Multi-agent Negotiation Model Based on RBF Neural Network Learning Mechanism

Aiming at the problem that negotiation agentpsilas learning algorithm is lack of learning ability for the negotiation history information, this paper introduces RBF neutral network technology in multi-Agent negotiation, establishes a Bilateral-Multi-Issue Negotiation Model, and defines a corresponding negotiation algorithm and utility evaluation functions. Negotiation agents learn to change the belief of the environment and other agents by using RBF neutral network, thus to determine the inference strategy in negotiation. It is proved by the experiment that this method can improve the efficiency of the mutual negotiation markedly.