Virtual power plant strategic bidding method in day-ahead market

The invention discloses a strategy bidding method for a virtual power plant in a day-ahead market, and the method comprises the steps: proposing a day-ahead market clearing target of the virtual power plant, and building a day-ahead market clearing model of the virtual power plant; building a single-agent participant-evaluator model based on a depth deterministic strategy gradient algorithm, and determining a bidding decision through an agent of reinforcement learning; establishing an optimal power flow model; and solving the optimal power flow model, and determining operation points of all internal generators and the production marginal cost of the virtual power plant. According to the embodiment, the virtual power plant can participate in strategic bidding of the day-ahead market, reasonable distribution of bidding strategies of the virtual power plant is facilitated, and the cost is saved. 本发明公开了日前市场下虚拟电厂策略性竞价方法，方法包括提出虚拟电厂日前市场出清目标，建立虚拟电厂日前市场出清模型；搭建基于深度确定性策略梯度算法的单代理参与者-评判者模型，通过强化学习的代理确定投标决策；建立最优潮流模型；求解最优潮流模