Multi-population artificial bee colony algorithm based on Lagrange interpolation

The artificial bee colony (ABC) algorithm, inspired by bees’ foraging, solves optimization problems well but converges slowly due to its search equation’s bias towards exploration. This paper proposes the MPABCLI based on Lagrange interpolation. It divides individuals into sub-populations by fitness, uses different variation strategies, limits candidate parents’ range, and adds Lagrangian interpolation local search (LSLI) to balance exploration and exploitation. Tests on CEC2014’s 30 functions show that for 10D, 30D and 50D problems, MPABCLI outperforms other ABC algorithms, with better performance and accuracy for global numerical optimization.