An Efficient Reinforcement Learning based Charging Data Delivery Scheme in VANET-Enhanced Smart Grid

Guangyu Li (1) , Chen Gong (1) , Lin Zhao (1) , Jinsong Wu (2) , Lila Boukhatem (3)

1 NJUST - Nanjing University of Science and Technology
2 Guilin University of Electronic Technology
3 LRI - Laboratoire de Recherche en Informatique

Insufficient and fragile delivery of enormous charging data imposes great challenges on the productive operations of smart grid systems. In this paper, we propose an efficient charging information transmission strategy (ECTS) for spatiotemporal coordinated vehicle-to-vehicle (V2V) charging services. Specifically, based on the concepts of mobile edge computing (MEC) and hybrid vehicular ad hoc networks (VANETs), an effective and scalable communication framework is firstly designed to decrease communication costs. In addition, by means of the derived model of wireless connectivity probability, an effective reinforcement learning based routing algorithm is proposed to adaptively select the optimal charging data delivery path in dynamic large-scale VANET environments. Finally, a series of simulation results are presented to demonstrate the effectiveness and the feasibility of our proposed ECTS scheme.

Domaines: Informatique [cs]

Guangyu Li, Chen Gong, Lin Zhao, Jinsong Wu, Lila Boukhatem. An Efficient Reinforcement Learning based Charging Data Delivery Scheme in VANET-Enhanced Smart Grid. 2020 IEEE International Conference on Big Data and Smart Computing (BigComp), Feb 2020, Busan, North Korea. pp.263-270, ⟨10.1109/BigComp48618.2020.00-64⟩. ⟨hal-03001823⟩