Multi-Agent And Its Application To Control And Fault Diagnosis Of Production System

Now multi-agent is hotspot in research field, and it belongs to front subjects in the world. Agent's concept arises from artificial intelligence, and it has several traits: Autonomy, Social ability, Reactivity, Pro-activity and Rationality. By means of agent we can realize the flexibility and the re-configurability of the system. Aiming at the actuality that control, monitoring and diagnosis are segregated one another in modern production system, in the dissertation we put forward the following points: take advantage of agent's high autonomy and easy integration in the control, monitoring and diagnosis of production system, take the packing line as the study object, and carry out the application research of agent technology. The main research work of the dissertation is as follows:1. Taking the method of multi-agent, previous diagnosis module, control module and monitoring module are encapsulated as single agents, composing multi-agent system. Then the problem that control, monitoring and diagnosis are segregated one another can be resolved. Making diagnosis, control and monitoring integrated, the aim of high diagnosis efficiency, flexible control and easily message exchange is reached.2. Put forward the point using CORE A as the middleware of data exchange among agents, resolve the problem of smoothlessness in data exchange, and reach the aim that data is exchanged in time among agents.3. Establish the model of distributed diagnosis system (DDS) using multi-agent concept. Using multi-agent concept, the agent model of DDS is established. The composing of thismodel and the function of each agent are analyzed in detail. Conflict resolution is put emphasis on.4. Establish the model of control and monitoring system using multi-agent concept.Using multi-agent concept, the model of control and monitoring system is established. The composing of this model and the function of each agent are analyzed in detail. Put forward the viewpoint that and cooperation is realized in multi-agent system through blackboard model.II5. Aiming at the complexity of production environment, put forward the point: develop a system that combines man and machine, takeman as an agent, and form a system of integrated proseminar.6. Apply multi-agent models to packing line.System of control, diagnosis and monitoring is designed by means of the model mentioned. Validity and rationality are demonstrated.