A Three-stage Incentive Scheme for Integrated Energy-Traffic Systems Using Deep Q-Network Tianyu Yang, Qinglai Guo, Hongbin Sun

Abstract

This paper solves the problem of dynamic pricing strategy in an urban integrated energy-traffic system. A three-stage incentive scheme has been introduced to enhance operational profits and mitigate the impact of system uncertainties. Pricing strategies in the three stages, namely day-ahead equilibrium references, hour-level prices, and surge prices, are used to incentivize human users. The proposed framework could yield a set of charging and service prices for IETS, which could improve in overall revenue and security.

Keywords Integrated energy-traffic system, dynamic pricing, reinforcement learning, Deep Q-Network