Improved Artificial Bee Colony Algorithm with Observed Subgroups for Optimization Problems

Artificial bee colony (ABC) algorithm is optimization technique that works well on complex optimization problems, but it’s potential is constrained by the shortcomings that insufficient local search and slow convergence. To alleviate these challenges, an improved ABC variant with observed subgroups (OSABC) is proposed. In this study, each food source has an observed subgroup that is determined by calculating its Euclidean distance from the other. And the subgroups’ size adaptively changes according to the ranking. Then, the new update equation is constructed by the food source from the subgroup. Additionally, to mitigate the scenario in which ABC faces strong selection pressure later on, we integrate a ranking-based selection mechanism with the fitness-based selection probability to design a dynamically adjusted selection probability. The numerical experimental results of OSABC with excellent ABC variants on optimization problems and their shifted versions show that OSABC has better solution accuracy and faster convergence rate. Meanwhile, OSABC’s practical applicability has verified on the wireless sensor network (WSN) coverage optimization problem. © (2024), (International Association of Engineers). All Rights Reserved.