Artificial Bee Colony Algorithm with Hierarchical Groups for Global Numerical Optimization

Abstract

Artificial Bee Colony (ABC) algorithm is a relatively new swarm-based optimization algorithm, which has been shown to be better than or at least competitive to other evolutionary algorithms (EAs). Since ABC generally performs well in exploration but poorly in exploitation, ABC often shows a slow convergence. In order to address this issue and improve its performance, in this paper, we present a novel artificial bee colony algorithm with hierarchical groups, named HGABC. In employed bee phase of HGABC, the population is divided into three groups based on the fitness values of the food source positions, and three solution search strategies with different characteristics are correspondingly employed by different groups. Moreover, in onlooker bee phase, onlooker bees conduct exploitation in the most promising area of search space, instead of around some good solutions. In order to demonstrate the performance of HGABC, we compare HGABC with four other state-of-the-art ABC variants on 22 benchmark functions with 30D. The experimental results show that HGABC is better than other competitors in terms of solution accuracy and convergence rate.