An adaptive multiobjective evolutionary algorithm for dynamic multiobjective flexible scheduling problem

doi:10.1002/int.23090

2022-10-14

Weiwei Yu

Li Zhang

Ning Ge *

DOI：10.1002/int.23090

原文链接

原文求助

摘要

En 中文

There are various uncertain disturbances in the actual manufacturing environment, which makes dynamic multiobjective flexible scheduling problem of flexible job shop (MDFJSP) become the research focus in the field of optimal scheduling. In this paper, MDFJSP in the environment of temporary order insertion uncertainty is studied, and a multiobjective dynamic scheduling scheme based on rescheduling index and adaptive nondominated sorting genetic algorithm (NSGA-II) is proposed. First, based on the actual manufacturing environment, the mathematical model of the traditional flexible job shop scheduling problem is improved, and the multiobjective dynamic rescheduling model of flexible work center is established. Then, the existing rescheduling mechanisms are summarized, and a rescheduling hybrid driving mechanism based on the rescheduling index is proposed to enable it to reschedule and drive according to the actual situation. Finally, the shortcomings of the traditional multiobjective scheduling algorithm NSGA-II are analyzed, the adaptive cross mutation strategy and the simplified harmonic normalized distance measure method are proposed to improve it, and an adaptive multiobjective dynamic scheduling algorithm NSGA-II (MDSA-NSGA-II) is formed. To analyze the performance of this algorithm, the performance of this algorithm is compared with five classical flexible job shop multiobjective scheduling algorithms in international general examples, and the effectiveness is verified by real aircraft production examples. The experimental results fully show that MDSA-NSGA-II has good performance in solving MDFJSP.

Keyword:

congestion measurement strategy

dynamic multiobjective flexible scheduling problem of flexible job shop

multiobjective optimization

NSGA-II

rescheduling mechanism