Real-time coordinated management of PHEVs at residential level via MDPs and game theory

This paper proposes a novel Markov decision process (MDP) with dynamic transition probabilities for the stochastic modeling of the charging process of plug-in hybrid electric vehicles (PHEVs). In the proposed dynamic MDP, PHEVs can be controlled in such a way that the effectiveness of the control strategy is maintained in the presence of uncertainties such as early departure events. Then a game theory based decentralized system is formulated to coordinate the PHEVs fleet in a residential network. The authors also proposed a decentralized coordinated optimization (DCO) algorithm to solve the formulated Nash game. Various simulations are carried out to verify the effectiveness of the proposed DCO approach. The results show that the DCO approach is robust in the face of uncertainties and is effective in enhancing both power quality and economic profits.