Journal of Bionic Engineering (2025) 22:1459–1483https://doi.org/10.1007/s42235-025-00675-y

A Multi-group Meta-heuristic Optimization with Dynamic Population Partition and Hybrid Strategies: Algorithm and Applications

Dongshuai Niu1  · Guangwen Yi1  · Long Chen2  · Zhenzhou Tang1

1 Wenzhou Key Laboratory for Intelligent Networking, Wenzhou University, Wenzhou 325035, China 

2 Key Laboratory of Intelligent Education Technology and Application of Zhejiang Province, Zhejiang Normal University, Jinhua 321000, China

Abstract 

To further improve upon the deficiencies of traditional algorithms in terms of population diversity, convergence accuracy, and speed, this paper introduces a Dynamic Multi-Population Hybrid Metaheuristic Algorithm (DHA). DHA dynamically categorizes the population into Elite, Follower, and Explorer subgroups, applying specific strategies: a novel dimensionwise Gaussian mutation combined with the Sine Cosine Algorithm (SCA) for the Elite, a randomized spiral search for the Explorer, and Lévy flight for the Follower. Rigorous testing on benchmark sets like CEC2005, CEC2017, and CEC2019, alongside practical application in Service Function Chain (SFC) mapping, underscores DHA’s superior performance and applicability. 

Keywords Dimension-wise Gaussian mutation · Random spiral search · SFC mapping