空间映射粒子群算法用于电磁优化计算  

Space-mapping particle swarm optimization on electromagnetic optimization

作　　者：楼群 田雨波[1] 邱大为[1] 刘东[1] 王兆尹 

机构地区：[1]江苏科技大学电子信息学院,江苏镇江212003 [2]南京大学电子科学与工程学院,江苏南京210046

出　　处：《电波科学学报》2015年第2期217-223,共7页Chinese Journal of Radio Science

摘　　要：将粒子群算法与空间映射算法相结合,提出了一种空间映射粒子群优化算法,并用于电磁问题的优化计算.算法实现过程中,将电磁仿真计算中精确网格剖分的计算模型作为精确模型,将粗糙网格剖分计算模型作为粗糙模型,在粒子群算法中计算粒子适应度前,使用粗糙模型结合基于卡尔曼滤波的映射关系,估计出粒子是否对算法最优解更新有效,并对有效的部分粒子做适应度计算,可以很大程度上减少算法的计算耗时.通过E型贴片天线和谐振腔缝隙天线的优化仿真说明了本算法的有效性.This paper proposes a kind of space-mapping particle swarm optimization by combining space-mapping with particle swarm optimization(PSO).In this algorithm,accurate-mesh simulation is regarded as accurate model,and coarse-mesh simulation is regarded as coarse model.Estimation of accurate result is gotten by Kalman filter mapping relation and coarse model,which can determine whether the particle is efficient for updating PSO or not,before calculating the fitness of each particle in PSO.This,in turn,can avoid the simulation of inefficient particle at some level,directly leading to reducing the computation of PSO.Simulations of E-kind patch antenna and cavity slot antenna proof the efficiency of the algorithm.

关 键 词：粒子群算法 空间映射 电磁优化 

分 类 号：TM15[电气工程—电工理论与新技术] TP18[自动化与计算机技术—控制理论与控制工程]