关键词: 多目标优化, 作业车间调度, 遗传算法, Giffler&Thompson算法

Abstract: To solve multi-objective Job Shop scheduling problem, a double-population genetic algorithm based on forward-backward scheduling approach and Giffler-Thompson algorithm was proposed. The method reduced the solution space by means of active scheduling. Forward-backward genetic scheduling algorithm was proposed to optimize different objective functions in different populations and multi-objective problems were decomposed into various single-objective problems. In each generation of the evolving process, individual migration operator was applied to accelerate the parallel searching. An elitist arena's principle was presented to improve the efficiency of constructing the Pareto optimal solutions. The experimental results of the Benchmark instances taken from literature demonstrated the effectiveness of the algorithm proposed on solving multi-objective job shop scheduling problems.

Key words: multi-objective optimization, Job-Shop scheduling problem, genetic algorithm, Giffler&Thompson algorithm