The Research Of Agent-Oriented Multi-Issue Automated Negotiation Based On IGA

With the rapid development of economic and social, the relations between economic entities is becoming increasingly close and complex, and how to effectively protect the rationalization of interests distribution and resolve the conflicts in cooperation and contacts, has become the focus of attention. As the effective means of allocation optimization and conflict mitigation, negotiation has become an indispensable component of economic activity. In order to improve the negotiation efficiency and ensure the fairness and rationality of the negotiations, the auto-negotiation techniques were the focus of many scholars at home and abroad study since it emerged after the mid-90s. Due to the difference of the focus and the theory and technologies application in study, the auto-negotiation techniques study was divided into negotiation rules, decision-making model and the Implementation technology of automatic negotiation, and many other study directions. Nowadays, the most of auto-negotiation techniques study focused on the relatively simple form and was in experimental and theoretical research stage lacking commercial application value, this seriously affected application and promotion process of the auto-negotiation technology.In this thesis, on the basis of the full reference to the results of recent correlative studies, the model of agent-oriented multi-issue automated negotiation based on improved genetic algorithms (IGA) was constructed for the specific characteristics of economic negotiation activities. This thesis also realized the negotiation system with the aid of computer technology, which laid a good foundation for the further application and promotion of auto-negotiation techniques.The main work of this thesis:(1) The logical reasoning and Utility Decision Theory of auto-negotiation were improved.In auto-negotiation technology, the improvement of negotiators'ability in logical reasoning and rational decision-making was the effective way to enhance the negotiation efficiency and ensure the fairness and rationality of the negotiations. This thesis realized the effective integration of the logical reasoning, multi-attribute utility decision theory and Agent technologies; it greatly improved the effect of the negotiations. (2) The genetic algorithm were improved and applied to auto-negotiation. On the basis of the full reference to the results of correlative studies in the improvement and application of genetic algorithms, this thesis improved genetic algorithm with the entropy theory of Claude Elwood Shannon's, and the convergence speed and stability of the algorithm had been improved, which had been by the relevant literature.(3) Agent-oriented multi-issue automated negotiation system was realized This thesis designed the class diagrams, sequence diagrams and collaboration diagrams of auto-negotiation system with JADE technology of Java technology.(4) With the specific issues of business negotiations, the simulation was made in this thesis, and the rationality and practicability of the negotiation were verified.