How should cities invest in green and grey stormwater infrastructure to withstand extreme rainfall? This project identifies the most damaging storms for a city’s infrastructure and develops robust investment guidelines that account for drainage capacity, flood risk, and budget constraints.

Our analysis and case studies based on Toronto, Manhattan, and New Orleans reveal that stronger infrastructure generally shifts the worst-case event toward shorter, more intense storms, while planning against rarer events shifts it toward longer storms. Grey retention facilities become particularly valuable under severe rainfall or tight budgets, despite the absorption advantages and urban co-benefits of green infrastructure.

How should water utilities coordinate recycling, storage, and primary water production as demand and wastewater inflows fluctuate? This project uses environmental signals to develop robust operating policies that dymically manage production/recycling costs and water shortage risks.

Using Melbourne Water data, we demonstrate improvements in out-of-sample costs and protection against severe shortages. Importantly, beyond the environmental benefits of water recycling, our work demonstrates its operational value as an intertemporal buffer that helps managers prepare for future misalignment between recyclable supply and service demand.

I design pooled testing strategies that balance resource use and detection performance. My group testing paper accounts for correlated infections; an ongoing score-based pooling project studies pool sizes and follow-up testing rules for newborn cystic fibrosis screening under a budget.

Using COVID-19 outbreak data from Hong Kong,  we show that our strategy could save up to 0.5 test per individual relative to the region's deployed testing policy. For the ongoing score-based pooling project, we will use newborn cystic fibrosis screening data from New York and California to evaluate our pooling designs.

How can repeated resource allocation processes be made fairer through changing the order of service? Motivated by non-profit and humanitarian resource-allocating operations, this project combines randomized service orders with allocation rules to balance fairness over time with protection against shortages in individual rounds.

We conduce a case study using Toronto shelter data to illustrate reductions in persistent service disparities and improvements in variant fairness metrics, demonstrating the value of service order as an operational decision.