The Research Of Job Shop Scheduling System Based On Multi-Agent

Production scheduling system is the core of Enterprise Resource Planning (ERP), also the bottleneck of ERP. In the situation of agile and global manufacturing, many new problem of production scheduling research develops. This research direction comes from a cooperate project named Resource Management System based on Distributed Object Technology with SUN company. The paper is the extended use in a discrate part produce company. A Multi-Agent Scheduling System (MASS) is presented in this paper to solve the agile scheduling problem.After deeply researching of the current research achievement, this paper built multi-agent model consisted of TMA, RMA, TA and RA. A hybrid type of Agent is built to make agent individual can think and active. The hybrid architecture of multi-agent integrated the hierarchy and the heterarchy is also proposed after analyzing. There are horizontal and vertical negotiations in the architecture of MASS in order to overcome drawbacks of hierarchy and heterarchy, and to make the MASS agile and flexible responses capabilities and a feasible, robust and near optimized schedulers.At last a multi-agent scheduling system is built up using distributed object. In the support of RMI technology, the communication of agents is implemented. We also did some experiments about agile production environment. The simulation results have successfully shown that the approach is capable of generating feasible schedulers in an agile manufacturing environment. The experiments have also demonstrated that it has incorporated the features of adaptive and agility to changes in production diversity and environment configuration.