基于引导素更新和扩散机制的人工蜂群算法

Artificial Bee Colony Algorithm Based on Inductive Pheromone Updating and Diffusion

人工蜂群算法是一种新型的搜索算法,其机理是通过模拟蜂群采蜜过程中体现出的智能行为来实现对问题的求解.在现有的蜂群算法中,蜂群间的信息交流仅使用单一的行为通信(跳舞),蜂群间的协作存在明显不足,影响了蜂群算法的求解性能.根据真实蜜蜂多模式传递信息的客观事实,通过引入基于引导素的化学通信方式,提出一种新的更忠实反映蜂群信息传递的蜂群算法,并应用于多维背包问题(MKP)的求解.新算法将行为通信和化学通信相融合,利用引导素的更新和扩散机制使蜂群能够更好地进行协作.MKP的仿真实验结果表明新算法优于传统的ABC算法.与其他一些元启发式搜索算法的比较同样显示了新算法的有效性.

Artificial bee colony （ABC） algorithm is a novel search algorithm which simulates the intelligent foraging behavior of honeybee swarm to solve the practical problems. However, there is only a behavior communication way （dancing） in the current ABC algorithm, which results in the lack and lag of collaboration among bees and influences the solving performance of ABC algorithm. Inspired by the objective fact of transinformation among real bees, a new ABC algorithm is proposed by introducing a chemical communication way based on inductive pheromone and applied to solve multidimensional knapsack problems （MKP）, which is more faithful to the transmission information of real bee colony system. With the combination of the behavior communication way and the chemical communication way, the new algorithm makes the honeybees cooperate with each other better by the scheme of inductive pheromone updating and diffusion. A number of simulation experiments and comparisons on benchmark datasets of MKP demonstrate that the performance of the new algorithm is superior over the original ABC algorithm. The performances of the new algorithm have also been compared with some typical meta-heuristic search algorithms, and the computational results show that the new ABC algorithm obtains better quality solutions than all the other approaches.