Abstract

The purpose of this study is to design and implement a real-time agent-based simulation for a snowstorm disaster response. The simulation was fully implemented from first principles using the GAMA platform, chosen for its spatial modeling capabilities, which are fundamental for realistic disaster scenarios. The project involved constructing the environment using geographic data, programming autonomous agents representing snow plows and rescue vehicles, and developing novel coordination algorithms for real-time decision-making. The model evolved through multiple iterations, from uncoordinated agent behaviors to a dynamic, hybrid system where agents adapt and collaborate based on environmental conditions like snow accumulation. Snowplows were guided by a custom scoring system to prioritize roads, while rescue vehicles adjusted their navigation speed in response to road conditions. Agent behavior was structured using elements of the BDI architecture to simulate autonomous, goal-driven decision-making. Performance was evaluated through tailored metrics and visualizations to ensure agents behaved realistically and effectively. This thesis demonstrates the potential of multi-agent systems for simulating complex, evolving disaster environments. It contributes a fully operational, spatially grounded simulation that reflects the adaptive, collaborative nature of real-world emergency response, and offers new insights into how multi-agent systems can be used for more intelligent and responsive disaster management systems.

Advisor

Palmer, Daniel

Department

Computer Science

Disciplines

Artificial Intelligence and Robotics

Publication Date

2025

Degree Granted

Bachelor of Arts

Document Type

Senior Independent Study Thesis

Download

Share

COinS
 

© Copyright 2025 Zaynab Ramou