Department of Computer Engineering, Dalian University of Technology Dalian, Liaoning.
This paper proposes a multi-agent scheduling system for dynamic Job-Shop problem. Firstly, a hybrid multiagent scheduling framework is introduced according to the characteristics of dynamic Job-Shop problem. Secondly the agents’ regulations are given to restrict the behaviors of agents. Thirdly intelligent algorithms and adaptive scheduling algorithm for resource agents are designed to achieve local objections. And multi-objective function considering tardiness, waiting time and cost is designed for task agents as an evaluation criteria of resources’ bids. Then the negotiation mechanism between agents controlling resources and tasks is discussed. In the end of this paper, the proposed framework is realized in java language and the communication between agents is based on RMI. A set of computer simulation experiments are conducted. The results obtained show the effective and good performance of this system in total flow time, total waiting time and makespan.
