The multi-layered job-shop automatic scheduling system of mould manufacturing for Industry 3.5

Abstract This study combined the First In First Out (FIFO) with the Earliest Due Date (EDD) heuristics and the on-site experience of actual production lines. According to such restrictive conditions as the available processing machine groups, the hierarchical relations among components, the due date of moulds as designated in the plan of mould manufacturing, this study conducted a quick analysis with an expert system to determine the optimal sequence and scheduling of machines. The machine management agent was then used for automatic pre-scheduling of the earliest available capacities of the machines to obtain preliminary scheduling. Firstly, the Genetic Algorithm (GA) was used to find the better scheduling sequences. After that, Ant Colony Optimization (ACO) was adopted to optimize the sequence determined by the expert system. It was able to effectively solve the problem of job-shop scheduling featuring a complicated hierarchical procedure for Industry 3.5 as a hybrid strategy between Industry 3.0 and to-be Industry 4.0. Practical mould cases with four-layer components were next used for a comparison of scheduling results. Three case studies show better results with integration of GA and ACO (GA + ACO) than GA or ACO alone. For the in-depth case study of 92 jobs, the scheduling based on the automatic scheduling of EDD spent 8% less work time than that based on FIFO. After the GA and ACO was adopted for the optimization of scheduling, the work span was further shortened by 10% with less computing time.