A distributed ADMM approach for energy-efficient resource allocation in mobile edge computing

Yangyang LI,Feng XUE,Wenchen ZHOU,Naixue XIONG,Weiwei FANG,Xuening YAO

doi:10.3906/elk-1806-112

Abstract

Mobile edge computing (MEC) is a new promising technique to provide cloud-computing capabilities at the edge of cellular networks in close proximity to mobile users. In this paper, we consider joint optimization of the communication and computation resources in a multiuser, multiserver MEC system. The objective of this optimization problem is to minimize the total energy consumption of mobile devices under the time-sharing constraint. Given the fact that no coordination is involved between mobile devices, we propose a light-weight and decentralized algorithm based on the alternating direction method of multipliers (ADMM) framework. Experimental results demonstrate that the proposed algorithm performs well in terms of convergence and outperforms the conventional centralized approach.

Highlights

With the prevalence of mobile computing, more and more new mobile applications, e.g., augmented reality, interactive gaming, and video streaming, are emerging and posing stringent requirements for intensive computation and tight latency [1]
Compared to mobile cloud computing, mobile edge computing (MEC) has more advantages in terms of communication latency, energy consumption, and back-haul load [4]. Since both computation latency and energy consumption are critical for resource-constrained mobile devices, the design of computation offloading schemes has attracted considerable attention for MEC systems
The main contributions are summarized as follows: 1- An optimization problem is formulated to minimize the total mobile energy consumption under the timesharing constraint. 2- For reasons of performance, scalability, and robustness, we develop a new algorithm to solve this problem in a distributed fashion based on the alternating direction method of multipliers (ADMM)

Summary

Introduction

With the prevalence of mobile computing, more and more new mobile applications, e.g., augmented reality, interactive gaming, and video streaming, are emerging and posing stringent requirements for intensive computation and tight latency [1]. Current mobile devices generally possess limited local resources, which are insufficient to support sophisticated applications To address this issue, mobile edge computing (MEC) was recently proposed as a promising paradigm that extends cloud computing capabilities to the edge of radio access networks [2, 3]. In [8], a dynamic computation offloading algorithm based on Lyapunov optimization was introduced to minimize the execution cost in a green MEC system with energy-harvesting mobile devices. A task offloading scheduling algorithm was proposed in [9], which considers minimization of mobile energy consumption under the constraints of resource capacity and computation latency. The goal of this paper is to propose a joint optimization of communication and computation resources for task offloading in a multiuser, multiserver MEC scenario.

System model

Algorithm design

Findings

Conclusion