SMART AGENT-BASED PRIORITY DISPATCHING RULES FOR JOB SHOP SCHEDULING IN A FURNITURE MANUFACTURING WORKSHOP

To meet the increasingly diversified demand, multi-variety and small-batch production has been widely acknowledged in manufacturing industry, which raises higher requirements for production scheduling. Priority dispatching rules have been widely used in job shop scheduling with many advantages, including low computational cost, high flexibility and good portability. However; it is usually hard to guarantee a globally optimal solution. Multi-Agent-based priority dispatching rules for scheduling are available to handle real complex and dynamic production environment, in the meanwhile hold a comprehensively satisfactory performance. In this paper, multi-agents are designed representing production jobs, production machines, and especially packaging process. A composite priority dispatching rule considering working conditions of job agent, machine agent and packaging agent is proposed, where the packaging agent plays a significant role in developing the proposed dispatching rule. Then, this priority rule is compared with four widely-applied rules in literature, FCFS, SPT, COVERT, and ATC. Four measures of performance, i.e. maximum completion time, tardiness time, equipment utilization, order waiting time, are employed The results show that the proposed priority dispatching rule performs better overall, particularly in reducing order waiting time. Finally, an agent-based simulation model of a furniture manufacturing workshop is constructed using AnyLogic, to verify the proposed priority rule in dealing with random order arrivals.