Abstract: In order to reduce the total delay time of multi workshop joint scheduling with serial parallel heterogeneous process constraints, a scheduling solution method based on knowledge guided genetic algorithm was proposed. Firstly, an extended process tree was used to describe the constraints of serial parallel heterogeneous processes, and the distribution of machines in multiple workshops was described based on an undirected graph. To address the constraints of the extended process tree during chromosome initialization and evolution, the concepts of the number of tight preceding steps and the number of remaining tight preceding steps were defined. Based on the number of remaining tight preceding steps, chromosome initialization and evolution methods were designed. In order to improve the evolutionary ability of genetic algorithms, the population evolutionary ability and the optimal individual evolutionary ability were used as knowledge to drive the evolutionary mode and direction of the algorithm, thus a solution method based on knowledge guided genetic algorithm was proposed. After experimental verification, the average total delay time of knowledge driven genetic algorithm scheduling is the smallest, at 30.8 hours, indicating that the algorithm has the best optimization performance in multi workshop scheduling. And the length of the total delay time box graph is the smallest, indicating that the stability of knowledge driven genetic algorithm is also good.

Key words: multi workshop collaboration, extended process tree, tight preceding steps, knowledge guided, genetic algorithm