I am primarily interested in the development and application of analytics and decision models to problems arising in healthcare operations and medical decision-making. More generally, I study problems of sequential decision making under uncertainty. My interests include
- Methodology: Markov decision processes, robust optimization, multiobjective optimization, infinite-dimensional optimization
- Applications: organ transplantation, treatment planning, scheduling/operations in clinical settings
Some of my ongoing and past research projects are described below:
- Incentives of federal policy in organ transplantation: Under regulations enforced by the Centers for Medicare and Medicaid Service (CMS) and the Organ Procurement & Transplantation Network (OPTN), transplant programs in the US are evaluated based on their patients’ post-transplantation survival outcomes. While well-intentioned. These regulations are believed to have had an unintended adverse effect on transplant access. My research analyzes the impact of these regulations on the behavior of transplant programs.
- Improving access in organ transplantation: The waiting list of transplant candidates is ever-growing, as the demand for donor organs far outweighs their supply. My research explores the effect of technological advances and alternative allocation strategies on increasing the number of successful transplants in the US.
- Multiobjective discrete optimization: Multiobjective optimization problems, as the name suggests, seek to simultaneously optimize multiple objectives over a common set of constraints. These problems are ubiquitous in applications, but challenging to solve, especially when (some of) the variables are constrained to be integers. I study relaxation methods and duality theory for these problems, which have the potential to lead to significant computational improvements.
- Robust dynamic programming: Solutions to optimization problems are sensitive to misspecifications in problem parameters, and robust optimization aims to mitigate this by a worst-case approach. I have developed approximate solution methods for infinite-dimensional Markov decision processes with uncertain transition probabilities. I also study applications of robust optimization, such as inventory management in the face of demand ambiguity, or robustness to parameter uncertainty in response-guided treatment planning.