Monitoring and Diagnosis of Complex Systems

In complex systems (such as electric grids, computer networks, automobiles, air traffic control systems, and others), a single faulty behavior could cause cascading failures in a very short period and potentially lead to unpredictable results. As a result, monitoring and diagnosis of complex systems could be extremely challenging because of the complicated interactions between subsystems, but they are very important to ensure safety operations. Such complex systems can be abstracted as discrete event systems (DESs) at a high level. This talk focuses on two related topics: the monitoring and diagnosis of a given DES equipped with limited sensors, and configuring a set of sensors of minimal cost in a given DES to achieve accurate monitoring and immediate fault diagnosis. Different methods/results are proposed to address these two problems and the talk concludes by highlighting future research directions
Yu Ru is currently a Ph.D. candidate in the Department of Electrical and Computer Engineering at the University of Illinois at Urbana-Champaign. He received the B.E. degree in industrial automation in 2002, and the M.S. degree in control theory and engineering in 2005, both from Zhejiang University, Hangzhou, China. His research focuses on monitoring, diagnosis, and control of discrete event systems, optimization, approximation algorithms, Petri net theory, and network tomography.