Author(s): Katharina Teuber
Linked Author(s):
Keywords: Large-language models (LLMs); Model calibration; Multi-agent systems; Urban drainage modelling; SWMM
Abstract: The contribution presents a multi-agent LLM framework that autonomously calibrates urban drainages models using EPA-SWMM. In a case study, the system reaches calibration performance comparable to an automatic calibration routine. Its advantage lies in the chat-like conversation structure that can enable non-programmers to use automatic calibration routines. This demonstrates the potential of LLM-based systems to accelerate calibration in urban drainage systems.
Year: 2026