基于多种群竞争差分进化算法的稀布线阵优化  

Synthesis of linear sparse arrays using muli-populationscompetition based differential evolution

作　　者：王莉 王旭健 康凯 田罗庚 WANG Li;WANG Xujian;KANG Kai;TIAN Luogeng(Institute of Operational Support,Rocket Force University of Engineering,Xi’an 710025,China;Test Center of National University of Defense Technology,Xi’an 710106,China)

机构地区：[1]火箭军工程大学作战保障学院,西安710025 [2]国防科技大学试验训练基地,西安710106

出　　处：《兵器装备工程学报》2024年第5期248-258,共11页Journal of Ordnance Equipment Engineering

摘　　要：为进一步降低差分进化算法及其改进算法优化稀布线阵时的峰值旁瓣电平,提出了一种基于多种群竞争的差分进化算法。用多个亚种群代替单一种群,借鉴竞争机制对亚种群进行评估,根据评估得分确定各亚种群产生新个体的数量,分数高的亚种群可以产生更多新个体,体现出高质量亚种群的优势。引入信息交流行为将这种优势扩散到其他亚种群,加快整个种群的收敛速度,提高解的质量。对亚种群评估和信息交流行为的机制分析以及实验验证表明,2种行为结合使用算法性能最佳。3种天线实例仿真实验表明:与对比算法中最优算法相较,所提方法50次实验的平均峰值旁瓣电平分别降低了4.31、1.31、0.41 dB。To reduce peak side lobe level when applying differential evolution to synthesis of linear sparse arrays,this paper proposes muli-populations competition based differential evolution algorithm(MPCDE).First,it generates several subpopulations.Then,these subpopulations are evaluated according to some indexes learning from teaching quality evaluation.The evaluation scores determine the number of new individuals each subpopulation can generate.The higher the evaluation score,the more individuals the subpopulation can generate,reflecting the privilege of good subpopulations.Next,the privilege can spread to other subpopulations through information exchange,accelerating the convergence speed of the whole population.The behavioral analysis of subpopulation evaluation and information exchange demonstrates the algorithm containing the two operations performs best.Simulations of synthesis of linear sparse arrays demonstrate that MPCDE can reduce peak side lobe level of antenna arrays.

关 键 词：阵列天线 稀布线阵 峰值旁瓣电平 差分进化 多种群竞争 

分 类 号：TN820[电子电信—信息与通信工程]