A Method for Predicting Elastic Cloud Computing Resources Based on Sarima-Wnn Model

Abstract

The invention discloses a method for predicting elastic cloud computing resources based on the SARIMA-WNN model, which uses the seasonal time series model (SARIMA) combined with the wavelet neural network (WNN) prediction model to realize complementary advantages and improve prediction accuracy; SARIMA is based on the ARIMA model The seasonal periodicity factor is added on the basis of , and the cloud resource demand data of the past period is input into SARIMA (q, d, q) (P, D, Q) s In the model, d, p, q, D, P, Q are respectively obtained; the SARIMA model is used to predict the code after the smoothed sequence, and the prediction result can be obtained through the prediction and the L residual value is marked as r t ; Use the training samples to train the WNN network to obtain a model that is consistent with the prediction of elastic cloud resources, and then for the residual sequence r t Prediction is carried out, and the prediction result is marked as; Finally, the result predicted by the SARIMA-WNN combination model is obtained. The present invention solves the problems of inaccurate single model and poor effect of other combination models.

Description

technical field [0001] The invention belongs to the technical field of cloud computing, and in particular relates to a method for predicting elastic cloud computing resources based on a SARIMA-WNN model. Background technique [0002] In view of the current popularization and wide application of cloud computing, the cloud computing center provides tenants with resources in the form of virtual machines (VMs). Sufficient resources are prepared, but service failures for tenants caused by resource allocation lags, frequent SLA breaches, and excessive resource supply lead to a large amount of waste of cloud computing resources. In order to achieve ideal supply and demand balance, so efficient algorithms related to elastic cloud resources emerge as the times require. [0003] Elastic algorithms widely used in cloud computing include Autoregressive Integral Moving Average Model (ARIMA), ARIMA-improve, Back Propagation Neural Network (BPNN), SARIMA, etc. Elastic cloud resource dema...

AI Technical Summary

Problems solved by technology

[0002] In view of the current popularization and wide application of cloud computing, the cloud computing center provides tenants with resources in the form of virtual machines (VMs). Sufficient resources are prepared, but service failures for tenants caused by resource allocation lags, frequent SLA breaches, and excessive resource supply lead to a large amount of waste of cloud computing resources. In order to achieve ideal supply and demand balance, so an efficient elastic cloud resource-related algorithm emerges as the times require

Images