Artificial bee colony algorithm based on hybrid search framework and novel search equations
Sibo He · 2024
The artificial bee colony algorithm (ABC) is a stochastic optimization method based on the intelligent foraging behavior of bees, which contains three core phase: employed bee phase, onlooker bee phase, and scout bee phase, where the employed bee phase focuses on improving the algorithm's exploration, while the onlooker bee phase serves to improve the algorithm's exploitation, and scout bee phase is responsible for randomly searching the search space in order to discover new possible solutions, thereby avoiding the algorithm falling into local optimization. The artificial bee colony algorithm has been shown to be an efficient and effective method for solving optimization problems. However, although the onlooker bee phase improves the exploitation, it is still insufficient and therefore the convergence of artificial bee colony algorithms tends to be slow. To improve this problem, this paper proposes a new ABC (ABC-EDD), in which a depth-first search framework and a new search equation are used to enhance the exploitation of the onlooker bee phase. Meanwhile, in order to better balance the global and local exploration capabilities, we subsequently used a breadth-first search framework and an elite dimensional learning strategy to enhance the global exploration capability in the employed bee phase. The deep-first search framework achieves efficient allocation of computational resources by prioritizing food sources with higher quality and greater potential for improvement. The elite dimension learning strategy utilizes the difference between random individuals and elite individuals in two random dimensions to improve the diversity of the population. The experimental results show that the improved algorithm is significantly better than other algorithms in terms of solution quality, robustness and convergence speed.