报告题目：Fault Diagnosis of Discrete-Event Systems: A Petri-Net Approach

报 告 人：李志武，澳门科技大学教授，IEEE Fellow

报告时间：2026年5月15日(周五) 15:30-16:50

报告地点：信息与控制工程学院B202

个人简介：Zhiwu Li received the B.S. degree in mechanical manufacturing, M.S. degree in mechatronics, and Ph.D. degree in Mechanical Engineering, all from Xidian University, Xi’an, China, in 1989, 1992, and 1995, respectively. Dr. Li held visiting professor position at the University of Toronto, Technion (Israel Institute of Technology), Martin-Luther University at Halle, University of Cagliari, Politecnico di Bari, Conservatoire National des Arts et Métiers (Cnam), King Saud University, and Meliksah University. He has published three monographs in Springer (2009; 2023) and CRC Press (2013). Currently he is a Chair Professor with the Macau University of Science and Technology. He is research was cited by leading business giants IBM, HP, ABB, Volvo, GE, GM, Mitsubishi, Ford Car, Hitachi and Huawei. His current interests include Petri net theory & application, production automation, and supervisory control of discrete event systems. He is a Fellow of IEEE (2016) and was selected as Thomson Reuters Highly Cited Researchers in the category of Engineering from 2014-2018.

报告摘要：

This report explores the fault diagnosis and diagnosability verification of discrete event systems based on labeled Petri nets. Firstly, it introduces the basic concepts of fault diagnosis for discrete event systems and the formal definition of labeled Petri nets. An online fault diagnosis method based on integer linear programming is proposed, which judges whether a fault has occurred by solving an optimization model constructed from observed labels. To address the state explosion problem, a basis reachability graph is introduced as a compact representation of system behavior, which can significantly reduce computational complexity while retaining all information required for diagnosis. By virtue of the basis reachability graph, diagnosability verification can be completed via a diagnoser or a verifier without constructing the complete reachability graph. For non-diagnosable systems, two approaches to enhancing diagnosability are further discussed. Finally, benchmark tests verify the efficiency of the basis reachability graph method, and future research directions including fault diagnosis under external attacks and dynamic system reconfiguration are prospected.

信息与控制工程学院

2026年5月13日