Energy efficient internet of things task scheduling framework in edge computing

Abstract

ABSTRACT newlineOver the years, computing systems have evolved to become faster, smarter, newlineand more efficient. Initially, computers were connected in groups to work newlinetogether and solve large problems more quickly. This approach was followed by newlinegrid computing, which allowed computers in different locations to share their newlineresources. Later, cloud computing gained popularity by offering services and newlinestorage over the Internet, accessible anytime and anywhere. However, with the newlinerapid growth of smart devices such as fitness trackers, smart home systems, and newlineconnected vehicles, there is an increasing need to process data quickly and closer newlineto where it is generated. This is where edge computing plays an important role. newlineInstead of sending all data to distant servers, edge computing processes it newlinenearby, saving time and reducing Internet traffic. newlineYet, this introduces new challenges. Devices at the edge vary widely in memory newlinecapacity, power, and processing speed. The volume and arrival rate of tasks are newlineconstantly changing. To keep the system running smoothly, we need intelligent newlinemethods to decide which device should handle which task and when. This process, newlinecalled task scheduling, is essential for making edge computing systems reliable and newlineefficient. Therefore, this thesis explores the challenge of task scheduling in edge newlinecomputing systems. It proposes adaptive, nature-inspired algorithms designed to newlineoptimize execution time, energy consumption, make span and average wait time, newlineaiming to make edge IoT systems more efficient and reliable. newlineSpecifically, the first contribution introduces a Coati optimization-based newlineenergy-aware task scheduling (CO-ETS) approach for efficient task assignment in newlineedge computing. The CO-ETS algorithm mimics coatis hunting and escaping newlinebehaviors in two phases: exploration for task distribution and exploitation for newlinerefinement of scheduling. It uses a specially designed fitness function that newlinebalances energy usage and task completion time. By optimizing task allocation newlinethrough this fitness function, CO-ETS signi