摘要
针对考虑工人技能水平差异和加工时间随机性的设备、工人双资源约束的柔性作业车间调度问题(Flexible job shop scheduling problem,FJSP),构建双资源约束的柔性作业车间鲁棒调度模型(Dual resource constraint flexible job shop robust scheduling problem,DRC-FJRSP)。将粒子群算法(Particle swarm optimization,PSO)的学习机制引入差分算法(Differential evolution,DE)的交叉部分,设计PSO-DE混合求解算法,实现问题的加速求解并有效避免搜索停滞的窘境。选取DE算法、遗传算法(Genetic algorithm,GA)、PSO算法作为对比算法进行仿真试验,试验结果显示,所提出的PSO-DE混合求解算法平均鲁棒性为5.331,在三个仿真算例调度makespan求解中表现最好。验证了DRC-FJRSP模型的合理性以及PSO-DE混合求解算法的鲁棒性。最后,以一个制造企业的FJSP为例进行求解,给出调度甘特图。研究解决了设备、工人资源不足给FJSP带来的加工冲突,降低了加工时间的随机扰动,优化了调度的makespan。
Aiming at the flexible job shop scheduling problem(FJSP)considering the difference in workers'skill level and the randomness of processing time,the dual resource constraint flexible job shop robust scheduling problem(DRC-FJRSP)model is constructed.The learning mechanism of particle swarm optimization(PSO)is introduced into the cross part of differential evolution(DE),and a PSO-DE hybrid solution algorithm is designed to realize the accelerated solution of the problem and effectively avoid the dilemma of search stagnation.DE,genetic algorithm(GA)and PSO are selected as comparison algorithms for simulation experiments.The experimental results show that the average robustness of the proposed PSO-DE hybrid solution algorithm is 5.331,which performs best in scheduling makespan solutions of three simulation examples.The rationality of DRC-FJRSP model and the robustness of PSO-DE hybrid solution algorithm are verified.Finally,taking the FJSP of a manufacturing enterprise as an example to solve the problem,the scheduling Gantt chart is given.The research solves the processing conflict caused by the shortage of equipment and worker resources to FJSP,reduces the random disturbance of processing time,and optimizes the makespan of scheduling.
作者
梁志珍
王晓佳
LIANG Zhizhen;WANG Xiaojia(School of Management,Hefei University of Technology,Hefei 230009)
出处
《机械工程学报》
EI
CAS
CSCD
北大核心
2024年第6期114-126,共13页
Journal of Mechanical Engineering
基金
国家自然科学基金资助项目(U2001201,61876055)。
