A novel method is proposed for grid-connected microgrid system with high penetration photovoltaic(PV)power that integrates large scale PV power and electric vehicles(EVs)efficiently. An optimization model is proposed to increase the self-consumption of PV power by smart charging of EVs and vehicle-to-grid technology. Without relying on the optimization of central control

a virtual cost theory is developed combined with EVs charging

discharging power

the system load and PV power output. On this basis

a non-cooperative charging and discharging(NCD)game model for EVs is proposed

and the consistency between the Nash equilibrium of the proposed model and the centralized optimization solution is also obtained. Hence the centralized optimization with several disadvantages is replaced by a decentralized and autonomous one. Taking a workspace grid-connected PV-battery microgrid as example

the practicality and effectiveness of the proposed optimization model for EVs scheduling are proved by the improved utilization of PV power and equivalent annual cost. This work is supported by National Natural Science Foundation of China(No. 51407160)and Zhejiang Provincial Natural Science Foundation of China(No. LY16E070005).