Evaluating Practical Automated Negotiation Based on Spatial Evolutionary Game Theory

Abstract

Over the past decade automated negotiation has developed into a subject of central interest in distributed artificial intelligence. For a great part this is because of its broad application potential in different areas such as economics, e-commerce, the political and social sciences. The complexity of practical automated negotiation – a multi-issue, incomplete-information and continuous-time environment – poses severe challenges, and in recent years many negotiation strategies have been proposed in response to this challenge. Traditionally, the performance of such strategies is evaluated in game-theoretic settings in which each agent “globally” interacts (negotiates) with all other participating agents. This traditional evaluation, however, is not suited for negotiation settings that are primarily characterized by “local” interactions among the participating agents, that is, settings in which each of possibly many participating agents negotiates only with its local neighbors rather than all other agents. This paper presents an approach to handle this type of local setting. Starting out from the traditional global perspective, the negotiations are also analyzed in a new fashion that negotiation locality (hence spatial information about the agents) is taken into consideration. It is shown how both empirical and spatial evolutionary game theory can be used to interpret bilateral negotiation results among state of the art negotiating agents in these different scenarios.