Parameter optimization in GA for job-shop scheduling problem

Job shop scheduling problem is a hot spot in today's intelligent manufacturing. Therefore, how to get JSSP optimal solution is a key problem in this field. The development of genetic algorithm has been relatively mature, and gets a better result in JSSP. However, the relevant research on how to further optimize on this basis is still relatively less. In order to improve the solution quality of genetic algorithm, we herein proposed a computational framework for job shop scheduling problem. Firstly, we use the design of experiment (DOE) method to get some samples and carry out a sensitivity analysis, and obtain the parameter factors with the greatest influence on this problem. Secondly, establish the approximation model and optimize it to get the optimal solution. Finally, we verified the results by an example and proved this method greatly improves the quality and efficiency of JSSP solution.