Dynamic Artificial Bee Colony Algorithm with Hybrid Initialization Method

Abstract

An improved basic artificial bee colony (ABC) algorithm with a self-adaptive technique, called dynamic ABC, is proposed. The dynamic ABC algorithm first uses a hybrid method combining the good-point-set method with the chaotic maps method to generate the initial population. Then, it applies self-adaptive population size at each generation, meaning that the population increases or decreases depending on some criteria, to enhance global convergence and avoiding local solutions. Experiments are carried out on a range of 10 popular benchmark functions. The results indicate that the dynamic ABC algorithm is superior to the basic ABC algorithm when considering both the speed and quality of the solution obtained.