doncel Vicente est maitre de conférences à l’Université du pays Basque dans le domaine de l’évaluation des performances et de l’optimisation dans les réseaux. Il a effectué sa thèse de doctorat au laboratoire LAAS de Toulouse, où il a utilisé la théorie des jeux et la théorie des files d’attente afin de résoudre des problèmes de performance et d’optimisation dans les réseaux (équilibrage de charge, recherche de chemin optimal). Il a ensuite effectué un post-doc à l’Inria Grenoble où il a utilisé la théorie des jeux en champ moyen pour étudier les dynamiques de propagation de virus (ou de diffusion d’information) dans des systèmes qui contiennent un grand nombre d’éléments.
Son programme :
- Séminaire vendredi 1er février 2019 – 10h30 – UVSQ – Bâtiment Descartes – salle 301 – 45 avenue des États-Unis – 78035 Versailles cedex
- Titre: Pricing in a single server with relative priorities
- Résumé:
We consider a game where users share the capacity of a single server. The allocated capacity to a user is directly proportional to its payment. Each user wants to minimize its payment while ensuring a certain quality of service. Due to lack of analytical expressions for the underlying queuing discipline, we are able to give the solution of the game only under some assumptions. For the general case, we develop an approximation based on a heavy-traffic result and we validate the accuracy of the approximation numerically.
Plan accès – une pièce d’identité est demandée à l’entrée
- Séminaire lundi 14 janvier 2019 – 11h – SAMOVAR – Telecom SudParis – 9, Rue Charles Fourier – 91011 Evry
- Titre: Size-Interval Routing to Parallel-Servers : economies of scale, performance balancing and asymptotical analysis
- Résumé:
The question of how to route optimally jobs in parallel-server systems is of great importance since they model a wide range of systems such as data-centers or web-servers. In the size-interval routing each host serves jobs whose service demand is in a designated range. This routing policy is interesting from the practical point of view since it does not require signaling between the routers and servers. First, we consider the economies of scaling up the arrival rate and the number of servers proportionally and we study how the degradation increases with the variability of the incoming job size distribution. Then, we present a size-interval routing that, instead of optimizing the performance of the system, balances the performance of the queues and we study the properties of this new routing policy. Finally, we analyze the performance of this system under an asymptotical regime where the system capacity grows linearly with the system demand to infinity and we present optimality results of this regime.
Plan accès
- Séminaire vendredi 14 décembre 2018 – 10h – Université Versailles Saint Quentin – Bâtiment Descartes – Salle 301 – 45 avenue des États Unis, 78035 Versailles (accès contrôlé, merci de prévoir votre pièce d’identité)
- Titre: On the convergence of Mean Field Games
- Résumé:
Mean-field games model strategic decision making of small interacting individuals in an infinite population. We study the convergence of the equilibria of N-player games to mean-field equilibria. We define classes of strategies over which any equilibrium converges to a mean-field equilibrium when the number of players goes to infinity. We also exhibit equilibria outside this class that do not converge to mean-field equilibria. This implies that mean-field games are not the limit of games with very large populations.
Plan accès
- Séminaire vendredi 30 novembre 2018 – 10h à 10h45 – LRI – Batiment 660 – Salle 1018 – Rue Noetzlin, 91190 Gif-sur-Yvette
- Titre: Size-Interval Routing to Parallel-Servers: economies of scale, performance balancing and asymptotical analysis
- Résumé:
The question of how to route optimally jobs in parallel-server systems is of great importance since they model a wide range of systems such as data-centers or web-servers. In the size-interval routing each host serves jobs whose service demand is in a designated range. This routing policy is interesting from the practical point of view since it does not require signaling between the routers and servers. First, we consider the economies of scaling up the arrival rate and the number of servers proportionally and we study how the degradation increases with the variability of the incoming job size distribution. Then, we present a size-interval routing that, instead of optimizing the performance of the system, balances the performance of the queues and we study the properties of this new routing policy. Finally, we analyze the performance of this system under an asymptotical regime where the system capacity grows linearly with the system demand to infinity and we present optimality results of this regime.
Plan accès campus UPSud – Orsay
Seules les personnes avec un identifiant peuvent lire le rapport de visite.