基于免疫粒子群-齿行法的桁架结构分层优化  

Layered optimization of truss structures based on I-PSO and zigzag method

作　　者：陆金钰[1,2] 沈圣[3] 牛畅[3] 杨湛[3] 

机构地区：[1]东南大学混凝土及预应力混凝土结构教育部重点实验室,南京210096 [2]东南大学土木工程学院,南京210096 [3]东南大学吴健雄学院,南京210096

出　　处：《东南大学学报（自然科学版）》2016年第2期335-340,共6页Journal of Southeast University：Natural Science Edition

基　　金：国家自然科学基金资助项目(51008065;51208263);东南大学优秀青年教师教学科研资助计划资助项目(2242014R30005)

摘　　要：针对考虑应力约束、位移约束和压杆稳定约束的桁架结构进行形状和尺寸优化,提出了一种基于免疫粒子群优化算法和齿行法的桁架结构优化算法.对形状和尺寸2类耦合变量进行分层处理.内层考虑各类约束条件,利用收敛速度较快的齿行法对给定形状的结构进行截面优化,使其自重最轻,将约束问题转化为非约束问题;外层则通过全局搜索能力较强的免疫粒子群算法对结构形状进行全局搜索,得出最终优化结果.通过37杆平面简支桁架桥和25杆空间桁架2个经典算例验证了所提算法的有效性.结果表明,所提算法采用内外层嵌套搜索,能够更好地处理2类变量的耦合关系,有效缩小了解空间的范围,具有良好的优化效果和较高的搜索效率.The shape and size optimization of truss structures with consideration of stress constraints,displacement constraints and pressure bar stability constraints is carried out,and an optimization algorithm based on the particle swarm optimization algorithm with immunity（ I-PSO algorithm） and the zigzag method is presented. Two coupled variables,the shape and the size,are hierarchically processed. In the inner layer,the zigzag method with a fast convergence rate is applied to optimize the cross-section of the structure with a certain shape to achieve the lightest weight under various constraints,turning the constrained problem into an unconstrained one. In the outer layer,the I-PSO algorithm makes use of its global searching ability to obtain the final optimized results. Two typical examples,a 37-bar plane truss bridge and a 25-bar spatial truss,are illustrated to validate the effectiveness of the proposed algorithm. The results showthat with inner and outer layer nesting searching,this algorithm can deal well with the coupled variables and reduce the solution space effectively.The proposed algorithm has a good optimization result and high efficiency.

关 键 词：桁架形状与尺寸优化 分层优化 免疫粒子群优化算法 齿行法 

分 类 号：TU323.4[建筑科学—结构工程]