基于GRU-RNN组合模型的云计算RLF仿真  

Simulation of Cloud Computing Resource Load Forecasting Based on GRU-RNN Combination Model

作　　者：胡应钢 宋泽瑞 姜静清[1,2] HU Ying-gang;SONG Ze-rui;JIANG Jing-qing(College of Mathematics and Physics,Inner Mongolia Minzu University,Tongliao Inner Mongolia 028000,China;College of Computer Science and Technology,Inner Mongolia Minzu University,Tongliao Inner Mongolia 028000,China)

机构地区：[1]内蒙古民族大学数理学院,内蒙古通辽028000 [2]内蒙古民族大学计算机科学与技术学院,内蒙古通辽028000

出　　处：《计算机仿真》2024年第10期513-516,共4页Computer Simulation

基　　金：国家自然科学基金(62162050,61662057)。

摘　　要：随着云计算的兴起和信息技术的快速发展,云计算数据中心存在能源消耗过高的问题,服务器的能源消耗主要是由CPU和内存等资源消耗过高引起的,这些资源利用率的不平衡是导致资源浪费和能源消耗的关键因素之一。对数据中心资源使用量进行预测,有助于进行资源管理,提高服务器资源利用率,从而缓解资源浪费和能耗过高的问题。本文建立一种基于门控循环单元与循环神经网络的组合预测模型GRU-RNN对负载进行预测。实验结果表明,提出的GRU-RNN模型相比于以往简单预测模型RNN、LSTM、GRU和现有的复合预测模型ARIMA-LSTM、GRU-LSTM等,有更高的预测精度。With the rise of cloud computing and the rapid development of information technology,there is a problem of high energy consumption in cloud computing data centers.The energy consumption of servers is mainly caused by the high consumption of resources such as CPU and memory.The imbalance of resource utilization is one of the key factors leading to resource waste and energy consumption.Predicting the usage of data center resources can help with resource management,improve server resource utilization,and alleviate ssues of resource waste and high energy consumption.This article establishes a combined prediction model CRU-RNN based on gated recurrent units and recurrent neural networks for load prediction.The experimental results show that the proposed GRU-RNN model has higher prediction accuracy compared to previous simple prediction models such as RNN,LSTM,GRU,and existing composite prediction models such as ARIMA-LSTM and GRU-LSTM.

关 键 词：门控循环单元 循环神经网络 负载预测 资源预测模型 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]