期刊名称：International Journal of Professional Business Review
电子版ISSN：2525-3654
出版年度：2018
卷号：3
期号：2
页码：172-186
DOI：10.26668/businessreview/2018.v3i2.49
出版社：Tutor Treinamento e Desenvolvimento
摘要：One of the discussed topics in scheduling problems is Dynamic Flexible Job Shop with Parallel Machines (FDJSPM). Surveys show that this problem because of its concave and nonlinear nature usually has several local optimums. Some of the scheduling problems researchers think that genetic algorithms (GA) are appropriate approach to solve optimization problems of this kind. But researches show that one of the disadvantages of classical genetic algorithms is premature convergence and the probability of trap into the local optimum. Considering these facts, in present research, represented a developed genetic algorithm that its controlling parameters change during algorithm implementation and optimization process. This approach decreases the probability of premature convergence and trap into the local optimum. The several experiments were done show that the priority of proposed procedure of solving in field of the quality of obtained solution and convergence speed toward other present procedure.
关键词：Optimization; Parallel Machines; Genetic Algorithm; Dynamic and Adjustment of controlling parameters
