Advanced Game Theoretical Models and Applications in Networks
University of Avignon
Abstract: New communication networks are more and more complex (heterogeneous, dynamic, etc) and mainly more decentralized. Performances of those networks are difficult to determine, even more when the network is composed of non-cooperative agents. Examples of such networks are: new wireless networks like « self-organized networks », social networks, transportation networks, etc. In the last ten years, Game Theory has become one of the more suitable mathematical tools to study and to optimize those decentralized complex networks. In this presentation, I will expose some works related to advanced game theoretical models and applications to decentralized resource allocation problems in networks. In fact, some particular networks assume strong properties about the interactions between the agents and on the network characteristics. Hierarchical games assume a multi-level structure between the decision makers. Important results can be obtained using this type of non-cooperative game for new wireless networks like Cognitive Radio Networks, based on opportunistic spectrum usage. Evolutionary games, inspired by Darwin's evolution principle, are able to propose models for complex networks with large number of players. Particularly, those models take into consideration robustness of the equilibrium against deviations of a fraction of individuals and also can serve as basement to build learning procedures that converge to equilibrium. Finally, routing games, initiated in transportation networks, propose facilities to model management problems in networks dealing with content distribution and to understand how pricing mechanisms can be used to control/to influence usage in different types of networks (social networks, transportation, content, etc).
About the Speaker: Yezekael Hayel is an Assistant Professor with tenure position at University of Avignon in France. He received one M.Sc. in Computer Science and one in Applied Mathematics from the University of Rennes 1, in 2001 and 2002 respectively. He had a Ph.D. in Computer Science from University of Rennes 1 and INRIA in 2005 on Network pricing. His research interests include performance evaluation of networks based on game theoretic and queueing models. He looks at applications in communication and transportation networks like: wireless flexible networks, bio-inspired and self-organizing networks, economics models of the internet and yield management. Recently, his application domains enlarge to transportation networks, social networks and complex networks. Since he has joined the networking group of the LIA/CERI, Yezekael Hayel participates in several national (ANR) and international projects (European, cefipra, etc) with industrial companies like Orange Labs, Alcatel-Lucent, IBM and academic partners like Supelec, CNRS, UCLA. For the academic year 2014-2015, he is a visiting professor at NYU School of Enginerring.