Stochastic Programming for Investment under Uncertainty
GPU-accelerated first-order methods for large stochastic programs, with application to investment decisions made before the future is known.
About
I am a Lieutenant Colonel in the United States Marine Corps and a fellow in the Marine Corps Doctor of Philosophy Technical Program (PHDP-T) at the Naval Postgraduate School. My work sits at the intersection of GPU computing, mathematical optimization, and decision-making under uncertainty.
The dissertation develops GPU-accelerated splitting methods for large-scale stochastic programs, as an alternative to classical scenario decomposition. The application class is investment under uncertainty: commitments made before a future is revealed, then adapted as information arrives. The research program extends toward multi-stage and nonlinear models. I have presented this work at the International Conference on Stochastic Programming (Paris, 2025), the SIAM Northern and Central California Sectional Conference (2025), and SIAM OP26 (Edinburgh, 2026).
Before this assignment I was an Operations Research Analyst at Manpower and Reserve Affairs, Headquarters Marine Corps, where I built analytical tools for the Marine Corps BAH appropriation program and applied optimization methods to officer talent management policy.
Research
Dissertation threads: methods, application class, and scale
Stochastic Programming for Investment under Uncertainty dissertation · in progress
Stochastic programming for investment decisions made before uncertainty is revealed. The dissertation treats this as a problem class — allocating resources under uncertainty, then adapting — rather than a single institutional instance.
GPU-Accelerated Splitting Methods for Stochastic Programs
Structure-exploiting first-order splitting methods for large stochastic programs. Current implementations use extensive-form structure on the GPU as an alternative to classical scenario decomposition; the research program extends toward multi-stage and nonlinear models. Presented at ICSP 2025 (Paris), SIAM NorCal 2025, and SIAM OP26 (Edinburgh).
Scale and Computational Evaluation
Computational study of structure-aware first-order methods on stochastic programs, including later work on inexact inner solves and multi-GPU scale-out. Comparisons are against structure-blind first-order solvers and classical decomposition.
Contact
Open to research collaboration and speaking invitations.
The best way to reach me is LinkedIn.
Message on LinkedIn