Resource Management for Edge Intelligence

Due to rapid growth of Internet of Things (IoT) applications that demand large amounts of computing power, resource-constrained IoT devices typically do not have enough computing capability to satisfy the computational requirements. For example, inference or training of large machine learning models, e.g., Deep Neural Networks (DNNs), are typically quite resource-intensive. Computational offloading for edge intelligence can help overcome resource constraints of IoT devices, reduce computing load and improve task processing efficiency. Edge Computing and Fog Computing aim to harness computing resources in the proximity of the IoT devices to provide services in a timely manner with reduced accesses to the cloud, which may be time-consuming. For example, a DNN can be partitioned and processed collaboratively between the device, edge server and the cloud to optimize objectives such as performance and power consumption, based on factors such as workload, hardware platform capabilities and wireless bandwidth. This special issue addresses recent developments in algorithms and techniques for computational offloading for edge intelligence. Topics of interest include, but are not limited to:

• Techniques for profiling and modeling of latency and power consumption on device/edge/cloud

• Algorithms for dynamic and adaptive resource management for both computation and communication objectives
• Intelligent decision making for computational offloading with Machine Learning, esp. Reinforcement Learning.
• Security, and privacy issues
• Application case studies

Submission Details:

General information for submitting papers to JSA can be found at https://www.journals.elsevier.com/journal-of-systems-architecture. Submissions should be made online at https://www.editorialmanager.com/jsa/. Please select the “VSI: RMEI20” option as the type of the paper during the submission process. Please direct all enquiries regarding this SI to Prof. Shaohua Wan.

JSA has adopted the Virtual Special Issue model to speed up the publication process, where Special Issue papers are published in regular issues, but marked as SI papers. Acceptance decisions are made on a rolling basis. Therefore, authors are encouraged to submit papers early, and need not wait until the submission deadline.

Important Dates

Deadline for submission: June 30th, 2020

First review notification: August 31th, 2020

Second review notification: October 31th, 2020

Final notification to authors: November 30th, 2020

Guest Editors

Prof. Shaohua Wan, Zhongnan University of Economics and Law, China, shaohua.wan@ieee.org

Prof. Huaming Wu, Tianjin University, China, whming@tju.edu.cn

Prof. Joarder Kamruzzaman, Professor of Information Technology, Federation University Australia, joarder.kamruzzaman@federation.edu.au

Prof. Sotirios Goudos, Aristotle University of Thessaloniki, Greece, sgoudo@physics.auth.gr