On linear optimization over Wasserstein balls
From MaRDI portal
(Redirected from Publication:2089797)
Abstract: Wasserstein balls, which contain all probability measures within a pre-specified Wasserstein distance to a reference measure, have recently enjoyed wide popularity in the distributionally robust optimization and machine learning communities to formulate and solve data-driven optimization problems with rigorous statistical guarantees. In this technical note we prove that the Wasserstein ball is weakly compact under mild conditions, and we offer necessary and sufficient conditions for the existence of optimal solutions. We also characterize the sparsity of solutions if the Wasserstein ball is centred at a discrete reference measure. In comparison with the existing literature, which has proved similar results under different conditions, our proofs are self-contained and shorter, yet mathematically rigorous, and our necessary and sufficient conditions for the existence of optimal solutions are easily verifiable in practice.
Recommendations
- Data-driven distributionally robust optimization using the Wasserstein metric: performance guarantees and tractable reformulations
- Conic programming reformulations of two-stage distributionally robust linear programs over Wasserstein balls
- Regularization for Wasserstein distributionally robust optimization
- Distributions with maximum spread subject to Wasserstein distance constraints
- Decomposition algorithm for distributionally robust optimization using Wasserstein metric with an application to a class of regression models
Cites work
- A framework for optimization under ambiguity
- Ambiguity in portfolio selection
- Data-driven distributionally robust optimization using the Wasserstein metric: performance guarantees and tractable reformulations
- Data-driven risk-averse stochastic optimization with Wasserstein metric
- Extreme points of a ball about a measure with finite support
- scientific article; zbMATH DE number 42272 (Why is no real title available?)
- scientific article; zbMATH DE number 765034 (Why is no real title available?)
- scientific article; zbMATH DE number 3274494 (Why is no real title available?)
- Infinite dimensional analysis. A hitchhiker's guide.
- Lectures on Stochastic Programming
- Measure theory. Vol. I and II
- On Choosing and Bounding Probability Metrics
- On the extreme points of moments sets
- Optimal Transport
- Quantifying distributional model risk via optimal transport
- Quantitative stability analysis for minimax distributionally robust risk optimization
- Wasserstein metric and subordination
Cited in
(11)- Markov decision processes under model uncertainty
- Distributionally robust optimization using optimal transport for Gaussian mixture models
- Nonconcave stochastic optimal control in finite discrete time under model uncertainty
- Sensitivity of functionals of McKean-Vlasov SDEs with respect to the initial distribution
- Robust approaches in portfolio optimization with stochastic dominance constraints
- Distributionally robust statistical verification with imprecise neural networks
- ALSO-X \#: better convex approximations for distributionally robust chance constrained programs
- Distributionally robust optimization
- First-order conditions for optimization in the Wasserstein space
- Convergent lifted Lasserre hierarchy of SDPs for minimizing expectation of piecewise polynomial loss over Wasserstein balls
- Algorithms for mean-field variational inference via polyhedral optimization in the Wasserstein space
This page was built for publication: On linear optimization over Wasserstein balls
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2089797)
