A phased intelligent algorithm for dynamic seru production considering seru formation changes

Product information is constantly changing in actual production; examples include batch due date changes, batch size changes, and the arrival of emergency batches. However, most studies of seru production have focused on the static seru systems. Therefore, this study investigated the total tardiness of the dynamic seru production considering seru formation changes. When product information changes, four decision processes were considered for optimal solutions: 1) selection of seru (s) for re-formation ( SSR ); 2) batch scheduling on the seru ( s ) waiting for re-formation ( BSSWR ); 3) seru re-formation; and 4) seru re-scheduling. A dynamic seru production model was established here based on the above process, and a phased intelligent algorithm was proposed for solving this problem. SSR was optimized using a genetic algorithm. The earliest due date ( EDD ) heuristic algorithm was used for BSSWR . Moreover, an adaptive cooperative coevolution algorithm was proposed, in which Q-learning was used for determining whether to optimize seru re-formation or seru re-scheduling. After extensive experiments, the dynamic seru production system considering the seru formation changes reduced the total tardiness by an average of 54 % compared with the original solution, and by an average of 15 % compared with the dynamic seru production system considering only the seru scheduling changes.