Print Email Facebook Twitter Data-driven distributionally robust optimization using the Wasserstein metric Title Data-driven distributionally robust optimization using the Wasserstein metric: performance guarantees and tractable reformulations Author Mohajerin Esfahani, P. (TU Delft Team Tamas Keviczky) Kuhn, Daniel (Swiss Federal Institute of Technology) Date 2017 Abstract We consider stochastic programs where the distribution of the uncertain parameters is only observable through a finite training dataset. Using the Wasserstein metric, we construct a ball in the space of (multivariate and non-discrete) probability distributions centered at the uniform distribution on the training samples, and we seek decisions that perform best in view of the worst-case distribution within this Wasserstein ball. The state-of-the-art methods for solving the resulting distributionally robust optimization problems rely on global optimization techniques, which quickly become computationally excruciating. In this paper we demonstrate that, under mild assumptions, the distributionally robust optimization problems over Wasserstein balls can in fact be reformulated as finite convex programs—in many interesting cases even as tractable linear programs. Leveraging recent measure concentration results, we also show that their solutions enjoy powerful finite-sample performance guarantees. Our theoretical results are exemplified in mean-risk portfolio optimization as well as uncertainty quantification. To reference this document use: http://resolver.tudelft.nl/uuid:dccfb770-41e8-476a-9c89-0553f5725fdb DOI https://doi.org/10.1007/s10107-017-1172-1 ISSN 0025-5610 Source Mathematical Programming, 171 (2018) (1-2), 115-166 Part of collection Institutional Repository Document type journal article Rights © 2017 P. Mohajerin Esfahani, Daniel Kuhn Files PDF 10.1007_s10107_017_1172_1.pdf 2.65 MB Close viewer /islandora/object/uuid:dccfb770-41e8-476a-9c89-0553f5725fdb/datastream/OBJ/view