基于粒子群算法的板结构离散优化设计  

OPTIMAL DESIGN OF PLATE STRUCTURES WITH DISCRETE VARIABLES BASED ON PARTICLE SWARM OPTIMIZERS

作　　者：刘锋[1] 张俊葆[1] 李丽娟[1] 

机构地区：[1]广东工业大学工程力学研究所,广州510006

出　　处：《工程力学》2009年第4期27-31,37,共6页Engineering Mechanics

基　　金：国家自然科学基金项目(10772052);广东省自然科学基金项目(06104655;8151009001000042)

摘　　要：该文将粒子群优化算法与有限元法结合,对板结构进行优化设计。采用映射函数法将离散变量按一定规律排序后,用离散值的序号代替其离散值进行搜索,从而将离散变量的优化问题转化为连续变量的优化问题。利用有限元方法结合粒子群方法开发了板结构优化模块,分别采用三种粒子群优化算法对三个算例进行了优化计算,计算结果表明:粒子群优化算法求解板结构的优化设计问题是有效可行的,而启发式粒子群算法是最有效的。Particle swarm optimizers （PSO） are combined with finite element method to optimize plate structures with discrete variables. The mapping function method is used to range the discrete variables and a sequential numbers is obtained. Thus, the searching for discrete variables is replaced by searching the sequential numbers of the discrete variables. Therefore a discrete optimization problem is converted into a continuous optimization one. The calculation module for plate optimal problem is developed with particle swarm optimizers based on finite element method. Three different optimizers are applied to three optimal design examples, showing the proposed method is feasible and the heuristic particle swarm optimizer （HPSO） is the most efficient method.

关 键 词：板结构 离散变量 微粒群算法 优化设计 有限元法 

分 类 号：TU311.4[建筑科学—结构工程]