Optimized Electric Vehicle Charging Scheduling Using a Cnn-Lstm Hybrid Deep Learning Model

The rapid adoption of electric vehicles (EVs) has resulted in considerable issues in the management of charging stations, especially in forecasting charging demand and optimising energy allocation. Conventional machine learning techniques fail to effectively capture the temporal and spatial connections in charging data, leading to suboptimal scheduling and grid instability. A CNN-LSTM hybrid deep learning model is suggested to enhance predicting accuracy by integrating convolutional feature extraction with sequential pattern learning. The model is trained using real charging datasets, attaining a Mean Absolute Error (MAE) of 2.13, a Root Mean Square Error (RMSE) of 3.25, and an R-squared (<tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{R}^{\mathbf{2}}$</tex>) score of 0.98, overcoming baseline models including Linear Regression, Random Forest, and independent LSTM. The suggested approach improves real-time energy distribution, minimises peak demand variations, and reduces grid overload. This research presents a stable, scalable, and adaptive strategy for optimising electric vehicle charging infrastructure, higher than current scheduling methodologies and facilitating the advancement of smart energy grids.