Hybrid artificial chemical reaction optimization algorithm for cluster analysis

Abstract

Heuristic algorithms have significant contribution in the clustering field. In present work, a hybrid version of the artificial chemical reaction optimization algorithm (HACRO) is proposed to optimize clustering problems. As exploration and exploitation are two major aspects that require balanced coordination among algorithmic steps. The artificial chemical reaction suffers from slower convergence speed due to its poor exploitation mechanism. Moreover, it requires more execution time. Henceforth, to enhance the convergence speed and to make balance among algorithmic space a hybrid version of ACRO is developed. In present work, the artificial chemical reaction optimization algorithm is incorporated with crossover and mutation operator of genetic algorithm. Further, the efficiency of the HACRO algorithm is examined on seven benchmark datasets and collated with ACO, PSO, K-means, GA, ICSO and ACRO clustering algorithms. The present investigation indicated that the proposed algorithm works efficiently in clustering field.