Best neighbor-guided artificial bee colony algorithm for continuous optimization problems

Abstract

As a relatively recent invented swarm intelligence algorithm, artificial bee colony (ABC) becomes popular and is powerful for solving the tough continuous optimization problems. However, the weak exploitation has greatly affected the performance of basic ABC algorithm. Meanwhile, keeping a proper balance between the exploration and exploitation is critical work. To tackle these problems, this paper introduces a best neighbor-guided ABC algorithm, named NABC. In NABC, the best neighbor-guided solution search strategy is proposed to equilibrate the exploration and exploitation of new algorithm during the search process. Moreover, the global neighbor search operator has displaced the original random method in the scout bee phase aiming to preserve the search experiences. The experimental studies have been tested on a set of widely used benchmark functions (including the CEC 2013 shifted and rotated problems) and one real-world application problem (the software defect prediction). Experimental results and comparison with the state-of-the-art ABC variants indicate that NABC is very competitive and outperforms the other algorithms.