Research on Multi-scenario Inter-provincial Power Transmission Strategy Based on Reinforcement Learning

With the advancement of electricity market reform, inter-provincial electricity trading faces many challenges. Traditional strategies for inter-provincial transmission are difficult to adapt to the complex environment, while reinforcement learning has the capabilities of autonomous learning, decision optimization, and environmental adaptability. Therefore, this paper proposes a multi-scenario method for inter-provincial transmission based on reinforcement learning. Firstly, different confidence level scenario sets are generated based on the multi-temporal source-load power prediction of LSTM (Long Short Memory) and the established evaluation index system. Secondly, an inter-provincial transmission strategy model based on the PPO (Proximal Policy Optimization) algorithm is established based on the benefit model for inter-provincial transmission, and the optimal strategy is obtained through training and optimization. Finally, through case analysis, the source-load power prediction results, the PPO algorithm training results, and the inter-provincial transmission results under different scenarios are presented, with the expectation of providing support for decision-making in the inter-provincial electricity market.