离散变量桁架结构拓扑优化的改进混合遗传算法

An improved hybrid genetic algorithm for discrete topology optimization of trusses

为了避免在结构拓扑优化过程中杆件和节点的增删带来计算上的麻烦,在对桁架结构进行受力分析的基础上设计了一些启发式准则来产生可能的拓扑结构形式,然后采用一种改进的混合遗传算法进行截面优化.混合遗传算法将离散复合形法引入到遗传算法中,一方面利用遗传算法为离散复合形法提供可行点;另一方面利用离散复合形法对遗传算法种群中的可行个体和不可行个体进行改进,从而提高了遗传算法的局部寻优能力,并对标准遗传算法在选择、交叉和变异操作上作了一些改进.它将两种算法的优点集中在一起,同时又弥补了两者的不足.算例的结果表明,该方法用于桁架结构拓扑优化是简单、快速和有效的.

In order to avoid the calculating trouble of additions and deletions of the unit and the node, a heuristic means for producing the topology-structure shape was proposed, based on the analysis of truss structure, and then an improved hybrid genetic algorithm was used to process cross-section optimization. By bringing discrete complex method into generic algorithm, the hybrid generic algorithm, on one hand, provides the feasible point for discrete complex method through generic algorithm; on the other hand, it improves the feasible individuals and the infeasible individuals of the generic algorithm through the discrete complex method, thus enhancing the local searching capability. Moreover, the simple genetic algorithm was improved in the operation of selection, crossover and mutation. It combines the advantages of the two methods and overcomes the disadvantages of both. The results by exemplification show that this method is simple, rapid and efficient for topology optimization of truss.