Negotiation in a non-cooperative environment

The area of automated negotiation has been of particular interest in artificial intelligence due to the important role negotiation plays in facilitating understanding and achieving co-operation among entities with differing interests. These entities may be individuals, organizations, governments, or automated agents. This paper presents methods for solving different aspects of automated negotiation: with whom to negotiate, evaluation of suggestions and the way to offer suggestions. These methods were successfully used to develop the system Diplomat, that may be one of the players in a board game, Diplomacy. This game is characterized by intense negotiation, a very large set of possible strategies and the absence of a trusted intermediary. Although Diplomacy players may break their promises, close co-operation is needed for a success.