Multi-Efficiency Based Resource Allocation for Cognitive Radio Networks with Deep Learning

In this paper, a multi-efficiency based scheme is considered, for not only the energy efficiency (EE) and spectrum efficiency (SE) of the primary users (PUs) and the secondary users (SUs), but also the computing efficiency (CE) of the deep learning based RA algorithm in the cognitive radio networks (CRN). Considering the OFDMA based resource allocation (RA) problem for the underlaying SUs, a weighted sum of the secondary interference power is introduced as the objective to minimize. To solve the problem, a deep learning based reweighted message-passing algorithm (ReMPA) is proposed. Compared to the previous scheme and the traditional scheme, the simulation results show that the proposed scheme has effective improvement on both the SE and EE for PUs and SUs.