.png){kind=icon}
.png){kind=icon}
Return
A robust method based on reinforcement learning and differential evolution for the optimal photovoltaic parameter extraction
.png){kind=icon}
PRE
AI
DOI：10.1016/j.asoc.2023.110916.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
It is crucial to identify the optimal parameters of Photovoltaic (PV) models with the purpose of evaluating, controlling, and improving PV systems. Lots of optimization algorithms have been developed to solve the problem and achieve better results. But, the performance of most algorithms is greatly influenced by their parameters. A Reinforcement Learning-based Adaptive Parameter and Mutation operator Differential Evolution algorithm (RLAPMDE) is developed, in which the parameters of algorithms are randomly set in [0,1], and the mutation operator is adaptively adjusted between Rand/1 and Best/1. Rand/1 is mainly responsible for exploration, while Best/1 is to implement exploitation. Exploration and exploitation are well balanced. RL technique is used to realize the process through interaction with the environment. The RLAPMDE is employed to identify the optimal parameters of PV. The dataset used in experiments is Single Diode Model (SDM), Double Diode Model (DDM) and PV module. The results of the best RMSE obtained by RLAPMDE are 9.8602E-4, 9.8248E-4, 2.4251E3, 1.7298E-3 and 1.6601E-2, which are the minimum when compared with the conventional DE, five DE variants, six latest meta-heuristic algorithms and some reported results. The proposed PLAPMDE algorithm is one of the alternative algorithms when estimating the parameters of PV systems.
Keywords:
Reinforcement learning
Differential evolution
Parameter extraction
PV systems
Journal
IF:
6.6
Papers:
1.4W
Citations:
4.8W
Organization
No organization information available