580 California St., Suite 400
San Francisco, CA, 94104
Key research themes
1. How does algebraic and automata theory provide foundational models and analysis tools for discrete event systems?
Research in this area focuses on formal mathematical frameworks to model, analyze, and control discrete event systems (DES). This includes the use of max-plus algebra, automata theory, and Petri nets to capture synchronization, event timing, and system states. Such algebraic and automata-theoretic models enable rigorous system representation, verification, and performance analysis critical for industrial and cyber-physical applications.
Key finding: This survey contextualizes max-plus algebra as a key algebraic tool enabling a 'linear system theory' for synchronization-dominated DES. It formalizes timed event graphs (TEGs) as max-plus linear systems, allowing DES timing... Read more
Key finding: This foundational text develops formal modeling of DES using automata and languages, detailing the role of finite automata in representing event-driven systems. It outlines operations on automata, determinism, and... Read more
Key finding: This work introduces GPenSIM, a Petri net-based simulation tool implemented in MATLAB for modeling, simulating, and controlling DES. It bridges Petri net theory and practical analysis by providing an extensible platform... Read more
Key finding: While focusing primarily on social and organizational events, this paper applies the Event System Theory (EST) to characterize impactful events with dimensions of novelty, change, and criticality — concepts foundational to... Read more
Key finding: This paper advances DES formalism by combining Timed Automata with Guards and max-plus algebra for modeling asynchronous networked control systems (NCS). It derives analytical formulas for response time bounds, validating... Read more
2. What methods exist for fault diagnosis, prognosis, and supervisory control to ensure safety and reliability in discrete event systems?
This research theme explores advanced diagnostic and predictive control methods for fault-tolerant operation of DES. The focus is on identifying fault occurrences (diagnosis), anticipating faults before occurrence (prognosis), and designing supervisory control strategies for safety. It covers modeling intermittent faults, online fault prognosis, and integrating these into supervisory control to proactively maintain system safety.
Key finding: This paper extends traditional fault diagnosis models by explicitly modeling intermittent faults with alternating fault and reset events, often unobservable. It introduces novel diagnosability notions and enhances diagnoser... Read more
Key finding: The study develops a prognostic-based supervisory control approach that anticipates fault occurrences rather than reacting post-fault. It introduces a prognosing controller to switch between nominal and degraded supervisors... Read more
Key finding: This work proposes a novel Intelligent Regulation Circuit modeled with Petri nets to make non-diagnosable DES diagnosable without altering system operation. By monitoring event sequences and reducing a 'relative distance'... Read more
3. How can real-time and hierarchical event stream models improve schedulability analysis and event density characterization in discrete event systems?
This theme investigates advanced event stream representations for real-time embedded systems, incorporating dependencies and hierarchical structures to accurately model event timing and interrelationships. These improved models enable tighter bounds on event densities and more precise schedulability and feasibility analyses, which support resource optimization and system performance guarantees in hard real-time environments.
Key finding: This paper introduces a model that accounts for task dependencies triggered simultaneously by the same event source in distributed real-time systems. Using the event stream model, it demonstrates that considering these... Read more
Key finding: The authors propose hierarchical event streams as an expressive model capturing internal events generated dynamically within tasks in embedded real-time systems. They provide an efficient feasibility test based on demand and... Read more
4. What are effective identification and reconstruction methods for event behavior and state estimation in discrete event systems under partial observability?
This theme covers identification of DES models from observed data, reconstructing unobserved or incomplete event logs, and state estimation in systems with shared resources and asynchronous observations. It addresses challenges due to partial event observability, unrecorded events, and the interactions among concurrent cases, enabling accurate performance analysis, root cause diagnosis, and systems modeling.
Key finding: This work advances DES identification by addressing practical data collection and model completeness in closed-loop plant-controller setups using binary signals. It shows that under specific observed language conditions, the... Read more
Key finding: The paper formulates the problem of reconstructing unobserved events and bounding their timestamps in systems with shared resources modeled as PQR-systems. It introduces multi-entity event logs and process models to capture... Read more