The Research Of Multi-Objective Decision-Making Based On Agent

The decision-making of enterprise is the core of enterprise management. Facing extremely fierce competition in the market. How to make forward-looking strategies and make correct decisions are the key to the success of enterprise.Since the 1970s, Decision Support System (DSS) has been applied to many enterprises, and has achieved a certain degree of development. But it still can't satisfy enterprises'need. Two reasons: on the one hand, DSS is powerful,but it is not highly intelligent;on the other hand , under the contemporary market economy, with the development of social economy and advancement of technology, all enterprises will face an increasingly complicated and changeful environment. Decision-making problem of enterprise becomes much more complicated. The decision-making problem changes from single object to multi-hierarchy multi-attribute and multi-object. It involves multifold professional acknowledge and has a mass of qualitative, unstructure information which will depend on multi-objective arithmetics to deal with. These factors bring a certain difficulty to decision- making system.In view of the difficulties, at first,this paper studies the structure of DSS, and introduces agent technology into DSS , Author put forwards an architecture of multi-objective decision-making system based on agent and gives a detailed design to some key agents. Author focuses on the study of model composite and automatic choice of model in this paper. Model composite is divided into internal composite within the model management agent and the composite among the model management agents, Author put forwards a composite arithmetic. Then, on the basis of the theory of multi-objective ,Author emphatically discusses of two multi-objective decision-making arithmetics ,They are AHP and TOPSIS. Further more, Author designs a TOPSIS whose weight is determined by AHP. Finally, This paper constructs a multi-objective decision-making system prototype based on agent in the application field of new product development".The system consists of database model base method base and knowledge base. Author uses relational database SQL Server 2000 to manage information,and ADO technology to access data. The system consists of seven agents, which are implemented by the language of Visual C + + 6.0 . As can be seen from the running process,the intelligence of the system has improved significantly.