Multi-Agent Gaming in Distribution Networks with Virtual Power Plants

This thesis builds a unified game framework around the distribution network scheduling problem with virtual power plants, considering the interaction among three types of subjects, namely, distribution network operators (DSOs), virtual power plants (VPPs), and demand response (DR): a master-slave game between DSOs and VPPs, which guides them to submit power plans or real-time adjustments; a master-slave game between VPPs and DRs, which incentivises load-side load response through the price mechanism; and a non-cooperative game among multiple VPPs, which competes for market gains. response; and multiple VPPs form a non-cooperative game to compete for market revenues. This hybrid game structure reflects the coordination of interests and strategies of multiple subjects in the previous day, which effectively improves the economy of dispatching and the flexibility of system operation.