Photovoltaic power prediction plays a crucial role in renewable energy management

as accurate forecasting significantly enhances grid stability and economic efficiency.This study proposes a hybrid photovoltaic power prediction method that integrates Genetic Algorithm (GA)-optimized Variational Mode Decomposition (VMD) with a Convolutional Neural Network-Gated Recurrent Memory Network (CNN-GRU).First

the genetic algorithm optimizes the key parameters of VMD to effectively decompose photovoltaic power sequences.Subsequently

the CNN extracts multi-scale local features

while the GRU captures long-term dependencies within the decomposed time series

thereby constructing a deep learning prediction model.Finally

the predictions from multiple components are integrated to generate the final output.Experimental results demonstrate that the proposed method outperforms traditional prediction approaches in accuracy

providing a novel and effective solution for photovoltaic power forecasting.