非侵入式市场竞争对手投标行为估计：面向网络化微电网的数据驱动逆优化方法

For a renewable energy-rich microgrid (MG) participating in a networked-MG local electricity market, the ability to submit strategic bids rather than passively offering marginal cost bids becomes increasingly attractive, which otherwise may harm its revenue. However, existing strategic bidding models all rely on the strong assumption that a strategic participant has perfect knowledge of its competitors’ bidding behaviors, which are not directly accessible in practice. To address this limitation, this paper proposes a non-intrusive estimation approach that enables a strategic MG to estimate the bidding behaviors (i.e., both bidding prices and quantities) of its competitors using only publicly available historical market clearing data, including the accepted power quantity of each unit, the nodal market clearing prices, and the nodal loads from past dispatch periods. This approach can thus provide a significant complement to existing strategic bidding models. Mathematically, the local market clearing problem is first formulated. The non-intrusive estimation approach is then cast as an inverse optimization model, developed by leveraging the strong duality between the primal market clearing formulation and its dual. Finally, case studies on two local markets of different scales are conducted to validate the effectiveness of the proposed inverse estimation approach.