A Survey of Methods and Architectures for Optimal Scheduling of PV-Battery-Electric Vehicle Loads Using a Scalable Quantum Non-Local Neural Network Approach

Abstract

The transformation of modern power systems into decentralized and sustainable smart grids has increased the complexity of energy management, particularly in the optimal scheduling of photovoltaic (PV) systems, battery energy storage systems (BESS), and electric vehicle (EV) loads. These systems introduce challenges due to renewable intermittency, stochastic demand, and nonlinear interactions, making traditional optimization methods insufficient for large-scale and dynamic environments. This review examines advanced artificial intelligence (AI) and optimization techniques for energy scheduling, with a focus on scalable quantum non-local neural networks (QNLNN). These models combine quantum-inspired computational principles with non-local learning mechanisms to capture global dependencies and enhance optimization performance.

The study explores various methodologies, including deep learning, reinforcement learning, and hybrid metaheuristic approaches, applied across smart grids and microgrids. It also highlights advanced architectures such as graph neural networks and transformer models for handling complex energy interactions. Datasets such as IEEE benchmark systems, smart meter data, and EV charging profiles are analyzed alongside simulation platforms. Findings indicate that QNLNN-based approaches improve scalability, convergence speed, and adaptability. Key challenges include computational complexity and integration issues, while future directions emphasize scalable, intelligent, and sustainable energy management solutions.