雾无线接入网中基于神经网络的资源分配方案

考虑雾无线接入网（Fog Radio Access Network，F-RAN）中的性能优化问题，提出一种基于深度神经网络（Deep Neural Network，DNN）的资源分配方案。该方案旨在通过资源分配策略来最大化经济频谱效率（Economical Spectral Efficiency，ESE）。为解决传统资源分配方案需要大量计算的问题，该方案借助神经网络模型，将ESE作为损失函数，使用更少的计算量来确定用户的波束赋形，从而实现实时处理。仿真结果表明，相比于基于传统凸优化功率分配方案或者是基于监督学习的CNN方法，所提出的方案的光谱效率（Spectral Efficiency，SE）和ESE的最大增益分别可以达到5%和20%。此外，该方案在执行时间上与CNN方案接近，明显优于传统算法。

Resource Allocation Framework Based on Deep Neural Network in Fog Radio Access Network

Consider performance optimization in fog radio access network, a resource allocation framework based on deep neural network is proposed. The framework aims to maximize the Economical Spectral Efficiency（ESE） through resource allocation strategy. In this framework, the resource allocation strategy to realize real-time processing is learned by means of a DNN while conventional resource allocation schemes require a considerable number of computations, and the beamforming can be determined using far fewer computations by a DNN. Through simulations, compared with the convex optimization based scheme and supervised learning based CNN scheme, the maximum gain of SE and ESE can be achieved respectively 5% and 20% in this framework. Furthermore, the execution time of the proposed framework is close to that of CNN scheme, which is much ahead of the traditional algorithm.