A new hybrid model for request rate prediction in mobile cloud computing

摘要

One of the key characteristics of mobile cloud computing compared to the traditional systems is its dynamicity that requires a scalable resource allocation scheme. Since the preparation of a new virtual machine in cloud is a time consuming task, the cloud provider must predict the resources that will request by users in near future. In this paper, we proposed an intelligent hybrid model that predicts the upcoming resource demands. The model uses neural networks and GARCH as statistical model to achieve high accuracy in demand prediction. Our hybrid model shows better results compared to the rival methods according to standard metrics.