Research
Three threads run through my work: where to put charging infrastructure, how populations evolve over decades, and how people actually decide to move.
2025 – Present · University of Washington
A mixed-integer linear optimization framework for siting megawatt-class DC fast-charging stations along major corridors that serve both long-haul electric trucks and private EVs. The model captures temporal origin–destination flows, queuing dynamics, and grid constraints to decide where shared charging hubs deliver the most value.
A Markov dynamic Bayesian-network demographic microsimulation module inside POLARIS that simulates household and individual life-stage transitions (aging, education, employment, marriage/divorce, birth/death, and migration) for long-horizon demographic forecasting.
A randomized A/B experiment evaluating whether AI-generated, in-app warnings reduce sidewalk riding on shared e-scooters, pairing computer-vision detection with behavioral nudges. Spoiler: they work.
Charging-infrastructure siting and EV-ownership dynamics, from corridor freight to household adoption.
Choice modeling of car ownership, mode choice, and ridehailing using econometrics and survival analysis.
Building and evolving synthetic populations so models can forecast decades, not just snapshots.