TC DES Tutorial Series 2022 | Kai Cai's Systems Control Group

IEEE CSS TC DES Tutorial Series 2022

The IEEE CSS TC DES is organizing a Virtual Talk Series on Discrete Event Systems throughout 2022 to enhance communications in our community during the current Covid-19 pandemic. In order to access past talks, you will need to register on the page below under "Registration."

Date & Time:

The virtual events take place via Zoom on the 3rd Thursday of each month in 2022 (except August, September, December) at 13:00 UTC (Paris 14:00, New York City 8:00, Beijing 21:00). Please see the detailed schedule below.

Scope:

This year we will organize a variety of events of different forms. The main part is still talks on cutting-edge topics. In addition we will hold a panel discussion, a PhD student forum, and talks by authors of recent TAC (IEEE Transactions on Automatic Control) papers.

Registration:

Registration is free. For security reasons we require pre-registration. Participants will receive log-in details to the virtual zoom meeting after registration.

Please register here

Schedule:

*January 20 | 13:00 UTC

(Samuel) Qing-Shan Jia

Tsinghua University, China

Title: Event-Based Learning and Optimization for Cyber Physical Energy Systems

Abstract:

TBD

*February 17 | 13:00 UTC

Spyros Reveliotis

Georgia Institute of Technology, United States

Polynomial-Time Optimal Liveness Enforcement for Guidepath-Based Transport Systems

Abstract:

Guidepath-based transport systems is a popular abstraction for many contemporary applications, from the Automated Guided Vehicle (AGV) systems and the overhead monorail systems that are used in many industrial facilities, to the physical transport and processing of the ionized atoms that are the primary information carriers in quantum computing. The management of the traffic that takes place in these environments must be controlled for time- based efficiency, like throughput maximization and delay minimization, but also for more qualitative objectives like the preservation of the system liveness, i.e., the avoidance of potential deadlocks and the preservation of the ability of the system agents to complete successfully their running assignments. These qualitative problems can be rigorously addressed using models and analysis tools borrowed from theoretical computer science and the control- theoretic area of Discrete Event Systems. Their formal investigation also enables the formulation of an “optimal control” version of these problems through the notion of maximal permissiveness. In general, the deployment of maximally permissive liveness-enforcing supervision (LES) is a computationally hard task for most contemporary applications, due to the very large size of the involved state spaces. In this talk, we shall present a rather surprising result that establishes that for a very large and very practical class of the aforementioned transport systems, maximally permissive LES can be attained in time polynomial with respect to the size of the underlying guidepath network.

*March 17 | 13:00 UTC