基于分段遗传算法的桁架拓扑优化设计

摘要

In the structural optimization design of aircraft and automobile, the truss structure should have enough strength,stiffness and sufficient reliability,and maintain the weight of their structure as light as possible and enhance their efficiency in optimization process. A GA operation process was divided into two phases,i. e. coarse search and fine search,the crossover probability (Pc) and the mutation probability (Pm) were adaptive setting in the process respectively,and the penalty function was introduced in the truss topology model to get rid of the restriction. Finally,the improved GA was applied to analyze the topological optimization design of the truss structures,the iteration time of the optimization design model is far less than that of the simple genetic algorithm. It is proved the improved genetic algorithm is practicality and feasibility.%在飞机、汽车等结构优化设计中,为使桁架结构具备足够的强度、刚度和可靠性,并同时实现结构轻量化和过程高效率的目标,将遗传算法的运行过程分为“粗搜”和“精搜”两个阶段,对交叉概率(Pc)和变异概率(Pm)进行自适应设置,同时在桁架拓扑建模中引入罚函数去约束化.桁架结构的拓扑算例验证表明:该优化设计方法的迭代次数远小于简单遗传算法,证明方案实用可行.