Georgia Institute of Technology
A Simulation-Based Methodology for Sequential Decision-Making in Dynamic and Uncertain Tactical Environments
Abstract
dc:description.abstractThis dissertation presents a simulation-based methodology for sequential decision-making in dynamic and uncertain tactical environments, with application to tactical wildfire suppression. The framework integrates three core components: an agent-based model to represent assets and their interactions, a GPU-enhanced fire spread model to simulate wildfire propagation, and a Monte Carlo Tree Search to explore allocation decisions over time and under uncertainty. Together, these components create a simulation environment that enables the evaluation of alternative courses of action in settings where system behavior evolves rapidly and where decisions must be made with incomplete information. The methodology is motivated by the hierarchical and system-of-systems nature of wildfire response. Tactical suppression involves multiple independent assets that must be coordinated in support of broader incident objectives. Rather than framing the problem as one of strict optimality, this work emphasizes structured exploration of tactical alternatives under changing environmental conditions. The framework incorporates uncertainty through stochastic weather forecasts and updates tactical decisions at discrete intervals to evaluate its impact on operational outcomes. The framework is demonstrated across multiple wildfire case studies across varying landscapes and uncertainty levels. Results show that the GPU-enhanced fire spread model provides substantial speed advantages while maintaining agreement with higher-fidelity benchmark simulations. Across the tested scenarios, the proposed MCTS-based framework consistently outperforms heuristic allocation strategies, including those with access to future information, demonstrating the advantage of sequential and structured exploration of the decision space. Beyond wildfire applications, the methodology offers a generalizable approach for studying hierarchical tactical decision-making in complex dynamic environments.
Degree
thesis:*- Name thesis:degree_name
- Aerospace Engineering, PhD
- Department thesis:degree_department
- Aerospace Engineering
- Grantor
- Georgia Institute of Technology
- Year dc:date.issued
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Abyad, Mohamad Omar
- Advisor dc:contributor.advisor
-
- Mavris, Dimitri N.
- Committee members dc:contributor.committeemember
-
- Jordan, Jenna
- Bagdatli, Burak
- Amadori, Kristian
- Schrage, Daniel
- ORCID person:identifier.orcid
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- https://hdl.handle.net/1853/81710