Artificial Bee Colony Based on Adaptive Search Strategies and Elite Selection Mechanism.

In the field of optimization algorithms, artificial bee colony algorithm (ABC) shows strong search ability on many optimization problems. However, ABC still has a few shortcomings. It exhibits weak exploitation and slow convergence. In the late search stage, the original probability selection for onlooker bees may not work. Due to the above deficiencies, a modified ABC using adaptive search strategies and elite selection mechanism (namely ASESABC) is presented. Firstly, a strategy pool is created using three different search strategies. A tolerance-based strategy selection method is used to select a sound search strategy at each iteration. Then, to choose better solutions for further search, an elite selection means is utilized in the stage of onlooker bees. To examine the capability of ASESABC, 22 classical benchmark functions are tested. Results show ASESABC surpasses five other ABCs according to the quality of solutions.