Description:

Phebe Vayanos
University of Southern California

Abstract: Robust optimization is a popular paradigm for modeling and solving two- and multi-stage decision-making problems affected by uncertainty. Most approaches assume that the uncertain parameters can be observed for free and that the sequence in which they are revealed is independent of the decision-maker's actions. Yet, these assumptions fail to hold in many real-world applications where the time of information discovery is decision-dependent and the uncertain parameters only become observable after an often costly investment (of time or money).
To fill this gap in the literature, and motivated from the problem of learning the moral priorities of policy-makers at the Los Angeles Homeless Services Authority, we first consider two-stage robust optimization problems in which (part of) the first stage variables decide on the uncertain parameters that will be observed between the first and the second decision stages. The information available in the second stage is thus decision-dependent and can be discovered (at least in part) by making strategic exploratory investments in the first stage. We propose a novel min-max-min-max formulation of the problem. We prove correctness of this formulation and leverage this new model to provide a solution method inspired from the K-adaptability approximation approach, whereby K candidate strategies are chosen here-and-now and the best of these strategies is selected after the portion of the uncertain parameters that was chosen to be observed is revealed. We reformulate the problem as an MILP solvable with off-the-shelf solvers and also generalize this popular approximation scheme to multi-stage problems. We demonstrate the effectiveness of our method compared to the state of the art.
This is joint work with Angelos Georghiou and Han Yu.

Biography: Phebe Vayanos is an Assistant Professor of Industrial & Systems Engineering and Computer Science at the University of Southern California. She is also an Associate Director of the CAIS Center for Artificial Intelligence in Society at USC. Her research aims to address fundamental questions arising in data-driven optimization (a.k.a. prescriptive analytics) with aim to tackle real-world decision- and policy-making problems in uncertain and adversarial environments. Her work is motivated by resource allocation problems that are important for social good, such as those arising in public health, public safety and security, public housing, biodiversity preservation, and education. She is also interested in issues surrounding fairness, efficiency, and interpretability in resource allocation. Prior to joining USC, she was lecturer in the Operations Research and Statistics Group at the MIT Sloan School of Management, and a postdoctoral research associate in the Operations Research Center at MIT. She holds a PhD degree in Operations Research and an MEng degree in Electrical & Electronic Engineering, both from Imperial College London.