改进的遗传算法在结构优化设计中的应用

Improved genetic algorithm and its Applications in optimization design of structures

遗传算法(GA)是基于达尔文进化论和遗传学说形成的一种崭新的优化算法。它具有全局收敛性和并行性;对先验知识要求较少,具有很强的适应性。针对结构优化设计方法中存在的局限性,将改进的遗传算法用于结构优化设计中。改进的GA采用以下措施提高搜索效率:(1)动态调整变量区间和GA参数;(2)在每一轮进化结束后重新初始化群体,开始新的进化;(3)将最优个体保留到下一轮。据此编制了计算机程序,并将其应用到一个桁架结构的优化实例中。运行结果表明,改进后的遗传算法用于结构优化设计能够有效地避免陷入局部最优解的现象,提高了搜索效率,具有较强的适应性。

As a newly developed optimization algorithm, the genetic algorithm(GA) is devived from the theory of Darwinism and genetics. It doesn't require prior knowledge of the problem and has the advantage of global convergence, parallelism and adaptability. An improved genetic algorithm is introduced to overcome the shortcoming of the traditional structure design optimization. The new GA improves the searching efficiency by two means: (1) adjust variables bound and GA's parameters dynamically; (2) reinitialize the population at the end of each evolution; (3) reserve the best individual. A computing program is developed and applied to a truss structure optimization problem. The example shows that the improved genetic algorithm can avoid local-convergence effectively and has good searching efficiency and adaptability.