A nature-inspired hybrid partitional clustering method based on grey wolf optimization and JAYA algorithm

Abstract

This paper presents a hybrid meta-heuristic algorithm that uses the grey wolf optimization (GWO) and the JAYA algorithm for data clustering. The idea is to use the explorative capability of the JAYA algorithm in the exploitative phase of GWO to form compact clusters. Here, instead of using only one best and one worst solution for generating offspring, the three best wolves (alpha, beta and delta) and three worst wolves of the population are used. So, the best and worst wolves assist in moving towards the most feasible solutions and simultaneously it helps to avoid from worst solutions, this enhances the chances of trapping at local optimal solutions. The superiority of the proposed algorithm is compared with five promising algorithms, namely, the sine-cosine (SCA),GWO, JAYA, particle swarm optimization (PSO), and k-means algorithms.The performance of the proposed algorithm is evaluated for 23 benchmark mathematical problems using the Friedman and Nemenyi hypothesis tests. Additionally, the superiority and robustness of our proposed algorithm is tested for 15 data clustering problems by using both Duncan's multiple range test and the Nemenyi hypothesis test.