基于改进遗传算法的可重构制造系统快速配置
Rapid Configuration of Reconfigurable Manufacturing Systems Based on Improved Genetic Algorithms
洪胡平 1张为民 1谢树联1
作者信息
- 1. 同济大学机械与能源工程学院,上海 201800
- 折叠
摘要
针对可重构制造系统的快速配置提出了一种多目标优化方法,考虑 3 个优化目标:设备负载均衡度、产品生产时间以及产品生产成本.在第三代多目标遗传算法(NSGA-Ⅲ)的基础上结合邻域搜索构建了多目标遗传-邻域搜索算法(NSGA-NS),以从配置解空间中寻找一系列前沿解.通过一个液压阀阀块的制造系统配置案例对算法进行验证,并与归档式多目标模拟退火算法、多目标粒子群算法以及一种改进的多目标粒子群算法进行了比较,结果表明,NSGA-NS与NSGA-Ⅲ在前沿解质量上显著优于其他算法,NSGA-NS在保持了NSGA-Ⅲ的全局寻优性能的同时显著提高了收敛速度和计算速度.
Abstract
A multi-objective optimization method is proposed for the rapid configuration of reconfigurable manufacturing systems,considering three optimization objectives:equipment load balance,product production time,and product production cost.A multi-objective genetic-neighborhood search algorithm(NSGA-NS)is constructed based on the third generation multi-objective genetic algorithm-Ⅲ(NSGA-Ⅲ)combined with neighborhood search to find a series of front solutions from the configuration solution space.The algorithm is validated by a configuration case of a hydraulic valve block manufacturing system,and compared with the ar-chived multi-objective simulated annealing,multi-objective particle swarm optimization,and an improved multi-objective particle swarm optimization.The results show that NSGA-NS and NSGA-Ⅲsignificantly out-perform each other in terms of front solutions quality,NSGA-NS significantly improves the convergence speed and calculational speed while maintaining the global search performance of NSGA-Ⅲ.
关键词
可重构制造系统/配置/多目标优化/混合启发式算法Key words
reconfigurable manufacturing systems/configuration/multi-objective optimization/hybrid heu-ristics引用本文复制引用
基金项目
国家重点研发计划(2022YFE0114100)
出版年
2024
NETL
NSTL
万方数据