基于自适应神经网络的云资源预测模型

为了避免容器云资源因资源供求不均衡而导致的资源利用率差等问题,需要对未来时刻的资源需求情况进行预测来进行更精准的调度和分配资源,因此,结合神经网络的高效学习能力与自适应调整的学习率,提出一种基于自适应神经网络的云资源预测模型。首先,融合卷积神经网络(convolutional neural network, CNN)和长短期记忆(long short-term memory, LSTM)网络的特点去挖掘历史数据的特征,预测未来的资源需求;然后,根据模型预测情况自适应调整学习率,提高模型预测的精度。使用Microsoft Azure公开数据集进行测试,相较于单一模型CNN、LSTM和未加入自适应学习率的神经网络模型,均方根误差分别下降了17.74%、18.27%和6%,证明了模型的有效性。

Cloud Resource Prediction Model Based on Adaptive Neural Network

In order to avoid the problem of poor utilization rate of container cloud resources due to the imbalance of resource supply and demand, it is necessary to forecast the resource demand in the future to schedule and allocate resources more accurately. Therefore, combined with the efficient learning ability of neural network and the adaptive learning rate, a cloud resource prediction model based on adaptive neural network is proposed. Firstly, the characteristics of historical data were mined by combining the characteristics of convolutional neural network (CNN) and long-term memory network (LSTM) to predict the future resource demand. Then the learning rate was adjusted adaptively according to the prediction of the model to improve the accuracy of model prediction. The test results with Microsoft Azure public dataset show that compared with CNN, LSTM and the neural network model without adaptive learning rate, the root mean square error decreases by 17.74%, 18.27% and 6% respectively, which proves the validity of the model.