Flexible flow shop scheduling by intelligent multi-agents

This paper presents some important improvements to a previously proposed intelligent production system dealing with a dynamic flexible flow shop scheduling problem under a multi-stage multi-machine factory environment. These improvements greatly help upgrade the overall system performance under stable demand situations as well as under fluctuated demand situations, build the system robust against demand increase, and raise the systems machine utilization rate. The research objective is to minimize the total earliness and tardiness penalties of all jobs during any given period of time. The system works on the basis of multi-agent feedbacks that are conducted by agents which collect realtime information, make decisions, and work interactively to give corresponding solutions to each job. Comparison between the previous system and the improved one has been carried out, and the experimental results demonstrate the effectiveness of the proposed improvements under various system situations.