A Note for CPS Data-driven Approaches Developed in the IDS Lab

Andreas A. Malikopoulos
2024-06-19
Abstract:The rapid evolution of Cyber-Physical Systems (CPS) across various domains like mobility systems, networked control systems, sustainable manufacturing, smart power grids, and the Internet of Things necessitates innovative solutions that merge control and learning [1]. Traditional model-based control methodologies often fail to adapt to the dynamism and complexity of modern CPS. This report outlines a comprehensive approach undertaken by the Information and Decision Science (IDS) Lab, focusing on integrating data-driven techniques with control strategies to enhance CPS performance, particularly in the context of energy efficiency and environmental impact. CPS are intricate networks where physical and software components are deeply intertwined, operating as systems of systems. These systems are characterized by their informationally decentralized nature, posing significant challenges in optimization and control. Classical control methods depend heavily on precise models, which often do not capture the full complexity of real-world CPS. As these systems generate large volumes of real-time data, there is a growing need for control algorithms that can leverage this data effectively. The IDS Lab is at the forefront of developing such data-driven approaches for CPS.
Optimization and Control
What problem does this paper attempt to address?