A grey artificial bee colony algorithm

• Published in: Applied soft computing 2017-11, Vol.60, p.1-17
• Main Authors: Xiang, Wan-li, Li, Yin-zhen, Meng, Xue-lei, Zhang, Chun-min, An, Mei-qing
• Format: Article
• Language: English
• Subjects: Artificial bee colony algorithm; Chaotic initialization; Grey relational analysis; Numerical optimization; Opposition-based learning
• ISSN: 1568-4946; 1872-9681
• DOI: 10.1016/j.asoc.2017.06.015

•Grey relational analysis is used to decide the relation of closeness of the employed bee with respect to its neighbor.•The chosen neighbor individual is employed to guide the evolutionary process.•A new perturbation scheme is presented to control the frequency of perturbation of parameters.•Two combinatorial solution search equations are presented to balance the exploration and the exploitation.•A comprehensive experimental study is carried out. Artificial bee colony (ABC) algorithm is a very popular population-based algorithm. Unfortunately, there exists a shortcoming of slow convergence rate, which partly results from random choices of neighbor individuals regarding its solution search equation. A novel scheme for the choice of neighbors is introduced based on grey relational degrees between a current individual and its neighbors to overcome the insufficiency. Then, the chosen neighbor is used to guide the search process. Additionally, inspired by differential evolution, a solution search equation called ABC/rand/2 is employed to balance the previous exploitation and a new perturbation scheme is also employed. What is more, solution search equations using information of the best individual, an opposition-based learning method and a chaotic initialization technique are also integrated into the proposed algorithm called grey artificial bee colony algorithm (GABC for short). Subsequently, the effectiveness and efficiency of GABC are validated on a test suite composed of fifty-seven benchmark functions. Furthermore, it is also compared with a few state-of-the-art algorithms. The related experimental results show the effectiveness and superiority of GABC.