Research On Production Scheduling System With Bottleneck Based On Multi-agent

With the increasingly fierce market competition,production plan for the upper management of enterprises is more and more largely affected by the market, and higher sensitivity of the time. In the face of customer requirements on delivery time, more products which need to be modified and the continuous adjustment of orders, the stipulatings for enterprises'plan are dependent on the market and the actual implementation of the job status, it can not be completely dependent on materials and storage to control production. In the context, Production Scheduling System (PSS) needs the following characteristics: it allows enterprises to not only deal with what has happened, but also expect to deal with what is going to happen, and dynamically deal with what is happening. This dissertation focuses on Production Scheduling System and PSS is established based on the multi-agent technologies; an extensive study on the system's crucial technologies is carried out as well. The main contents and achievements of the dissertation are as follows:1. The architecture of Production Scheduling Systems is constructed based on multi-agent. There are three levels that are composed kinds of agents in the system. The main agents are introduced, and the unified agent architecture is structured. The architecture of system is presented. The multi-agent based architecture is satisfied with manufacturing enterprises that have kinds of requirements.2. Aimed at the resources allocation problem, this dissertation uses the degree of membership function for the fuzzy set as the attribute of the rough set, obtains the classified knowledge of the system through the fuzzy truncation set; uses the sample fuzzy degree of membership to constructure fuzzy quality synthetic evaluation matrix, the importance degree of attribute to constructure weight matrix, and the fuzzy assessment method to carry on the appraisal to the resources collections; finally obtains the best resources. This dissertation unifies the qualitative analysis and the quantitative evaluation, realizes the stationary index well quantitative, demonstrates the relative importance of the evaluation criteria, and enables the manufacture resources evaluation result to have the strong objectivity.3. Aimed at the imbalance problem of resource capacity in Production Scheduling System, this dissertation finds out the bottleneck resource in the production line, analyses the inherent mechanism of bottleneck, and describes the production scheduling process based on bottleneck resource, as follows: after Bottleneck Resource Agent(BRA) finds out the bottleneck resource, the remain resources are considered as Non-Bottleneck Resource Agents(NBRA) , and at the assumption that processing capacities of these non-bottleneck resources are infinite, Bottleneck Decomposition Agent(BDA) harmonizes the relationship of jobs'arrival time and transfer time in BRA and NBRAs. After the jobs in BRA are scheduled with the production scheduling decomposition algorithm based on bottleneck resource, the jobs in NBRAs are scheduled with Heuristic Rules, so the problem is simplified as the single machine scheduling of every resource. If there is the urgency job inserted in the process, BRA identifies the bottleneck newly, and then, BDA harmonizes the relationship of jobs'arrival time and transfer time in BRA and NBRAs to realize the final scheduling. The simulation result demonstrates that it is effective to solve the production scheduling problem with higher bottleneck degree.4. This dissertation discusses the design and implementation of Production Scheduling System. Its framework is integrated on the platform of Jini technology. On basis of theories presented in this dissertation, the mold system is actualized primarily.