基于灰熵并行分析法的多目标作业车间调度优化

摘要

提出利用信息熵理论与灰色关联分析法并行地处理多目标优化问题，将多目标优化的目标值构成数据序列，挖掘序列关系实现多目标优化。首先，并行的对目标值序列计算灰关联系数以及熵值权重，之后将信息熵与灰关联系数结合计算灰熵并行关联度，建立灰熵并行分析法。最终，利用灰熵并行关联度作为优化算法的适应值计算策略，以该策略引导智能优化算法进化。建立作业车间调度问题的三目标优化模型，以灰熵并行分析法为基础，分别应用差分算法、遗传算法解决三目标作业车间调度问题，验证新方法的可行性。实验表明：新方法均能使两算法收敛且得到分布均匀的 Pareto 前端，表明其有效和可靠。同时，差分算法得到的解较遗传算法的解具有明显的优势。%In this paper, the multi-objective optimization problem was solved with the theory of information entropy and gray correlation analysis method in parallel. The objective function values were used to structure a data sequence. The multi-objective optimization was complicated by using data sequence relation model. Firstly, the grey relational coefficient and the entropy weight were calculated in parallel based on multi-objective value sequence. Then, the information entropy and the grey relational coefficient were combined and used to calculate the grey entropy parallel relational degree (GEPRD), that is, the grey entropy parallel analysis method was built. Finally, the GEPRD was used as the fitness value calculation strategy to guide the evolution of the heuristic algorithm. The Tri-objectives optimization model of job shop scheduling problem was established. In order to verify the feasibility of the new method ,the grey entropy parallel analysis method was testified with differential algorithm and genetic algorithm respectively to solve the Tri-objectives job shop scheduling problem. Experimental results show that this method is effective ,with this method, the convergent and uniform distribution of Pareto can be obtained by this two algorithms. Indicated that it was effective and reliable.The solutions obtained by the difference algorithm are better than those of genetic algorithm.