Deep learning based Energy-Efficient Handover in Multi-Operator Heterogeneous Electric Power Networks

In this paper, we propose a novel energy-efficient soft handover scheme by invoking the deep learning method (DL) based on a handover preference analysis approach for multi-operator heterogeneous networks (HetNets) for electric power scenarios. Considering the historical handover strategy of electric meter equipment (EME), firstly, the preference of the different operators can be estimated by the back propagation neural network (BPNN), then based on the preference parameters, we invoking the deep belief network (DBN) to generate the handover strategy to maximize the average data rate of EMEs and minimize the system power consumption and handover frequency. Simulation results show that our proposed scheme can achieve effective soft handover in ultra-dense HetNets.