An LSTM Enabled Dynamic Stackelberg Game Theoretic Method for Resource Allocation in the Cloud

Abstract

Resource allocation is essential in cloud computing because it affects the performance, functionality, and development of cloud services. Resource allocation in the cloud based on economic and pricing approaches has the potential to increase the infrastructure suppliers’ revenue and improve the service providers’ efficiency and quality of services. However, in practice, information may be imperfect, and resulting resource allocation is not fair. Therefore, there is a need for a more appropriate model which could leverage imperfect information. In this paper, we propose an imperfect information based game theoretic method for resource allocation in the cloud. In this method, both an information control strategy and an LSTM model are used to predict the market status and to optimize bidding strategy for achieving maximal profits while maintaining fairness and profits for tenants with huge demands. The simulation results demonstrate the feasibility of the simulation framework as well as the effectiveness of the proposed method.