复合形遗传算法在离散变量桁架结构拓扑优化设计中的应用

Application of Complex Genetic Algorithm to 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的奇异解,设计了一些启发式准则来产生结构可能的拓扑结构形式,再采用混合遗传算法———复合形遗传算法进行截面优化。把复合形法嵌入到遗传算法中,利用复合形法对群体中的可行个体和不可行个体分别进行处理,可提高遗传算法种群的质量,有利于最优解的搜索。该方法既有复合形法快速高效的特点,又有遗传算法全局性好的特点。算例的结果表明,该方法用于桁架结构拓扑优化设计是有效的。

In order to avoid odd solution resulted from additions and deletions of units and nodes in the process of topology optimization, one kind of heuristic mean for producing the topology-structure shape is proposed, and hybrid genetic algorithm (complex genetic algorithm) is used to process cross-section optimization. Complex method is inserted into genetic algorithm, to process the feasible and infeasible individuals in order to improve the qualities of the colony. It is beneficial for searching the best solution. The method has the advantage of high efficiency of complex form method, and has the advantage of globally searching capability of genetic algorithm. Examples show that this method is effective for topology optimization of truss.