Optimizing Electric Vehicle Charging in Smart Grids: An IoT‐Based Approach With TARNN‐NO

ABSTRACT The integration of Internet of Things (IoT) technology has significantly simplified the scheduling of electric vehicles (EVs) in everyday life, leading to a transformative impact on the electric load curve. This manuscript presents a hybrid method for optimizing EV charging in smart grids. The proposed hybrid approach is the joint execution of both the Narwhal Optimizer (NO) and Time Aware Recurrent Neural Networks (TARNN). Hence, it is named as TARNN‐NO approach. The primary objective is to develop a radial EV charging system to reduce overall energy costs and maximize self‐consumption. The TARNN is used to predict the EVs' behaviors to make charging decisions. The NO algorithm is used to optimize the power flow of the Electric Vehicle Charging Station (EVCS). The proposed method is assessed utilizing the MATLAB platform and contrasted with other existing techniques, including differential evolution optimization (DEO), reinforcement learning (RL), and deep deterministic policy gradient (DDPG). Experimental outcomes establish that the TARNN‐NO approach achieves the highest self‐consumption rate of 0.926 and a maximum cost reduction of 24.23%, highlighting its effectiveness for intelligent EV charging management in smart grids.