特殊工艺约束下并行机多目标调度问题

Research on Multi-objective Scheduling Problem Subjected to Special Process Constraint on Parallel Machines

针对并行机多目标调度问题的特点,提出了一个多目标调度模型,设计了一个基于向量组编码的遗传算法,并对其初始种群、选择、交叉和变异等方法进行了研究。此算法的编码方法能有效地反映实际调度方案,并引入免疫算子,保证了种群的多样性和种群的质量,加快了收敛速度。仿真结果表明,此算法是有效的,优于没引入免疫算子的遗传算法。

In connection with MOSP, a multi-objective scheduling model was presented, a genetic algorithm based on a vector group coding method was proposed, and the method of its initial population, selection, crossover, and mutation were studied. With its simple encoding method, the algorithm can effectively reflect the virtual scheduling policy. Meanwhile, an immune operator was adopted in order to guarantee the varieties and qualities of the colony. Numerical experiments showed that it is efficient, and it is superior to genetic algorithm without immune operator.