Centralized and Decentralized Pricing Strategies for Optimal Scheduling of Electric Vehicles

Abstract

Three novel pricing schemes based on the centralized and decentralized architectures for day-ahead optimal scheduling of Electric Vehicles (EVs) are proposed in this paper. A novel <inline-formula> <tex-math notation="LaTeX">$\psi $ </tex-math></inline-formula>-Iterative Pricing Scheme (<inline-formula> <tex-math notation="LaTeX">$\psi $ </tex-math></inline-formula>-IPS) is proposed for the decentralized model. In contrast, the Bi-level Pricing Scheme (BPS) and Multi-objective Pricing Scheme (MPS) are proposed for the centralized model. The issues of load valley-filling and rebound peak occurrence are addressed effectively while fulfilling the objectives of multiple players, i.e., the EV owners, the aggregator, and the Distribution System Operator (DSO). The aggregated load, i.e., predicted non-EV load and EV charging requests, served by the distribution transformer, is used for the scheduling process. Solar PV integrated residential load (non-EV load) is predicted using the Random Forest technique. The input load and weather data are utilized after data cleaning and features selection. The <inline-formula> <tex-math notation="LaTeX">$\psi $ </tex-math></inline-formula>-IPS and BPS are implemented using game-theoretic approaches, while MPS utilizes a multi-objective formulation. The three schemes based upon the two architectures, i.e., centralized and decentralized, are compared, and the key benefits of the schemes are highlighted.