论文部分内容阅读
混合变量桁架结构形状优化中采用并行计算思想,在单台计算机上实现了一种类似分层遗传算法的拟分层遗传算法。该算法能够产生更加平等的竞争机会,提供更多的优良个体,提高了种群多样性,同时不用人为的控制信息交换,再加上多层分级控制,一定程度上避免了标准遗传算法容易出现的“早熟”现象,加快了收敛速度,具有很高的搜索效率。用拟分层遗传算法解决25空间杆桁架结构形状优化问题的结果表明,这是一种解决具有连续、离散混合变量的桁架结构优化设计问题的很有效方法。
Hybrid variable truss structure shape optimization using parallel computing ideas in a single computer to achieve a hierarchical genetic algorithm similar to the hierarchical genetic algorithm. The algorithm can generate more equal competition opportunities, provide more excellent individuals and improve the diversity of the population, without the need for artificial control of information exchange, coupled with multi-level hierarchical control, to a certain extent, to avoid the standard genetic algorithm prone to “Premature” phenomenon, speeding up the convergence rate, with high search efficiency. The results of solving the shape optimization problem of 25-space truss with quasi-hierarchical genetic algorithm show that this is a very effective method to solve the optimal design of truss structure with continuous and discrete variables.