基于Pareto GA的机械手结构/控制多目标优化设计

针对机械／控制系统并行设计的多变量、多目标优化特点，以单臂机械手为对象，研究机电融合优化设计模型和多目标并行优化的算法．根据单臂机械手的动态特性和控制性能要求，采用极点配置方法构造了基于PD反馈控制的机电融合优化设计模型，通过Pareto遗传算法（PGA）对带约束的机电融合多目标优化问题的并行求解进行了研究，并对得到的Parato可行解集进行了分析．在PGA优化过程中采用了目标函数的归一化处理、分级罚函数法将约束问题转化为无约束优化、群体分级和Pareto解集过滤器等技术，并采用聚类方法对Pareto解集中解的个数、解集的特性和解的多样性进行控制，使优化解收敛到一个非劣点集．

The Multiobjective Optimization of Mechatronical Sysytem of Single Manipulator Based on Pareto GA

The integrated structure/control modeling of single-manipulator and concurrent optimal methodology for multiobjective were studied. According to the requirement of dynamical and control stability performance, the optimal model with multiobjective and constraints was proposed by combination of the dynamics of flexible structure and poles configuration, and solved by Pareto GA(PGA). In the process of PGA optimization, the technique of regularization of objectives, graded penalized function technique to remove constraints, Pareto-rank of population and Pareto set filter are used. To protect the feature and multiformity of Pareto set, a clustering analysis technique is used to limit the number of individual in Pareto set.