Multi-energy coupling ground source heat pump building energy supply system scheduling method based on reinforcement learning

The invention discloses a multi-energy coupling ground source heat pump building energy supply system scheduling method based on reinforcement learning. The method comprises the steps that a multi-energy coupling ground source heat pump building energy supply system digital twinborn model is established; based on the digital twinborn model of the multi-energy coupling ground source heat pump building energy supply system and the constructed multi-energy coupling ground source heat pump building heat load and cold load influence factor knowledge graph, factors influencing building heat load and cold load changes are obtained; building a building load prediction model by adopting reinforcement learning and deep learning algorithms in combination with building thermal load and cold load historical data and influence factors; building load requirements are obtained on the basis of the building load prediction model, a multi-energy coupling ground source heat pump building energy supply system economic dispatching