A hybrid grey Wolf–arithmetic optimization algorithm for high-accuracy parameter identification of photovoltaic models

Abstract

Introduction In recent years, the hybridization of metaheuristic algorithms has been widely recognized as an effective strategy for overcoming the limitations of conventional optimization approaches, particularly their slow convergence behavior and susceptibility to premature convergence in complex search spaces. Methods A novel hybridization of the Grey Wolf Optimizer (GWO) and Arithmetic Optimization Algorithm (AOA), called H-GWOAOA, is proposed. The method selectively incorporates the arithmetic operators of AOA into GWO’s hierarchical leadership model by embedding AOA operators into the α-agents’ updating phase to improve the balance between exploration and exploitation without increasing algorithmic complexity. Discussion The effectiveness of the proposed approach is experimentally confirmed through testing on 23 benchmark functions, achieving values of 0.1887 (F1) and 0.0331 (F2) and demonstrating improved accuracy, stability, and robustness compared with GWO, AOA, PSOGWO, SSA, and SCA. Its generalization ability is further evaluated on four biomedical datasets (XOR, Iris, Breast Cancer, and Heart), achieving accuracy rates of 100% for XOR and 99.14% for Breast Cancer. Moreover, H-GWOAOA is applied to parameter estimation for the Single-Diode Model (SDM), Double-Diode Model (DDM), Triple-Diode Model (TDM), and Photovoltaic Model (PMM), achieving a minimum Root Mean Square Error of 5.51 × 10 −4 for SDM. Discussion The results demonstrate that the proposed H-GWOAOA provides competitive optimization accuracy, convergence stability, and robustness while preserving the computational efficiency of the original GWO framework.