Modeling and simulation of improved artificial bee colony algorithm with data-driven optimization

To balance the exploration and exploitation and to enhance the convergence rate of an artificial bee colony (ABC) algorithm, the driving force of using additional data during searching process is studied in this paper, and an improved ABC algorithm with data-driven optimization (DDABC) is proposed. First, to speed up convergence rate, the searching process is driven by directional guiding data. Therefore, a bee colony would learn from the directional guiding data, instead of picking up a random direction. Second, to enhance the exploitation capability of the onlooker bees, the searching process is driven by local data of onlooker bees. Every onlooker bee would search independently for multiple times to generate local data applied into optimization. Comparisons are made with a number of other ABC-based and nature-inspired algorithms. The results show that the proposed DDABC achieves improvements in both exploitation capability and convergence rates.