A Dynamic Resource Management Strategy for Cloud Based on Sequential Decision Model and Demand Forecast

In order to improve the utilization of finite resources, lower the cost of electrical power consumption and provide high QoS (Quality of Service) for customers, resource management approaches should be chosen carefully and dynamically. In this paper, we model the resource management process as a sequential decision problem. In addition, we apply the grey model to forecast the future demands. The experiments show that by using the grey model to predict the demands and dynamic sequential decision model to decide which actions should be performed, our approach can always achieve a good performance for different load models and different forecast accuracy rates.