A Survey of Trust Models in Agent Applications

Trust is an important requirement for sustained interactions between agents. It is relevant to the research of many fields, such as Psychology, Sociology, Economics and a multitude of fields within Computer Science. This manuscript is intended to provide a comprehensive overview of trust models for those who wish to implement trust in an intelligent agent system, or for those who would like to familiarize themselves with the field. We will attempt to cover trust and its issues from the view points of the aforementioned fields, with regard to implementing trust in agent-based organization modeling. Due to the broad range of issues related to trust, we focus here on issues related to risk analysis, centralized versus decentralized management, situational context of the decision, boundaries of a trust value, the transient nature of trust and the concept of modeling reciprocity. We will then discuss the potential implications of these issues in designing agent-based models of organizations and their information sharing policy. Contact: Barbara Pusey Department of Computer Science and Engineering The Pennsylvania State University University Park, PA 16802 Tel: (256) 585-0144 Email: [email protected]