Autonomous decentralized scheduling system for flowshop problems with storage and due-date penalties

An autonomous decentralized scheduling system, which aims at minimizing the sum of changeover cost, storage cost, and due-date penalty, is proposed fo r flowshop problems. In the proposed system the scheduling system belonging to each production stage gradually generates a plausible schedule through repeated generation of schedules for the stage and repeated exchange of dat a among the production stages. When intermediate storage cost is included i n the objective function, calculation of the optimal starting times of jobs is fairly difficult, even if the production sequence of jobs at each stage is fixed. In this study, an algorithm fur effective calculation of the sta rting times of jobs, which can minimize the storage cost, has been develope d. By embedding this algorithm in the scheduling algorithm, an autonomous d ecentralized scheduling system, which realizes Just-in-Time production at a ll production stages, is developed. It is also shown that the proposed syst em is applicable to the case in which the space available for intermediate storage is limited. The results of example problems show that the schedules obtained by the proposed system are significantly better than those obtain ed by conventional methods.