基于小生境粒子群算法的公差多目标优化设计  

Optimization design of multi-objective tolerances based on niche PSO algorithm

作　　者：刘海博[1] 刘检华[1] 何永熹[1] 郭崇颖[1] 蒋科[1] 

机构地区：[1]北京理工大学先进加工技术国防重点学科实验室,北京100081

出　　处：《计算机集成制造系统》2015年第3期585-592,共8页Computer Integrated Manufacturing Systems

基　　金：国家自然科学基金资助项目(51275047);"十二五"国防基础科研资助项目(A0420132503)~~

摘　　要：为解决产品设计中的公差优化问题,提出一种基于小生境粒子群算法的公差多目标优化方法。以加工成本、质量损失成本和公差敏感性为优化目标,以装配功能要求和加工能力为约束条件,建立了公差多目标优化模型。对标准粒子群算法进行改进,根据小生境数和Pareto优劣性确定孤立粒子,并通过个体历史最优粒子与孤立粒子的变异、选择操作更新粒子的个体历史最优位置;利用Pareto支配数排序更新粒子群的全局最优位置。利用改进的粒子群算法对公差多目标优化模型进行求解,得到分布均匀的Pareto前沿。设计并开发了原型系统,通过实例验证了该方法的有效性。To solve the problem of tolerance optimization in product design, an optimization approach for multi-objec- tive tolerance based on niche Particle Swarm Optimization （PSO） algorithm was proposed. By taking manufacturing cost, quality loss cost and tolerance sensibility as the design objectives, a multi-objective model of tolerance optimi- zation was presented which were subject to the assembly functional requirement and the machining capability. The traditional particle swarm optimization was improved, and the outlier particle was identified by niche number and Pa- reto dominance of every particle, with which the mutation and selection operation of individual optimal particle and outlier particle were applied to update the optimal position for the particles. The global optimum of particle swarm was updated by Pareto dominance number. The Pareto front of uniform distribution was obtained by using niche PSO algorithm to solve the multi-objective model of tolerance optimization. A prototype system was developed and an example was tested to verify the feasibility of the proposed approach.

关 键 词：公差优化 多目标 PARETO最优解 小生境粒子群算法 产品设计 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]