More and more distributed systems include agents that interact in various and varying degrees of collaboration and competition. For example, in telecommunication protocols, inter-domain routing often involves routes that belong to commercial companies whose benefits are function of the involved traffic, and hence function of a global competition. As another example, in the computational grid, tasks must often be scheduled on machines, that may belong to several organizations that have their own objectives. This may lead to globally sub-optimal load balancing situations resulting of the competition between organizations that have to optimize their own objectives. This workshop is devoted to problems related to learning equilibria techniques and the study of the properties of solution points in networks and distributed systems.

  • Algorithmic Game Theory
  • Evolutionary Game Theory and learning techniques
  • Algorithms for repeated or dynamic games
  • Reinforcement learning
  • Distributed gradient descent based algorithms
  • Comparison of centralized versus distributed algorithms
  • Stochastic approximation of dynamics
  • Convergence properties of evolutionary games
  • Evaluations of quality of solutions
  • Notion of stability in distributed System
  • Convergence results for distributed system

Registration is free and includes full access to the technical program, lunches (June 20th and 21st) and the conference dinner (June 20th).

Registrations are now closed.