Multi-strategy parameter adaptive differential evolution algorithm based on stage grouping

Aiming at the problems that differential evolution algorithm is easy to fall into local optimum and low solution accuracy in the evolution process, a multi-strategy parameter adaptive differential evolution algorithm based on stage grouping (MSDE_SG) is proposed. Firstly, in the first stage, triplet grouping and intra-group strategy allocation boost the algorithm's exploration capabilities, while in the second stage, fitness-based descending order grouping and inter-group strategy allocation enhance the algorithm's exploitation abilities. Then, to ensure rational resource allocation, the second stage introduces a strategy allocation probability mechanism. Finally, the parameter adaptation mechanism dynamically adjusts group parameters to meet the requirements of each stage. Comparative results of the proposed algorithm against seven advanced DE variants on 30 benchmark functions show significant superiority in solution accuracy and convergence speed. This indicates that the algorithm can achieve a more effective balance between exploration and exploitation.