A cooperative control method of multiple CSPS system based on part-flow information

The collaborative optimization control problem of a multiple conveyor-serviced production station (CSPS) system is concerned in this paper. The objective is to improve the load balance, and to improve the part-processing rate of the whole system at the same time, by choosing a suitable control strategy for each station. Firstly, partial part information is introduced to describe the state of the production system. A part-flow evaluating method is given to modify the performance function of each station from the perspective of processing the part flow. Then, a simulation model for strategy learning is built using Q-learning algorithm based on performance potential. Finally, the learning effects with either fixed or Erlang-distributed processing time are analyzed through simulation experiment. The evaluation results show the effectiveness of the proposed mechanism.