基于遗传算法的网格结构的拓扑优化

遗传算法是一种新兴的基于遗传进化机理的寻优技术。它与常规算法的不同之处在于不受初始值影响、从多个初始点开始寻优，并采用交叉、变异和移民算子避免过早地收敛到局部最优解，可获得全局最优解。该算法不必求导计算，编程简单、快捷，尤其适用于具有离散变量的结构优化设计。本文利用遗传算法对在应力、位移约束下的网格结构进行拓扑优化，利用对称性对杆件进行分组，使优化后的结构模型仍然保持对称，具有工程应用价值，并达到降低造价的目的。计算实例表明了该算法的有效性。

Topological optimization of reticulated structures based on Genetic Algorithms

Genetic Algorithms (GA) is recently developed search strategy based on the rules of genetic evolution. It differs from existing mathematical programming methods in that the GA uses a population of initial points in contrast to the single-point approach by the traditional ones. The optimal solution by GA is no longer dependent on initial search values, and will guarantee a global optimum. Derivative calculation is not necessary and computer program is simple. Genetic algorithm is best suited for structural optimizations which have discrete variables. In this paper, the algorithm developed is used to deal with structure topological optimization problems with stress and displacement constraints. A numerical example shows that it is of high efficiency and good adaptability.