부산대학교 도서관

Escalating the performance of Discrete Event Simulations (DESs) in manufacturing factory environ-ments plays a valuable role towards harnessing their real-time predictive and analytical capabilities. However, this attraction of performance-enhancement, whilst justified, has often caused the oversight of examination into how other notable metrics such as energy -efficiency and Total Cost of Ownership (TCO) affect the feasibility of DES management and implementation in factories. Hence, this work investigates how a line-balancing DES combination of performance, energy and TCO varies between low-resource edge devices, high-resource edge devices and remote High Performance Computing (HPC) clusters in the cloud. The findings demonstrate that, although an HPC cluster is noticeably more viable in terms of per-formance, it is not a consistently advantageous option for energy-efficiency and TCO. Alternately, we argue that a high-resource edge device displays itself as the preferred factory-appropriate hardware-choice to complement all three metrics-of-interest considered in this work.