Data-driven distributionally robust optimization using the Wasserstein metric: performance guarantees and tractable reformulations
Open accessThe DOI matches its Crossref record, with a small difference in title or year; 50 of the 51 references it cites were found in a registry or an index.
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.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Distributionally robust optimization [273] provides a promising way to achieve worst-case probabilistic performance by incorporating all sample-generating distributions into an ambiguity set.”
From Edge artificial intelligence for 6g: vision, enabling technologies, and applications (2022) · cites this paper as [273]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Boyd, S., & Vandenberghe, L. (2004) · doi:10.1017/cbo9780511804441 · Cited by 56,601
Wu, Y. (1999) · Technometrics · doi:10.1080/00401706.1999.10485951 · Cited by 27,221
Optimization of conditional value-at-risk
Rockafellar, R. T., & Uryasev, S. (2000) · The Journal of Risk · doi:10.21314/jor.2000.038 · Cited by 6,725
Bertsimas, D., & Sim, M. (2004) · Operations Research · doi:10.1287/opre.1030.0065 · Cited by 4,557
Topics in Optimal Transportation
Villani, C. (2003) · Graduate Studies in Mathematics · doi:10.1090/gsm/058 · Cited by 4,412
Ben-Tal, A., El Ghaoui, L., & Nemirovski, A. (2009) · Princeton University Press eBooks · doi:10.1515/9781400831050 · Cited by 2,895
Delage, E., & Ye, Y. (2010) · Operations Research · doi:10.1287/opre.1090.0741 · Cited by 1,893
On the rate of convergence in Wasserstein distance of the empirical measure
Fournier, N., & Guillin, A. (2014) · Probability Theory and Related Fields · doi:10.1007/s00440-014-0583-7 · Cited by 1,655
Distributionally Robust Convex Optimization
Wiesemann, W., Kuhn, D., & Sim, M. (2014) · Operations Research · doi:10.1287/opre.2014.1314 · Cited by 947
Lang, S. (1993) · Graduate Texts in Mathematics · doi:10.1007/978-1-4612-0897-6 · Cited by 807
Robust Solutions of Optimization Problems Affected by Uncertain Probabilities
Ben-Tal, A., den Hertog, D., De Waegenaere, A., Melenberg, B., & Rennen, G. (2013) · Management Science · doi:10.1287/mnsc.1120.1641 · Cited by 748
Convex Optimization Algorithms
Bertsekas, D. P. (2015) · Cited by 689
Distributionally Robust Optimization and Its Tractable Approximations
Goh, J., & Sim, M. (2010) · Operations Research · doi:10.1287/opre.1090.0795 · Cited by 668
Worst-Case Value-At-Risk and Robust Portfolio Optimization: A Conic Programming Approach
Ghaoui, L. E., Oks, M., & Oustry, F. (2003) · Operations Research · doi:10.1287/opre.51.4.543.16101 · Cited by 662
A Min-Max Solution of an Inventory Problem
Scarf, H. (2005) · Herbert Scarf’s Contributions to Economics, Game Theory and Operations Research · doi:10.1057/9781137024381_3 · Cited by 552
Integral Probability Metrics and Their Generating Classes of Functions
Müller, A. (1997) · Advances in Applied Probability · doi:10.1017/s000186780002807x · Cited by 548
Data-driven chance constrained stochastic program
Jiang, R., & Guan, Y. (2016) · Mathematical Programming · doi:10.1007/s10107-015-0929-7 · Cited by 485
On Wasserstein Two-Sample Testing and Related Families of Nonparametric Tests
Ramdas, A., Trillos, N., & Cuturi, M. (2017) · Entropy · doi:10.3390/e19020047 · Cited by 443
On Duality Theory of Conic Linear Problems
Shapiro, A. (2001) · Nonconvex optimization and its applications · doi:10.1007/978-1-4757-3403-4_7 · Cited by 423
Kullback-Leibler divergence constrained distributionally robust optimization
Hu, Z., & Hong, L. J. (2012) · Cited by 402
Multistage Stochastic Optimization
Pflug, G. C., & Pichler, A. (2014) · Springer series in operations research/Springer series in operations research and financial engineering · doi:10.1007/978-3-319-08843-3 · Cited by 384
The Optimizer’s Curse: Skepticism and Postdecision Surprise in Decision Analysis
Smith, J. E., & Winkler, R. L. (2006) · Management Science · doi:10.1287/mnsc.1050.0451 · Cited by 308
Ambiguous chance constrained problems and robust optimization
Erdoğan, E., & Iyengar, G. (2006) · Mathematical Programming · doi:10.1007/s10107-005-0678-0 · Cited by 302
On Complexity of Stochastic Programming Problems
Shapiro, A., & Nemirovski, A. (2006) · Applied Optimization · doi:10.1007/0-387-26771-9_4 · Cited by 301
Challenging the empirical mean and empirical variance: A deviation study
Catoni, O. (2012) · Annales de l'Institut Henri Poincaré, Probabilités et Statistiques · doi:10.1214/11-aihp454 · Cited by 298
Ambiguity in portfolio selection
Pflug, G., & Wozabal, D. (2007) · Quantitative Finance · doi:10.1080/14697680701455410 · Cited by 261
Quantitative Concentration Inequalities for Empirical Measures on Non-compact Spaces
Bolley, F., Guillin, A., & Villani, C. (2007) · Probability Theory and Related Fields · doi:10.1007/s00440-006-0004-7 · Cited by 248
Deriving robust counterparts of nonlinear uncertain inequalities
Ben-Tal, A., den Hertog, D., & Vial, J.-P. (2015) · Mathematical Programming · doi:10.1007/s10107-014-0750-8 · Cited by 246
Robust sample average approximation
Bertsimas, D., Gupta, V., & Kallus, N. (2018) · Mathematical Programming · doi:10.1007/s10107-017-1174-z · Cited by 244
The 1/N investment strategy is optimal under high model ambiguity
Pflug, G. C., Pichler, A., & Wozabal, D. (2012) · Journal of Banking & Finance · doi:10.1016/j.jbankfin.2011.07.018 · Cited by 220
Models for Minimax Stochastic Linear Optimization Problems with Risk Aversion
Bertsimas, D., Doan, X. V., Natarajan, K., & Teo, C.-P. (2010) · Mathematics of Operations Research · doi:10.1287/moor.1100.0445 · Cited by 214
Distributionally Robust Stochastic Programming
Shapiro, A. (2017) · SIAM Journal on Optimization · doi:10.1137/16m1058297 · Cited by 214
On the Relation Between Option and Stock Prices: A Convex Optimization Approach
Bertsimas, D., & Popescu, I. (2002) · Operations Research · doi:10.1287/opre.50.2.358.424 · Cited by 203
Representation Theorems in Hardy Spaces
Mashreghi, J. (2009) · Cambridge University Press eBooks · doi:10.1017/cbo9780511814525 · Cited by 189
Ambiguous Risk Measures and Optimal Robust Portfolios
Calafiore, G. C. (2007) · SIAM Journal on Optimization · doi:10.1137/060654803 · Cited by 165
Empirical risk minimization for heavy-tailed losses
Brownlees, C. T., Joly, É., & Lugosi, G. (2015) · The Annals of Statistics · doi:10.1214/15-aos1350 · Cited by 145
A framework for optimization under ambiguity
Wozabal, D. (2010) · Annals of Operations Research · doi:10.1007/s10479-010-0812-0 · Cited by 134
Block splitting for distributed optimization
Parikh, N., & Boyd, S. (2013) · Mathematical Programming Computation · doi:10.1007/s12532-013-0061-8 · Cited by 132
TRACTABLE ROBUST EXPECTED UTILITY AND RISK MODELS FOR PORTFOLIO OPTIMIZATION
Natarajan, K., Sim, M., & Uichanco, J. (2010) · Mathematical Finance · doi:10.1111/j.1467-9965.2010.00417.x · Cited by 131
Tests of goodness of fit based on the $L_2$-Wasserstein distance
Cuesta-Albertos, J. A., Matrán, C., Rodríguez-Rodríguez, J. M., & del Barrio, E. (1999) · The Annals of Statistics · doi:10.1214/aos/1017938923 · Cited by 129
A comment on “computational complexity of stochastic programming problems”
Hanasusanto, G. A., Kühn, D., & Wiesemann, W. (2015) · Mathematical Programming · doi:10.1007/s10107-015-0958-2 · Cited by 95
Robustifying Convex Risk Measures for Linear Portfolios: A Nonparametric Approach
Wozabal, D. (2014) · Operations Research · doi:10.1287/opre.2014.1323 · Cited by 90
Empirical risk minimization for heavy-tailed losses
Brownlees, C. T., & Joly, É. (2014) · Cited by 89
Simple Bounds for the Convergence of Empirical and Occupation Measures in 1-Wasserstein Distance
Boissard, E. (2011) · Electronic Journal of Probability · doi:10.1214/ejp.v16-958 · Cited by 87
Computationally Tractable Counterparts of Distributionally Robust Constraints on Risk Measures
Postek, K., den Hertog, D., & Melenberg, B. (2016) · SIAM Review · doi:10.1137/151005221 · Cited by 78
Models and algorithms for distributionally robust least squares problems
Mehrotra, S., & Zhang, H. (2013) · Mathematical Programming · doi:10.1007/s10107-013-0681-9 · Cited by 52
Robust Data-Driven Dynamic Programming
Hanasusanto, G. A., & Kühn, D. (2013) · Infoscience (Ecole Polytechnique Fédérale de Lausanne) · Cited by 38
Monotone Approximation of Decision Problems
Chehrazi, N., & Weber, T. A. (2010) · Operations Research · doi:10.1287/opre.1100.0814 · Cited by 25
Ambiguous probabilistic programs
Hu, Z., Hong, L. J., & So, A. M. (2013) · Cited by 8
Robust Sample Average ApproximationPreprint
Bertsimas, D., Gupta, V., & Kallus, N. (2014) · arXiv (Cornell University) · doi:10.48550/arxiv.1408.4445 · Cited by 8
Research works printed but not linked: 5
- Bertsekas, D.P.: Convex Optimization Theory. Athena Scientific, Belmont (2009)
- Kallenberg, O.: Foundations of Modern Probability, Probability and its Applications. Springer, New York (1997)
- Kantorovich, L.V., Rubinshtein, G.S.: On a space of totally additive functions. Vestn. Leningr. Univ. 13, 52-59 (1958)
- Rockafellar, R.T., Wets, R.J.-B.: Variational Analysis. Springer, Berlin (2010)
- Zhao, C.: Data-Driven Risk-Averse Stochastic Program and Renewable Energy Integration. PhD thesis, University of Florida (2014)
Cited by
Computational Optimal Transport with Applications to Data Sciences
Gabriel, P., & Marco, C. (2019) · Foundations and Trends® in Machine Learning · doi:10.1561/2200000073 · Cited by 2,319
Computational Optimal Transport
Peyré, G., & Cuturi, M. (2019) · now publishers, Inc. eBooks · doi:10.1561/9781680835519 · Cited by 847
Edge artificial intelligence for 6g: vision, enabling technologies, and applications
Letaief, K. B., Shi, Y., Lu, J., & Lu, J. (2022) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2021.3126076 · Cited by 829
“Distri-butionally robust optimization [273] provides a promising way to achieve worst-case probabilistic performance by incorporating all sample-generating distributions into an ambiguity set.”
BackgroundDistributionally Robust Neural Networks
Sagawa, S., Koh, P. W., Hashimoto, T., & Liang, P. (2020) · International Conference on Learning Representations · Cited by 780
Distributionally Robust Stochastic Optimization with Wasserstein Distance
Gao, R., & Kleywegt, A. (2023) · Mathematics of Operations Research · doi:10.1287/moor.2022.1275 · Cited by 630
Metadata
- Authors
- Mohajerin Esfahani, P., & Kuhn, D.
- Venue
- Mathematical Programming
- Published
- 2017-07-07
- Volume
- 171
- Issue
- 1-2
- Pages
- 115-166
- ISSN
- 0025-5610 (ISSN-L)
- 1436-4646
- OpenAlex
- W1712265143
- Cited by
- 136
- References
- 51
- Citations by year
- 2025: 6 · 2024: 16 · 2023: 19 · 2022: 10 · 2021: 15 · 2020: 22 · 2019: 14 · 2018: 19 · 2017: 10
- FWCI
- 15.75
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Locations
- Mathematical Programming (publishedVersion, cc-by)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- Delft University of Technology
- Topics
- Social Sciences › Decision Sciences › Management Science and Operations Research › Risk and Portfolio Optimization
- Physical Sciences › Mathematics › Numerical Analysis › Advanced Optimization Algorithms Research
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Optimal Power Flow Distribution
- Keywords
- distributionally robust optimization, Wasserstein distance, Wasserstein ball, stochastic programming, convex reformulation, linear programming reformulation, concentration measurement, uncertainty quantification, data-driven optimization
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.