Research On Multi-agent Based Multi-issue Negotiation Conflict Network Model And It’s Evolution

With the development of the electronic commerce and artificial intelligence, the combination of them becomes a new technology.The electronic commerce negotiation system which based on agent artificial intelligence can satisfy the needs of customer, especially upon its increasing presence. In consequence, this technology promoting the development of intellectualized electronic commerce. However the negotiation conflict is ubiquitous in the intellectualized negotiation system which consisting of multi agent, one agent could forming negotiation conflict with multiple agent on account of the dependency between multi issue, thus the negotiation conflict network is formed between the agent. The negotiation conflict network which is complex and dynamic will bring about unsteady running of the agent negotiation system. Therefore, to carry through the research about the formation and the evolution of negotiation conflict network is practical and significant.Based on negotiation conflict network, this dissertation makes researches around with its formation and the mechanism of its evolution, and aimed at enhancing the efficiency of the negotiation. The primary works and innovations are summarized as follows.First of all, this dissertation depicts negotiation issue and gives the definition of negotiation strategy in the basis of negotiation model. Negotiation issue and negotiation strategy are the basic elements of negotiation conflict, then the dissertation puts forward the concepts of negotiation conflict, as well as demonstrates the process of negotiation conflict through agents negotiating. According to the characteristics of multi issue negotiation which the agent presents in intellectualized negotiation system, the dissertation puts forward the concepts of negotiation conflict network and constructs a negotiation conflict network model. In addition, the dissertation takes the multilateral negotiation in electronic commerce as example, demonstrates the sample of negotiation conflict network, and then analyzes the components of negotiation conflict network.Second, since the negotiation conflict set is the pivotal issue in negotiation conflict network, the dissertation applies artificial immune algorithm to solve these sets which rely on the strength of agent’s intelligent action. After defines the solution of negotiation conflict, the dissertation contrives a negotiation conflict solving model by the theory of artificial immune algorithm, which has validity to solve the negotiation conflict set and the solution could be the initialization conditions for agent to making decision.Finally, the dissertation uses the finite state automata theory to analyze the dynamic decision making process of agent, and then contrives a negotiation conflict network evolution model. The agent makes decision by learning neighbor vertices’ negotiation strategies, which affects the dynamic evolution of the network. According to the agent’s strategy, the dissertation analyzes the state transfer between the sets of negotiation conflict and devises the update mechanism about negotiation strategy. These works could reflect the dynamic characteristics of negotiation conflict network, provide effective solution to control the evolution path of negotiation conflict, improve the stability of the intellectualized negotiation system and enhance the efficiency of agent negotiation.