基于强度Pareto进化的注塑机注射性能多目标优化  

Multiobjective injecting performances optimization of injection molding machinery based on strength Pareto evolutionary algorithm

作　　者：李中凯[1] 谭建荣[1] 冯毅雄[1] 裘乐淼[1] 

机构地区：[1]浙江大学CAD&CG国家重点实验室,浙江杭州310027

出　　处：《计算机集成制造系统》2007年第11期2162-2168,2183,共8页Computer Integrated Manufacturing Systems

基　　金：国家863/CIMS主题资助项目(2003AA413310);国家自然科学基金资助项目(60573175)。~~

摘　　要：为实现大型注塑机注射性能的优化设计,构建了注射压力、注射速率和注射功率优化模型,应用多目标进化算法,系统分析了影响注射性能的各方面因素。改进强度Pareto进化算法,引入模糊C均值聚类,加快外部种群的聚类过程。采用约束Pareto支配和浮点数、二进制混合染色体编码策略,一次运行就能求得分布均匀的Pa-reto最优解集,并使用基于集合理论的方法选择一个最优解。试验分析表明:结合了强度Pareto进化算法与模糊C均值聚类方法的混合算法在提高注射综合性能的同时,能够获得比线性加权法分布性更好的Pareto前沿;且与强度Pareto进化算法相比,显著缩短了运算时间,具有较高的效率与鲁棒性。To optimize the injecting performances for large injection molding machinery,the optimization model for injection pressure,speed and power was established,and the factors affecting injection performances were analyzed by using multi-objective evolutionary algorithms.The Strength Pareto Evolutionary Algorithm（SPEA） was improved by introducing the Fuzzy C-Means（FCM） clustering algorithm to accelerate the clustering procedure of the external population.With the constraint Pareto dominance concept and the float,binary chromosome representation scheme,a well-distributed Pareto optimal set was achieved in a single operation.Then a solution was extracted as the best compromise one based on set theory.Experimental results illustrated that FCM-SPEA could improve general injection performance,and acquire the Pareto front with better distribution than linear combination at the same time.Compared to SPEA,the FCM-SPEA drastically reduced the computational time with better efficiency and robustness.

关 键 词：强度Pareto进化算法 模糊C均值聚类 多目标优化 大型注射成型机 注射性能模型 

分 类 号：TH45[机械工程—机械制造及自动化]