An Enhanced Artificial Bee Colony Algorithm Based on Elimination History and Elite Correction

Artificial bee colony (ABC) algorithm is often challenged by slow convergence, poor accuracy, and premature convergence in handling complex medium-scale optimization problems, due to its biased search equation and the high assimilation rate of bees within the colony. To trade off the ABC for global exploration and local exploitation of complex problem landscapes, this paper proposes an enhanced ABC based on elimination history and elite correction (HeCABC). Given the bias effects of the superior solutions and the historical inferior solutions eliminated on the search behavior of ABC, HeCABC separately formulates an exploration equation oriented by the historically eliminated inferior solutions and an exploitation equation upon multi-elite information fusion for employed bees and onlook bees, to regulate their exploration and exploitation intensity of the solution space. Meanwhile, HeCABC couples an elite correction strategy for fine-tuning the quality of the elites based on the update signal of these elites within the colony. HeCABC is experimented on various complex CEC 2014 test functions of 30 dimensions. The experimental results showcase its superior performance over five state-of-the-art ABC variants and two advanced swarm optimizers.