Wheeling Pricing Calculation and Allocation using Deep Reinforcement Learning (DRL) Approach

Wheeling pricing calculation and allocation play a vital role in ensuring equitable and efficient utilization of transmission networks in electricity markets. Traditional methods often suffer from limitations due to simplified models and assumptions, leading to suboptimal outcomes. In this research, we propose a novel approach based on Deep Reinforcement Learning (DRL) to address these challenges. Our objective is to leverage the power of DRL algorithms to learn an optimal wheeling pricing policy that considers dynamic power flows, network congestion, and market conditions. By utilizing historical data and reinforcement learning techniques, our approach aims to maximize system efficiency while ensuring fairness and compliance with regulatory constraints. We present the formulation of the wheeling pricing problem as a reinforcement learning task, defining the state space, action space, and reward function for the DRL agent. Through extensive simulations using realistic power system scenarios, we compare the performance of our DRL-based approach with traditional methods, evaluating metrics such as system efficiency, congestion management, and fairness. We also discuss the computational aspects of implementing DRL for real-time wheeling pricing and highlight practical implementation challenges and regulatory considerations. This research contributes to the field by demonstrating the potential of DRL algorithms by comparing previously well known particle swarm optimization (PSO) based algorithm using IEEE 30 -bus system in improving the accuracy and fairness of wheeling pricing in electricity markets, paving the way for future studies in advanced AI techniques for power system operations and market design.