Mobile Edge Computing Enabled Intelligent IoT

Document Type

Article

Publication Title

Wireless Networks (United Kingdom)

Abstract

The proliferation of the number of IoT devices, the ever-increasing computation intensive applications pose great challenges on resource allocation and offloading. In this chapter, to address spectrum sharing and edge computation offloading problems in SDN-based ultra dense networks, we propose a second-price auction scheme for ensuring the fair bidding for spectrum rent, which enables the MBS edge cloud and SBS edge cloud to occupy the channel in cooperative and competitive modes. Then, a novel deep reinforcement learning (DRL)-based network structure is proposed to jointly optimize task offloading and resource allocation. Finally, we propose two pervasive scenarios including single edge scene and multiple edge scenes. In the single edge scenario, a novel deep reinforcement learning (DRL)-based framework is invoked for collaboratively optimizing the task scheduling, transmission power, and CPU cycle frequency under metabolic channel conditions. Meanwhile, we propose a multi-agent aided deep deterministic policy gradient (MADDPG) algorithm to alleviate interference in multiple edge scenarios.

First Page

271

Last Page

350

DOI

10.1007/978-3-031-26987-5_6

Publication Date

2-6-2023

Keywords

Computing offloading, Mobile edge computing, Multi-agent aided deep deterministic policy gradient, Ultra-dense networks

Comments

IR Deposit conditions:

OA version (pathway b) Accepted version

12 months embargo

Publisher's Bespoke License Published source must be acknowledged with citation Must link to publisher version with DOI Post-prints are subject to Springer Nature re-use terms

Link to Full Text

Share

COinS