Intelligent optimization method and device for cloud edge-end collaborative federated learning

The invention discloses an intelligent optimization method and device for cloud edge-end collaborative federated learning, and relates to the field of artificial intelligence. According to the method, a federal learning framework topological structure based on a cloud edge end is reasonably constructed through a lightweight training device, the state of the current round of environment is constructed based on model parameters, training time, training power consumption and communication time, and actions including the edge aggregation frequency and the number of terminal training round times are generated through an intelligent agent decision model. The edge and the terminal equipment are trained according to the action, meanwhile, information is collected to form a next round of state, and the federal learning framework and the decision model continuously interact to generate a large amount of decision trajectory information for updating the decision model until the model converges; the trained agent decision