Clustering analysis using a novel locality-informed grey wolf-inspired clustering approach

Abstract

Grey wolf optimizer (GWO) is known as one of the recent popular metaheuristic algorithms inspired from the social collaboration and team hunting activities of grey wolves in nature. This algorithm benefits from stochastic operators, but it is still prone to stagnation in local optima and premature convergence when solving problems with a large number of variables (e.g., clustering problems). To alleviate this shortcoming, the GWO algorithm is hybridized with the well-known tabu search (TS). To investigate the performance of the proposed hybrid GWO and TS (GWOTS), it is compared with well-regarded metaheuristics on various clustering datasets. The comprehensive experiments and analysis verify that the proposed GWOTS shows an improved performance compared to GWO and can be utilized for clustering applications.