基于自适应学习率CNN的物联网网络资源分配方法研究  

Research on Resource Allocation Method for Internet of Things Network Based on Adaptive Learning Rate CNN

作　　者：李丹 江霞 柳琳 郭林霞 LI Dan;JIANG Xia;LIU Lin;GUO Linxia(Henan University of Finance and Economics,Zhengzhou 451464,China)

机构地区：[1]河南财政金融学院,郑州451464

出　　处：《智能物联技术》2025年第1期26-29,共4页Technology of Io T& AI

摘　　要：固定学习率在训练神经网络时存在收敛速度慢且易于过拟合的问题,降低了资源分配的效率。因此,研究基于自适应学习率卷积神经网络(Convolutional Neural Network,CNN)的物联网网络资源分配方法。设定目标函数,并结合约束条件,建立物联网网络资源分配的模型。在该模型中,引人自适应学习率机制,根据训练过程中的误差变化率动态地调整学习率,从而加速网络的收敛并有效避免过拟合。利用模拟数据进行训练后,CNN模型可输出物联网网络资源的最优分配方案。实验结果显示,该方法显著提高了物联网资源的利用率,增加了系统的吞吐量,进而提高了资源分配的效率。The fixed learning rate has the problem of slow convergence speed and easy overftting when training neural networks,which reduces the efficiency of resource allocation.Therefore,research on resource allocation methods for Internet of Things networks based on adaptive learning rate Convolutional Neural Network(CNN).Set the objective function and combine it with constraint conditions to establish a model for resource allocation in the Internet of Things network.In this model,an adaptive learning rate mechanism is introduced to dynamically adjust the learning rate based on the rate of error change during the training process,thereby accelerating the convergence process of the network and effectively avoiding overfitting.After training with simulated data,the CNN model can output the optimal allocation plan for Internet of Things network resources.The experimental results show that this method significantly improves the utilization rate of Internet of Things resources and system throughput,thereby enhancing the efficiency of resource allocation.

关 键 词：自适应学习率 卷积神经网络(CNN) 资源分配 系统吞吐量 

分 类 号：TN929.5[电子电信—通信与信息系统]