In order to overcome some shortcomings including the premature convergence and slow convergence speed at later evolution stage of conventional cuckoo search (CS) algorithm for the sidelobe suppressions of antenna arrays, an improved CS with reverse learning and invasive weed operators (ICSRLIWO) is proposed. First, ICSRLIWO algorithm generates reverse populations through opposition‐based learning method, then selects better individuals in the mixed population consists of the original and reversed population to form a high‐quality initial population. Second, ICSRLIWO introduces the reproducing, space diffusion and survival competition operators of invasive weed optimization to improve the population diversity and global search ability of conventional CS algorithm. Simulations are conducted based on 30 test functions in the CEC 2014 test suit to verify the performance of the proposed approach, and the results show that ICSRLIWO has higher solution accuracy and faster convergence speed than other comparison algorithms. In addition, ICSRLIWO also has advantages in solving the array antenna beam pattern optimization problems.

中文翻译：

---

具有反向学习和侵入性杂草算子的改进的布谷鸟搜索，可抑制天线阵列的旁瓣电平

为了克服传统布谷鸟搜索算法在天线阵旁瓣抑制方面的早期收敛和较慢收敛速度等缺点，提出了一种具有逆向学习和入侵杂草算子（ICSRLIWO）的改进CS。首先，ICSRLIWO算法通过基于对立面的学习方法生成反向种群，然后从原始种群和反向种群组成的混合种群中选择更好的个体，以形成高质量的初始种群。其次，ICSRLIWO引入了入侵杂草优化的繁殖，空间扩散和生存竞争算子，以提高常规CS算法的种群多样性和全局搜索能力。在CEC 2014测试套件中基于30个测试功能进行了仿真，以验证所提出方法的性能，结果表明ICSRLIWO具有比其他比较算法更高的求解精度和更快的收敛速度。此外，ICSRLIWO在解决阵列天线波束方向图优化问题方面也具有优势。