<![CDATA[The development of Industry 5.0 demands intelligent, adaptive, sustainable computing systems oriented toward human-machine collaboration. In this context, edge computing plays a crucial role because it can process data in real time near the data source, supporting industrial applications such as smart manufacturing, predictive maintenance, and cyber-physical systems. However, the increasing use of edge devices results in high energy consumption, limited computing resources, increased operational costs, and an increased carbon footprint of digital industrial systems. Therefore, energy efficiency in edge computing has become a strategic and pressing issue. The problem is further complicated by the fact that most edge resource management currently relies on static or heuristic approaches that are less adaptable to the dynamics of industrial workloads. Artificial Intelligence (AI)-based approaches have the advantage of accurately predicting workloads, but generally fail to guarantee optimal resource allocation. In contrast, mathematical optimization methods such as Linear Programming (LP) are capable of producing optimal solutions but are less adaptable to changing dynamic system conditions. This research aims to develop a hybrid AI-Optimization model based on Linear Programming to improve the energy efficiency of edge computing in Industry 5.0. AI models are used to predict workloads and computing demands, while LP is utilized to determine optimal resource allocation by minimizing energy consumption without violating the Service Level Agreement (SLA). The research methods include collecting workload datasets, developing machine learning prediction models, formulating LP models, and integrating the two into an adaptive system]]>
Copyrights © 2026
