Parameter optimization of mechanical workshop scheduling parallel genetic algorithm based on orthogonal test

ZHANG Sheng-fang, WANG Guo-qing, MA Fu-jian, LIU Yu, YANG Da-peng, SHA Zhi-hua
(School of Mechanical Engineering, Dalian JiaoTong University, Dalian 116028, China)

Abstract: Aiming at the different performances affected by parameters in the scheduling parallel genetic algorithms, an orthogonal experiment method was used. The effects of eight factors on the computing time of the scheduling parallel genetic algorithm were studied, including the number of neutrons, the number of individuals, the generation gap, the crossover rate, etc. The primary and secondary order of factors was studied with the analysis of variance and range analysis. The interaction between generation gap and variation rate, crossover rate and variation rate was analyzed and the optimal horizontal combination of parameters was determined. The FT class typical scheduling problem with selected parameters was tested. The results indicate that using orthogonal experiment instead of the full factor test to optimize the parameters of parallel genetic algorithm can effectively improve the evolutionary process, reduce the computing time on the basis of ensuring the quality of the algorithm.

Key words: mechanical workshop production scheduling; parallel genetic algorithm; orthogonal test; parameter optimization