A framework for comparison of trust models for multi agent systems

Agents technology plays an important role in the development of many major service applications. However, balancing between the flexible features agents provide, and their vulnerability to many security oriented attacks are considered a great challenge. In this paper we review trust models that are proposed in the literature to provide trustworthiness and security to Multi Agent Systems (MAS). We subsequently develop a framework for comparison of the various different trust models. Trust models are first compared and classified according to types of evaluations used, weight assignment, consideration of inaccurate evaluations and architecture. They are also compared according to suitability to MAS.