HABCDE: A hybrid evolutionary algorithm based on artificial bee colony algorithm and differential evolution

Abstract

Artificial bee colony (ABC) algorithm is one of the most recent swarm intelligence based algorithms which has been proven to be competitive with other population based algorithms. However, there is still an insufficiency in ABC regarding its solution search equation, which lacks the guidance of better solutions and much more exchange of information between the old solution and new solution. Inspired by gbest-guided ABC (GABC), a new solution search equation with the direction of better solutions, is introduced and combined with the original one. Moreover, many more dimensions of an old solution are perturbed to enhance the level of information exchange between the two solutions (social learning). And then a modified differential evolution (DE) is also incorporated into the modified ABC in view of the fast convergence speed of DE. Subsequently, a new population catastrophe scheme is introduced in order to further achieve better compromise between the exploration and the exploitation. Based on the above explanation, this paper presents a novel hybrid evolutionary algorithm named hABCDE, which integrates a modified ABC and a modified DE to solve numerical optimization problems. Finally, the experimental results tested on a set of 20 benchmark functions show that the hABCDE algorithm can outperform ABC, DE and a few other state-of-the-art DE variants in most cases.