chapter  6
Learning in Constrained Robust Games
ByHamidou Tembine
Pages 24

This chapter is concerned with the concept of constrained equilibrium and satisfactory solution in constrained-like games. In addition to the Nash equilibrium, where no player can improve its payoff by unilateral deviation, which can be seen as players are interested in selfishly maximizing its payoff function, we present the concept of satisfactory solution, which models the case where players only aim to guarantee a minimum satisfaction level. At a satisfactory solution, whenever it exists, each player is able to guarantee its minimum satisfaction level which corresponds to a certain quality of experience (application layer) or a quality of service (network layer) needed by the player. Then, we investigate an efficient satisfactory solution and establish a connection with the constrained equilibrium. Under this setting, we develop a fully distributed algorithm to be close to the set of satisfactory solutions in long-run interactions.