改进正交边界交叉算法及其在高频电磁场逆问题中的应用  

An Improved Normal Boundary Intersection Method for Multiobjective High Frequency Inverse Problems

作　　者：安斯光[1] 杨仕友[2] 刘磊[2] 

机构地区：[1]中国计量学院机电工程学院,浙江省杭州市310018 [2]浙江大学电气工程学院,浙江省杭州市310027

出　　处：《中国电机工程学报》2015年第21期5607-5613,共7页Proceedings of the CSEE

基　　金：国家自然科学基金项目(51407172);浙江省自然科学基金项目(LQ14E070003)~~

摘　　要：针对高频电磁场逆问题分析复杂、求解缓慢等现有问题,以给定方向增益最大化,同时副瓣电平最小化为目标,建立10个同性线性点源天线阵模型,并提出一种改进的正交边界交叉(normal boundary intersection,NBI)算法。为克服原NBI算法收敛速度慢、一次运行只得到一个Pareto解等固有缺陷,该文将理想平面分化为若干子区域,将每个种群个体投影到理想平面上,根据其在理想平面上的投影进行分类及评估,并将分类及评估结果体现在母代选择中,以实现一次运行搜索得到全部Pareot解平面;其次,为避免早熟现象和增加种群多样性,提出双重外部档案概念,并引入一种新的精英策略。给出的典型数学函数验证了该文算法的快速性和有效性,高频电磁场逆问题的数值计算结果说明了该文算法的优越性和工程应用价值。An improved normal boundary intersection（NBI） method is proposed for the high frequency inverse problems. To deal with the complicated situations and time-consuming computations of the high frequency electromagnetic inverse problems, a model of a 10 linear antenna array with the same nature sources is built which is to maximize the direction gain in the fixed direction and minimize the side slope level（SLL） at the same time. To overcome the shortage of original NBI methods in finding one Pareto solution in a single run and make the Pareto solutions can evolve at the same time, the utopia plane is divided into sub-domains, project each individual on the ideal plane, classify and assessment the individuals based on the projections which is used in parents selection. To avoid premature and enhance the diversity of the population, dual external archives and a new elite strategies are proposed. The results of mathematical functions show the effectiveness and efficiency of the proposed method. The promising results of the high frequency inverse problem demonstrate the engineering value of the propose algorithm.

关 键 词：正交边界交叉算法 电磁场逆问题 线性天线阵列 矢量进化算法 优化 

分 类 号：TM155[电气工程—电工理论与新技术]