Machine Learning and Optimization for Edge Computing based Internet-of-Things

Dear Colleagues,

The concepts of Internet of Things (IoT) are providing services across different sectors including smart cities, healthcare, manufacturing, agriculture, industrial internet, automobile and smart supply chains. However, the explosive growth of smart devices, applications, and volume of data traffic becomes a major concern in IoT environment, which is important to address before widely deploying the IoT services. Edge computing is a capable technology that enhances the quality and performance of the IoT services. The edge computing technology perfectly fits the architecture and features of IoT systems. Edge computing is the modern, distributed computing architecture that brings data storage and computation closer to the data source. This helps save bandwidth and improve the response time. Simply put, edge computing involves fewer processes running in the cloud. It also moves those computing processes to edge devices, such as IoT devices, edge servers, or users’ computers. This way of bringing computation closer, or at the network’s edge, reduces long-distance communication between a server and a client. Therefore, it reduces bandwidth usage and latency. Edge computing, over the years, has become an important architecture to support distributed computing and deploy storage and computation resources close to the same geographical location as the source. Although it employs decentralized architecture, which can be challenging and requires continuous control and monitoring, edge computing is still effective in solving advancing network issues like moving large

Journal of Autonomous Intelligence data volumes in less time than other computing methods. The unique architecture of edge computing aims to solve three main network challenges – latency, bandwidth, and network congestion. Edge computing finds applications in various industries. It is used to aggregate, process, filter, and analyze data near or at the network edge. Some of the areas where it is applied are: IoT Devices, Optimizing Network, Smart Healthcare, Intelligent Manufacturing, Transportation, and Intelligent Energy, etc. However, edge computing is still a novel concept and faces many challenges specifically related to integration and universal adoption, availability, safety, privacy and efficiency. The given challenges motivate us to explore various possible solutions for the intelligence IoT environment. Machine learning and optimization come into play as the leading solution strategies recently. This theme is concerned with machine learning and optimization techniques addressing various tasks in edge computing based IoT.

Prof. Bo Cheng

Prof. Shuai Zhao

Guest Editors