A Trust-Based Model for Multi-agent Systems

Trust, Reputation, Multi-agent system

Trust is very important aspect in our everyday interaction with people, groups and institutions in our society. We have trust in our environment, people and institutions as well. We are often rated and judged on the basis of our trustworthiness and this defines a different manner of the interactions in our social life. We behave more openly towards subjects on account of the strong confidence and this subjects can access different types of information which can be confidential. In the case of abuse of the information, the trust of the subject rapidly decrease and it is very hard to restore it. Some research in this area are aimed to use these trust principles from real-word and shift it into digital environment nowadays. In autonomous multi-agent systems, where agents are operated in a networked environment, it is particular possible to use these principles of trust to establish protocols for agents interaction. If we try to shift this real-word trust principles to multi-agent systems, we meet with some essential problems.   This paper try to show a possible approaches to the basic problems with multi-agent systems based on trust. It presents the way to simulate the trust or the reputation from a  viewpoint of application and safety in multi-agent systems.