My current research focuses on modeling observational studies as discrete optimization problems, understanding the complexity of the resulting model, and developing exact and approximation algorithms for solving the optimization problems. This research seeks to extend and generalize the methods currently being used by statisticians and applied researchers for drawing causal inferences from observational data.

I am also working with several other researchers to develop new branch and bound search strategies for NP-Hard problems. We're focusing on problems with an exponential number of variables so column generation is needed. Our goal is to identify search strategies that are able to either preserve the structure of the pricing problem after branching, or minimize the number of times the pricing problem needs to be resolved.

Finally, I am supervising several undergraduate students at the University of Illinois in their work on an election forecasting website. We began forecasts for the 2016 presidential election this past spring, and will have Senate forecasts available at the end of August. One new feature that we have added this year is the ability to create custom forecasts, which can be found here.

Links that provide an overview of operations research can be found here.

Last modified: 2019-06-05