Prisoner’s dilemma game model for e-commerce

Abstract

This study investigates how the user’s exchanges affect people’s trust in electronic commerce and empower them to interact with foreign customers. Online exchanges often occur between strangers who cannot rely on past behavior or the prospect of future interactions to establish mutual trust. Game theorists have formalized this problem as a prisoner’s dilemma and predict mutual noncooperation. In this paper, we introduce a social network based model to promote the cooperation in the prisoner’s dilemma game. For this study, we implemented an agent-based simulation framework, which models different types of behaviors in online auctions. Agents who represent buyers or sellers are located in the nodes of a small world graph. Each link weight between the agent and its neighbor symbolizes how much it trusts this neighbor. A link with trust inferior to some cut-link threshold will be removed from the graph. Our results show that the evolved structure of the graph induces considerable variation in the level of cooperation and the profit in the e-commerce system. This shows that the outcome of our co-evolutionary game, in terms of cooperative behavior, strongly depends on topology but also on the update rule used in the trust between users. Co-evolution of game dynamics in interconnected networks is also studied, where two networks are interconnected, and players have interactions not only with others in the same network, but also with players in the other network. We will demonstrate that the interdependence between networks promotes the cooperation in both networks. But, the degree of promotion changes as a function of interdependence degree.