Trust based cooperative system formation and evolution

Referring to the trust relationship models of social people, the trustworthiness of agents was evaluated by utilizing Bayesian method, and a kind of trust mechanism-based multi-agent cooperation system was presented. By updating the trust-based connection between agents, a robust multi-agent cooperation system was formed. The main properties of the proposed cooperative system were analyzed based on random graph model. Simulations prove that the proposed cooperation model can not only improve the efficiency of cooperation of agents and increase the ratio of successful execution of tasks, but also cluster nodes around the trustworthy nodes which have the strong ability of providing resource and assurance the reliability of cooperative agent selection.