My research focuses on the design and analysis of decision-making models under uncertainty with limited information. It is grounded in applications, in particular portfolio and revenue management, where stochasticity is often difficult to characterize using traditional probabilistic techniques because of fast-changing environments, for stock prices, or of limited data, for new products. My overarching goal is to incorporate uncertainty in an intuitive, tractable manner, which (i) is well-suited to the information available, i.e., does not require managers to make restrictive assumptions on the underlying probability distributions, (ii) leads to tractable formulations that can be solved efficiently for the large data sets encountered in practice, and (iii) yields theoretical insights into the structure of the optimal solution, to foster a better understanding of the computer-generated strategy. These features play an important role in practitioners adopting quantitative approaches. To achieve this research’s goals, I use tools from robust optimization as well as data-driven optimization.

For more information on my research, click here.