Exploration–exploitation balance in Artificial Bee Colony algorithm: a critical analysis

Abstract

Artificial Bee Colony (ABC) algorithm is a popular metaheuristic due to its simplicity yet a stronger search mechanism. However, some researchers have reported that ABC algorithm lays more emphasis on exploration in comparison with exploitation, its performance also deteriorates gradually as the dimensions of the problems increase and the algorithm may occasionally stop proceeding towards the global optimum. Hence, the algorithm runs the risk of missing out on true global optima. This study critically analyses the functional behaviour of ABC algorithm in the context of above reports and finds that the scout bee operator may turn redundant while dealing with high dimensional problems. Thus, in contrast to the popular view, the study suggests that the ABC algorithm may be poor in exploration ability too for high-dimensional problems. Further, the study offers an explanation for the above-reported observations by other researchers. The findings of the study may be quite useful for designing better performing variants of ABC algorithm.