Low-carbon demand response method based on combined online learning

The invention discloses a low-carbon demand response method based on combined online learning, and belongs to the technical field of power load management, and the method comprises the steps: constructing a carbon emission model corresponding to a user power consumption behavior, and obtaining a linear quantitative relation-node carbon potential between a user power adjustment amount and a corresponding carbon reduction amount; dynamic node carbon potential is sent to a user through a carbon table to advocate the user to perform energy conservation and emission reduction; the method comprises the following steps: constructing a low carbon demand response problem under the influence of node carbon potential into a Multi-Armed Bands (MAB) model, introducing a Contextual Multi-Armed Bands (CMAB) theory, proposing a Linear User Confidence Bound LinUCB algorithm to solve the provided model, continuously updating a user selection strategy according to context information (node carbon potential) and a user carbon re