Multiagent Based Manufacturing Grid Resource Management And Scheduling

In the context of economic globalization, Virtual Enterprise will become the main operation model of manufacturing enterprises in the future. In the field of information technology, the Grid technology is committed to provide the information infrastructure for supporting the Virtual Enterprise. It has become a trend to use Grid and its related technology to construct the Manufacturing Grid in manufacturing enterprise for networked manufacturing. Using Grid and its related technology to construct the Manufacturing Grid is a trend of manufacturing enterprise to meet requirements for networked manufacturing.Resource management and scheduling are one of the core technology of the Manufacturing Grid. Based on the analyzing the requirements of the Manufacturing Grid resource management and scheduling, a Multiagent model of the resource management and scheduling for the Manufacturing Grid is proposed in this paper. Three kinds of Agents are contained in the model. The first agent is the Resource Agent which is used to be on behalf of one manufacturing resource. The second is the Broker Agent which is used to be on behalf of one user. The third is the DF(Directory Federation) which is to present information service. The structure, the specific function and the interaction process of all kinds of Agent in the proposed model are discussed in detail. The price(P), the time(T), the quality(Q) and the reliability (R) are the evaluation criterion of resources in the Manufacturing Grid. Customers have multi-QoS requirements of PTQR for resources selection and need to search for the optimal combination of resources to complete tasks. In order to satisfy the multi-QoS requirements of customers, a genetic algorithm is designed as the optimal selection algorithm for the Broker Agent.The running mode and the Communication Principle of JADE(Java Agent Development Framework) are introduced in the paper. The installation process and the configuration method are illustrated. Based on the JADE and the JGAP(Java Genetic Algorithms Package), a simulation example for the Manufacturing Grid scheduling model is designed and an experiment is conducted. Simulated results show that the model and the algorithm which have been proposed are feasibility and effective. At last, conclusions are sumurized and the future works are proposed.