Ruili Xu
Incoming MSc Student, Electrical and Software Engineering, University of Calgary
Agentic Simulation, LLM Agents, Wildfire Evacuation
Biography
Ruili Xu is an incoming MSc student at DENOS Lab, holding a BSc from the University of Toronto. Ruili first joined the lab as a summer intern in 2026 and continues as a graduate student in the Department of Electrical and Software Engineering.
Ruili’s research sits in the lab’s work on agentic simulation. During the internship Ruili connected AgentEvac, the lab’s wildfire evacuation simulator, with OASIS, an LLM-agent framework for simulating social media platforms, so that simulated residents decide whether and when to evacuate under official announcements, peer behaviour, and platform recommendation algorithms. Preliminary results on a Halifax and Tantallon scenario suggested that the choice of underlying language model shifts evacuation timing at least as much as the recommendation policy does. Dr. Steve Drew presented the abstract at IEEE SMC 2026 in Bellevue, Washington.
The MSc research will build on that study, asking how the information a community shares with itself during a disaster changes how fast it can get out.
Projects
-
Agentic Simulation for Wildfire Evacuations
Connecting the lab's AgentEvac wildfire evacuation simulator with the OASIS LLM-agent social simulation framework to study how official warnings, peer behaviour, and recommendation algorithms shape when residents leave.
Publications
- IEEE International Conference on Systems, Man, and Cybernetics (SMC), abstract2026
Simulating Socially Mediated Wildfire Evacuation Decisions with LLM Agents