Research on optimization strategy of centralized bidding and trading of pumped storage units based on reinforcement learning theory

In recent years, the medium and long-term market trading volume in China's power trading market accounted for about 90%, if the pumped storage power plant as an independent subject to participate in the power market transactions, can be determined through the medium and long-term market trading strategy, significantly reduce the transaction risk. The medium and long term transactions in the centralized bidding transaction can best reflect the openness and fairness of the market, but also the highest transparency of the transaction, so this paper is based on the portfolio optimization method, to build the strategy set of market members; at the same time, based on the theory of reinforcement learning, and constantly simulate the market members of the intelligent decision-making behavior, and ultimately constructed the centralized bidding transaction optimization strategy of the pumped storage unit, and finally, through the simulation of the example, to validate the model's Finally, the model is validated through example simulation, which provides a reference for pumped storage energy to participate in power market transactions.