基于混合灰狼算法的阵元失效校正方法  

Element failure correction method based on hybrid GWO algorithm

作　　者：祁峥东 解效白 孔玥 张羽 QI Zhengdong;XIE Xiaobai;KONG Yue;ZHANG Yu(School of Information Engineering and Artificial Intelligence,Nanjing Xiaozhuang University,Nanjing 211171;The Eighth Research Academy of CSSC,Nanjing 211153;Topxgun(Nanjing)Robotics Co.,Ltd.,Nanjing 211100;Shanghai Topxgun Robotics Co.,Ltd.,Shanghai 201306)

机构地区：[1]南京晓庄学院信息工程人工智能学院,南京211171 [2]中国船舶集团有限公司第八研究院,南京211153 [3]拓攻(南京)机器人有限公司,南京211100 [4]上海拓攻机器人有限公司,上海201306

出　　处：《雷达与对抗》2024年第3期31-36,共6页Radar & ECM

摘　　要：针对经典灰狼算法在处理复杂非线性优化问题时存在搜索不均衡、易陷入局部最优而出现早熟、寻优精度低等缺点,提出一种基于非线性双收敛因子策略与局部跳出机制相结合的混合灰狼算法,用于解决阵元失效校正问题。首先,依据平均适应度值将种群分为侦察狼和捕猎狼,两类狼依据不同的非线性递减收敛因子寻优来平衡算法的全局勘探和局部挖掘能力;其次,引入一种基于Sine混沌函数的跳出机制,提高了头狼个体的局部抗停滞能力,提升了算法的种群多样性。算法在目标函数中设计增加公共激励约束条件来降低天线系统的维护成本、提高系统的稳定性。对比现有基于群智能算法的校正技术,在更为苛刻的约束条件下,算法综合所得公共激励占比由0分别上升至10.7%,且校正前后方向图的辐射特性有一定提升,证明了所述算法的有效性。In view of the shortcomings of the classic grey wolf optimizer(GWO)algorithm in dealing with complex nonlinear optimization problems,such as the unbalanced exploration,which is prone to falling into local optimum,leading to prematrue convergence and low optimization accuracy,a hybrid GWO algorithm is proposed by utilizing nonlinear dual-convergence factor strategy and local optimum escape mechanism for element failure correction.Firstly,the wolve pack can be divided into hunting wolves and reconnaissance wolves according to the average fitness value,and different decreasing convergence factors are used by them to optimize and balance the exploitation and exploration.In addition,an escape mechanism is introduced based on the Sine chaos function,enhancing the local anti-stagnation capability of the alpha wold individual and population diversity of the algorithm.The maintenance cost of the antenna system is reduced and the system stability is improved by incorporating common excitation constraints into the objective function.Compared with the existing correction technologies based on the swarm intelligence algorithms,the percentage of common excitation obtained by the algorithm increases from 0 to 10.7%under more stringent constraint conditions,and the radiation characteristics of the patterns before and after correction are improved,proving the effectiveness of the algorithm.

关 键 词：阵列信号处理 灰狼优化算法 阵元失效校正 方向图综合 双收敛因子 

分 类 号：TN957.2[电子电信—信号与信息处理]