基于Pareto改进的混合算法求解多目标柔性车间调度问题  

Multi-objective flexible shop scheduling problem with hybrid algorithm based on Pareto

作　　者：赵勇 史亚斌[1] 何军红[2] 刘赛[2] 马国伟 ZHAO Yong;SHI Yabin;HE Junhong;LIU Sai;MA Guowei(Xi'an High Voltage Apparatus Research Institute Co.,Ltd.,Shaanxi Xi'an 710077,China;School of Marine Science and Technology,Northwestern Polytechnical University,Shaanxi Xi'an 710072,China)

机构地区：[1]西安高压电器研究院有限责任公司,陕西西安710077 [2]西北工业大学航海学院,陕西西安710072

出　　处：《工业仪表与自动化装置》2021年第3期10-15,49,共7页Industrial Instrumentation & Automation

摘　　要：针对多目标柔性作业车间调度问题,该文建立优化目标为最大完工时间、机器平均相对空闲率以及机器总负荷最小化的数学模型,并设计一种基于Pareto改进的自适应混合算法(NGA-PSO)。其算法采用分层结构相结合,底层采用基于隔离的小生境技术(Niche genetic algorithm,NGA),上层采用粒子群算法(Particle swarm optimization,PSO)。为提高算法的收敛效率和求解精度,提出了改进策略,采用适应度值分配策略作为种群选择的评价标准;设计动态的交叉变异概率,使算子在迭代过程自适应地对种群的寻优操作进行调整。最后,针对10个单目标基准案例与3个多目标典型案例进行仿真求解,通过与其他前沿算法进行对比验证NGA-PSO算法的优越性。In order to solve the multi-objective flexible job shop scheduling problem,the paper establishes a mathematical model that optimizes the maximum completion time,the average relative idle rate of the machine,and minimizes the total load of the machine,and designs an improved adaptive hybrid algorithm based on Pareto(NGA-PSO).The algorithm adopts a hierarchical structure,the bottom layer adopts isolation-based niche technology(Niche genetic algorithm,NGA),and the upper layer adopts particle swarm optimization(PSO).In order to improve the convergence efficiency and solution accuracy of the algorithm,an improved strategy is proposed.The fitness value allocation strategy is used as the evaluation standard for population selection;the dynamic cross-mutation probability is designed to make the operator adaptively optimize the population during the iterative process Make adjustments.Finally,10 single-objective benchmark cases and 3 multi-objective typical cases are simulated and solved,and the superiority of the NGA-PSO algorithm is verified by comparison with other cutting-edge algorithms.

关 键 词：多目标优化 混合算法 小生境技术 粒子群算法 

分 类 号：TH165[机械工程—机械制造及自动化]