Research On Agent-based Purchasing Management Decision Support System

Decision Support System (DSS) is an important tool which can make decision for enterprise and enhance the key competitiveness of enterprise, and it has improved the quality and efficiency of decision making by offering various decision information and solution to the commercial problems for the enterprise. Agent is a new developing technology in Distributed Artificial Intelligence (DAI), and it provides a new way to build DSS, because of its intelligence and flexibility, suitable for dealing with complex, synergistic and unpredictable problems.In the foundation of researching on Agent-based DSS, the thesis analysises the classic Contract Net Protocol which is used in Multi-Agent Cooperation, some shortcomings are found, which are the redundant communication and high load of node. It proposes the improved Contract Net Protocol based on auction mechanism, through introducing the mental state parameter of agent, refusing bids strategy and auction mechanism, which reduces costs of communication, balances the total system burden and economizes system resource.The thesis also discusses the method of Vendor Selection, and analysises the advantages and disadvantages of Analytic Hierarchy Process (AHP) and Fuzzy Comprehensive Evaluation (FCE). It proposes combining the AHP and FCE to apply to Vendor Selection, making more reasonable and more scientific, which bring benefit for enterprise.Finally, the thesis analysises the function and requirments of the system, which is purchasing management Decision Support System of Dalian DHI&DCW Group, it proposes the level architecture of Agent-based DSS, describes the inner-structure and function of Agent, designs and implements the prototype system of purchasing management Decision Support System of steel based on Agent.