Distributed Games

The Internet exhibits forms of interactions which are not captured by existing models in economics, artificial intelligence and game theory. New models are needed to deal with these multi-agent interactions. In this paper we present a new model--distributed games. In such a model each players controls a number of agents which participate in asynchronous parallel multi-agent interactions (games). The agents jointly and strategically control the level of information monitoring by broadcasting messages. As an application we show that the cooperative outcome of the Prisoners' Dilemma game can be obtained in equilibrium in such a setting.