A New Swarm-Based Simulated Annealing Hyper-Heuristic Algorithm for Clustering Problem

Abstract

Hyper-heuristic is a new generation of heuristic algorithms which include a set of heuristics that could generate a set of heuristics or automate the process of selecting a batch of heuristics to address different optimization problems. Most of the heuristic and meta-heuristic algorithms work in solution space and are highly problem-dependent and they are limited to use certain number of heuristic models only. Hyper-heuristic consists of two different sections, namely high-level heuristic and low level heuristics; to deal with optimization problems and work in heuristic spaces instead of solution spaces in selecting appropriate heuristics. In this article, we propose a new novel swarm-based simulated annealing hyper-heuristic model to solve clustering problem for analyzing data. The performance of proposed model has been evaluated by several benchmark and real data sets and compared with its state of the art in the literature. The experimental results show that the proposed algorithm provide better results than the alternative methods.