A modified crow search algorithm with niching technique for numerical optimization

Abstract

This paper proposes a modified crow search algorithm with local search and niching technique. The primitive crow search algorithm is a newly developed population-based algorithm which gained attention from the researchers of many fields as it needs only one parameter to be tuned. Despite its easy implementation, crow search algorithm has weakness to find global optima and suffers from slow convergence rate in multi-modal optimization problems. The search agent of the primitive crow search algorithm does not always follow the best solution obtained so far. Another disadvantage is that the search agents updates its location randomly. In order to enhance its searching capacity, a global search operator is introduced. Also, The proposed method modifies the search techniques and incorporates niching method to increase exploration capacity. The proposed technique is tested on 23 benchmark functions. The results of the proposed method demonstrated faster convergence rate and better solution in most cases when compared with the standard crow search algorithm.