Customized Rebate Pricing Mechanism for Virtual Power Plants Using a Hierarchical Game and Reinforcement Learning Approach
Wen Chen, Jing Qiu, Junhua Zhao, Qingmian Chai, Zhao Yang Dong
Abstract
In the transition to a two-sided electricity market, energy users are turning into prosumers who own the flexible distributed energy resources (DERs) and have the potential to provide services to the power system. Virtual power plants (VPPs) aggregate DERs to join the electricity market and respond to system signals. It is urgent to develop a new pricing mechanism for VPPs to allocate the payoff from the electricity market to prosumers. This paper proposes a customized rebate package pricing mechanism for a VPP retailer to reward prosumers for supporting the power system. The retailer’s pricing strategies are determined based on a Stackelberg game, considering the heterogeneous prosumers’ dynamic selecting process based on an evolutionary game. The extended replicator dynamics is proposed to take the future payoff into account and guarantee the evolutionary equilibrium. Moreover, a new reinforcement learning algorithm based on the Cross learning model is developed to solve the evolutionary game with less computational effort. The simulation results verify the effectiveness of the proposed customized rebate package pricing mechanism, which can efficiently reward prosumers’ flexible resources in supporting the system while maximizing the retailer’s utility to achieve a win-win outcome.