@article{SUN2026, 
author = {Rongfu SUN and Tianbo ZHU and Kangyang YU and Yueyao ZHOU and Qinzhe LIU and Hongyang LI and Xiaohan LI and Jingrong GUO and Zesen WANG and Yunpeng XIAO},
title = {Bi-level Stackelberg Game Model for Distributed Photovoltaic Aggregators Participating in Day-Ahead Energy and Reserve Joint Market Trading},
year = {2026},
journal = {Distributed Energy},
volume = {11},
number = {5},
pages = {98-108},
keywords = {distributed photovoltaic aggregato, electricity market, Stackelberg game, discrete control, data-driven bilevel reconstruction algorithm},
url = {https://www.sciopen.com/article/10.16513/j.2096-2185.DE.26110039},
doi = {10.16513/j.2096-2185.DE.26110039},
abstract = {With the acceleration of new power system construction, distributed photovoltaic aggregators are able to participate in the day-ahead energy and reserve joint market trading and obtain profits. However, due to differences in their interests, distributed photovoltaic aggregators, market trading centers, and distribution system operators exhibit complex market trading game behaviors. Accordingly, a single-leader multi-follower mixed-integer Stackelberg game framework is constructed for distributed photovoltaic aggregators participating in the day-ahead electricity market. The upper-level model aims to maximize the profit of the distributed photovoltaic aggregator (leader) by optimizing bidding strategies, while the lower-level model involves the market trading center (follower 1) conducting day-ahead joint market clearing, and the distribution system operator (follower 2) performing security verification on the market clearing results based on discrete control measures such as transformer tap changers and capacitor switching. To solve this multi-agent game model, the Karush-Kuhn-Tucker conditions and the Big-M method are first used to equivalently transform the follower 1 problem. Subsequently, a data-driven bilevel reconstruction algorithm is employed to solve the leader-follower game model with continuous and discrete variables. Finally, the accuracy and effectiveness of the game model and its solution algorithm are validated using a practical transmission and distribution system in a certain region.}
}