Research & Development Of Multi-Agent Based Wastewater Treatment Decision Direction Expert System

Agent Theory is a hot research area, syncretizing and exploiting the conceptions of Object-Oriented technology, Distributed Computation and Artificial Intelligence. Its recognitions bring new ideas to software programming, give ES designation support in technology and thinking, and provide possibility to simulate the intelligent activities of human being on a higher level.Our discussion centers round the research and development of the Multi-Agent Based Urban Wastewater Treatment Decision Direction Expert System. This system will be applied at Gongping Wastewater Treatment Plant, Fengjie, Chongqing. It is a paradigm of economical and convenient environment infrastructure contribution technologies of cities and towns in West China, which is a project sponsored by the government of Holland. This paper firstly discusses several aspects of Agent Theory, such as basic concepts, recognition model, architecture, MAS communication mechanism and knowledge-share framework etc. Then it analyzes the difficulties faced with traditional wastewater treatment expert systems, and points out combining the Agent Theory with traditional expert system technology is a way to resolve these problems. After that, the Agent-Oriented (AO) methodologies are discussed. We choose MaSE to direct our work after the comparison of some prominent AO methodologies. Several technical issues are emerged during the analyzing, designing and implementation phases in order to easily model and implement our system. We present a roadmap concerning how to use agentTool ( a modeling tool of MaSE) for analysis and design purposes and implement agents taking JADE as a platform, as well as how to combine JESS with JADE to create an intelligent agent.After the testing of the prototype system, we give some ways to improve its performance, such as using ANN to enhance the robust and self-learning ability, better management of knowledge base and so on. Considering that the Agent Theory and standards are not very well-rounded, and also our level and conditions are limited, many of the thesis are needed to study further. Thus, our aim here is to share this experience with future MAS developers, give them a new approach that can be used to reference while developing open, integrated, distributed intelligent systems.