Multi-population and dynamic-iterative cuckoo search algorithm for linear antenna array synthesis

Nature inspired algorithms have served as the backbone of modern computing technology and over the past three decades, the field has grown enormously. CS is a recent addition to optimization computing and suffers from problem of local optima stagnation and poor exploration. This paper presents a new version of CS namely extended CS (ECS) by employing the concept of multi-population adaptation, adaptive switching, dynamic iterative search, and GWO inspired global search phase, for solving benchmark problems and real-world design problem of linear antenna array (LAA) optimization. The algorithm aims at providing a better framework for optimization problems by keeping the original structure of CS intact. For performance evaluation, the algorithm has been firstly applied on two highly challenging datasets namely CEC 2015 and CEC 2017 benchmark problems and performance evaluation is done in comparison with other algorithms. Further to test the ECS algorithm on real world application, it is applied for synthesis of uniform and non-uniform LAA. Experimentally, the ECS algorithm is found to provide better performance in comparison to basic CS and others state-of-the-art algorithms. Statistical tests and radiation patterns further validate the results.