改进的群搜索优化算法及其应用  

An improved group search optimizer algorithm and its application

作　　者：李丽娟[1] 张雯雰 徐小通[1] 刘锋[1] 

机构地区：[1]广东工业大学土木与交通学院,广东广州510006

出　　处：《空间结构》2010年第2期17-23,共7页Spatial Structures

基　　金：国家自然科学基金资助项目(10772052);广东省自然科学基金资助项目(06104655,8151009001000042,9151009001000059)

摘　　要：群搜索(GSO)算法是一种新的群智能优化算法,适用于结构优化设计.本文通过对GSO算法进行改进,简化了算法的计算过程,提高了优化性能.对算法的改进主要有二个方面:一是采用随机搜索,放弃了按角度搜索的方式;二是在生成个体新位置时,增加了一个随迭代次数递减的控制变量——分量变异概率,用于限制允许变异的维的数量.通过对经典桁架算例的优化以及与标准GSO算法的计算结果比较,可以看出改进后的群搜索优化算法(SGSO)具有更好的收敛速度和收敛精度,SGSO算法的结构比GSO算法更简单、易于实现并且计算用时更少.Group Search Optimizer（GSO） is a novel swarm intelligence algorithm,which has a superior performance on structural optimization problems.An improved group search optimizer（SGSO） is presented in this paper to simplify the calculation process and improve the optimal performance.The modification has two main aspects.Firstly,random search strategy is employed to replace the visual search strategy.Secondly,mutation probability,which may decrease by the iteration times,is introduced to limit the number of dimensions of allowed variations.SGSO is conceptually simple,timesaving and easy to implement.The SGSO is verified and compared with the GSO algorithm used for the designs of several planar and spatial truss structures.The results show that the SGSO has preferable convergence rate and convergence accuracy.

关 键 词：改进的群搜索优化算法 结构优化 收敛速度 收敛精度 

分 类 号：TU12[建筑科学—建筑理论]