Urbanization and climate change pose two major challenges related to urban stormwater management: flooding of the urban area, and the environmental impact on receiving waters from urban catchment runoff. These challenges cause a significant increase in the cost of constructing, maintaining, and upgrading urban stormwater infrastructure for water utility companies to ensure efficient and safe stormwater discharge. Currently, most stormwater systems are static systems involving no control at all. And if control is applied, those are still manually designed rule-based controllers taking the current state into account but not, for example, the weather forecast of the near future. In this presentation I will showcase recent research results and future challenges on modeling urban stormwater infrastructure as hybrid Markov decision processes, obtaining near-optimal controllers for them using reinforcement learning, and using formal methods to obtain shields for guaranteed safe learning.
Översikt
- Datum:Startar 15 november 2024, 10:00Slutar 15 november 2024, 11:00
- Plats:EDIT-rummet 3364
- Språk:Swedish and English
Biography:
Martijn Goorden is an assistant professor at the Department of Mechanical Engineering, Eindhoven University of Technology, the Netherlands, since October 2024. He obtained his Ph.D. degree in Mechanical Engineering at Eindhoven University of Technology, the Netherlands, in 2019 and joined Aalborg University as a postdoc in 2020. In 2022 he became assistant professor at Aalborg University. Since 2020, he is also a senior advisor on synthesis-based engineering within Rijkswaterstaat, part of the Dutch Ministry of Infrastructure and Water Management, the Netherlands. His research interests are in the area of supervisory control theory, synthesis-based engineering, and formal methods in control.
Martijn is a committer to the open-source Eclipse project ESCET (Eclipse Supervisory Control Engineering Toolkit) that is used both in academia and industry for developing supervisory controllers in a synthesis-based approach, with notable applications in infrastructural systems and lithography. In this role, he actively contributes to the development of ESCET with new algorithms and code implementations as well as reporting and solving issues.
