Self-adaptive position update in artificial bee colony

Abstract

Artificial bee colony (ABC), which is one of the leading swarm intelligence based algorithm, dominates other optimization algorithms in some fields but, it also has the drawbacks like premature convergence and slow convergence in the later stages due to unbalanced exploration and exploitation abilities. In this paper, we propose a novel variant of ABC, namely Self-adaptive Position update in ABC (SPABC), in which three position update strategies are incorporated in employed bee phase based on the fitness of the solutions. Each employed bee checks its fitness and accordingly adopts one of the position update strategies of standard ABC, Gbest guided ABC (GABC), and modified ABC (MABC). In this way, ABC with a set of solution update strategies of different characteristics can improve the quality of newly generated solutions and hence can improve the convergence speed of ABC. During solution generations, the suitable position update strategy is self-adapted according to the fitness of the solution. The performance of the SPABC is reported on the set of 15 real parameter benchmark test problems and is compared with standard ABC and its recent variants, namely BSFABC, GABC, and MABC.