An adaptive artificial bee colony algorithm for global optimization

Artificial bee colony algorithm (ABC) is a recently introduced swarm based meta-heuristic algorithm. ABC mimics the foraging behavior of honey bee swarms. Original ABC algorithm is known to have a poor exploitation performance. To remedy this problem, this paper proposes an adaptive artificial bee colony algorithm (AABC), which employs six different search rules that have been successfully used in the literature. Therefore, the AABC benefits from the use of different search and information sharing techniques within an overall search process. A probabilistic selection is applied to determine the search rule to be used in generating a candidate solution. The probability of selecting a given search rule is further updated according to its prior performance using the roulette wheel technique. Moreover, a memory length is introduced corresponding to the maximum number of moves to reset selection probabilities. Experiments are conducted using well-known benchmark problems with varying dimensionality to compare AABC with other efficient ABC variants. Computational results reveal that the proposed AABC outperforms other novel ABC variants.