Dynamic Bayesian Game Based Power Allocation in Mobile Edge Computing with Users’ Behaviors

In this paper, we investigate the power allocation problem based on users' behaviors in mobile edge computing (MEC) networks. Due to the information limitation about the behavior types of mobile users, the power allocation is quite challenging. To deal with this problem, we propose a dynamic Bayesian game based power allocation algorithm to maximize the utility function of MEC network. In Bayesian game model, the MEC server imposes a price per unit power on mobile users. And the MEC server modified the price based on the Bayesian probability information about the behaviors of the users. The mobile users dynamically adapt the behavior types and the required power according to their own performance and cost. We investigate the existence of Bayesian Nash equilibrium and simulation studies are carried out to demonstrate the effectiveness of the proposed algorithm.

I. Introduction

Mobile cloud computing (MCC) is known as an effective solution to bridge the gap between the resource-constrained mobile devices and the resource-demanding applications by offloading the compute-intensive tasks to the remote clouds [1] . However, MCC faces a large communication latency issue because of the long distance between mobile devices and remote clouds. Hence, for real-time and delay sensitive applications, MCC is not suitable for computation offloading. Mobile edge computing (MEC), which move cloud resource to the vicinity of the mobile network, can address the high latency issue [2] . Since MEC servers are located at the edge of mobile networks, the response time of cloud-based mobile applications is greatly reduced. Besides offering computing resource, the MEC server is able to provide storage and caching service, and help base station to allocate power and bandwidth resource to mobile users [3] .