Electric Vehicle Cluster and Scheduling Strategy Based on Dynamic Game

Abstract

Aiming at the problem of excessive dimensionality of massive electric vehicles participating in grid peak shaving scheduling, a scheduling strategy using electric vehicle agents (EVA) as an intermediary and electric vehicles participating in peak shaving in a cluster manner is proposed. A leader-follower game model based on Stackelberg theory is constructed. The upper layer takes the peak shaving demand and peak shaving cost of distribution system operator (DSO) as the optimization objectives, and uses an improved multi-objective particle swarm optimization algorithm to obtain the game strategy set of DSO. The lower layer constructs a calculation model of vehicle owners' willingness to participate in the network based on fuzzy neural networks, calculates the willingness value of electric vehicles to participate in peak shaving based on the electric vehicle state of charge, the adjustable duration of vehicle charging and discharging, and the lower layer optimal electricity price. The vehicles are clustered based on the willingness value of vehicle owners to participate, and the EVA returns the cluster electricity to the DSO to readjust the optimal strategy. Simulation results of the dispatching strategy show that the peak-to-valley ratio decreased by 13.22%,which indicates that the EVA can effectively reduce its own operating costs while meeting the peak shaving requirements of the power grid, and vehicle owners can obtain benefits from participating in the peak shaving process.

Keywords

• power grid peak shaving
• fuzzy neural network
• stackelberg game
• optimized scheduling
• electric vehicle agent(eva)