Enhanced artificial bee colony algorithm through differential evolution

•When producing the initial population, the chaotic opposition-based population initialization method is employed to enhance the global convergence.•A learning strategy is developed with an attempt to use more prior information of previous search experience.•A hybrid approach that combines DE with gbest-guided ABC, is designed to improve the performance of ABC.•The experiment results demonstrate the good performance of the proposed algorithm.

Artificial bee colony algorithm (ABC) is a relatively new optimization algorithm. However, ABC does well in exploration but badly in exploitation. One possible way to improve the exploitation ability of the algorithm is to combine ABC with other operations. Differential evolution (DE) can be considered as a good choice for this purpose. Based on this consideration, we propose a new algorithm, i.e. DGABC, which combines DE with gbest-guided ABC (GABC) by an evaluation strategy with an attempt to utilize more prior information of the previous search experience to speed up the convergence. In addition, to improve the global convergence, when producing the initial population, a chaotic opposition-based population initialization method is employed. The comparison results on a set of 27 benchmark functions demonstrate that the proposed method has better performance than the other algorithms.