Spherical search optimizer: a simple yet efficient meta-heuristic approach

In these years, more meta-heuristic approaches have been proposed inspired by nature. However, the search mode has not been researched deeply. In this paper, we find that search style and individual selection mechanism for interaction are the core problems for a meta-heuristic algorithm. In particular, we focus on search style and have studied the principle of basic hypercube search style and basic reduced hypercube search style. Inspired by them, we propose a spherical search style. Furthermore, we design a spherical search optimizer by the spherical search style and tournament selection method. And then, theoretical analysis of it is provided. To validate the performance of the proposed method, we compare our approach against nine state-of-the-art algorithms. The CEC2013, CEC2014, CEC2015 and CEC2017 suites and the data clustering optimization problem in the real world are used. Experimental results and analysis verify that it is a simple yet efficient method to solve continuous optimization problems.