I develop scalable optimization methods for allocating limited resources when future demand is uncertain and historical data may not accurately represent the underlying distribution.
My initial work formulated a large-scale stochastic optimization problem across 254 Texas counties. I developed an exact cutting-plane decomposition algorithm in C++ and CPLEX that solved large instances 7× faster than the full formulation.
Building on this work, I developed a distributionally robust optimization framework and have been working on its application and the corresponding algorithm.
Methods: Distributionally robust optimization · Stochastic programming · Decomposition algorithms · C++ · CPLEX
Developed a network optimization model that matches independent couriers with delivery tasks through decentralized smart lockers. The solution combines a rolling-horizon heuristic with column generation to determine deliveries.
20% reduction in delivery cost
1,000+ packages evaluated per day
Published in Computers & Industrial Engineering, 2025
Developed and deployed a large-scale mixed-integer programming model for production and maintenance scheduling across high-volume manufacturing lines.
$440K annual labor-cost savings
4% reduction in yield loss