Research On Multi-attributes Negotiation Model Based On Preference And Time Constraints

Along with the rapid development of economic globalization and e-commerce, the traditional commercial negotiation, which has high cost and low efficiency, can’t adapt to the current negotiation needs of rapid-pace and high-efficiency. As the complexity of the negotiation environment and the diverse needs of people, people doesn’t negotiate for only one attribute, but for many attributes, such as price, warranty time and delivery date and so on. Agents hold limited information resources. Each agent’s preference information is private, not being understood by other agents. Considering the three factors of multi-attributes goods, the private preference information and the limit mind of human, it is the focus of negotiation study field that how to give the agents learning ability to adapt to the changes of dynamic negotiation environment and to make reasonable concession to avoid deadlock in the dynamic negotiation environment which cooperation and competition are both existed in.Making the above issues as the starting point, this paper built a multi-attributes negotiation cognitive model based on restrictions of preference and time. It expressed the cognitive mind state of the negotiation agent about himself and his negotiation opponent. It showed the decision-making mechanism of negotiation agent. It provided the data, information and method for the basis and the decision support. It guided the negotiation agent to adopt the best behavior in negotiation interactive process.This paper built a multi-attributes negotiation model based on artificial immune algorithm. Because the negotiation agents pursue their respective interest, the negotiation is sometimes at a stand. To solve this question, the model used the artificial immune algorithm and multi-attribute utility theory, considered the negotiation goal as antigen and considered the candidate proposal as antibody. By the immune process of antigen and antibody, the model enabled the Agents to achieve a negotiation solution which maximizes the overall interests. The simulation experiment proved that this model can solve the impasse during the process of negotiation and improve the efficiency and adaptability of negotiation.This paper built a multi-attributes negotiation model based on restrictions of preference and time. If the above negotiation solution cannot meet agents’ interest demands, the agents will communicate with each other to reach an agreement. This paper combined artificial immune algorithm with Bayesian learning algorithm to learn the preference information of the opponent agent and proposed a concession strategy based on the constraints of preference and time. The model makes agents take suitable concession strategy in the dynamic negotiation environment, ensuring that each interaction is effective in the predicting bargaining range. The model can reduce the negative dialogue in interaction and improve the efficiency and adaptability of the negotiation.Finally, based on the above study, we developed a prototype to verify the effectiveness and the practicality of the model.