EvacuAIDi: AI-Driven Evacuation Framework

An AI-driven, causal-informed framework for probabilistic and disability-inclusive evacuation guidance.

📖 About

This is the presentation website for the PhD dissertation "EvacuAIDi: An AI-Driven, Causal-Informed Framework for Probabilistic and Disability-Inclusive Evacuation Guidance" by Amir Rafe at Utah State University (2025).

🚀 Live Demo

Visit the live presentation: https://evacuaidi.vercel.app

🔬 Research Overview

This dissertation addresses critical gaps in evacuation science through four interconnected research contributions:

1. AI Knowledge Extraction: Automated parameter extraction from fire safety codes
2. Disability-Inclusive Modeling: Enhanced agent-based models for heterogeneous populations
3. Causal Evaluation: Quantifying the impact of AI guidance on evacuation performance
4. Probabilistic Risk Assessment: Bayesian framework for uncertainty quantification

📊 Key Findings

• 22.5% reduction in evacuation time with AI guidance
• 15.9% greater benefits for individuals with disabilities
• 92%+ accuracy in automated parameter extraction
• 57% of risk variance attributed to occupant load

📧 Contact

Amir Rafe
Civil & Environmental Engineering
Utah State University
Email: amir.rafe@usu.edu
Advisor: Dr. Patrick Singleton

📄 License

This research presentation is part of academic work conducted at Utah State University.