An Improved Artificial Bee Colony Algorithm with Diversity Control

Artificial Bee Colony algorithm is a swarm intelligence based technique that has shown a satisfactory performance for complex optimization problems even when compared with other well known techniques. Although, as a fact, the literature describes some flaws in the algorithm, such as poor convergence speed and exploitation capabilities. Different mechanisms can be explored for the construct of a more efficient algorithm. This work proposes a modification on the search equation altogether with a more careful observation of the swarm diversity to improve the convergence and, at the same time, intensify the exploration of the algorithm at more suitable moments. Experimental tests are carried out with six classical optimization functions in order to show the effectiveness of the proposed approach and statistical analysis of the results support our claim about the ABC model proposed.