betaFeedback

WorksMathematics

Geometry, Local Minimax Risk, and the Limits of Sample Average Approximation in Linear Stochastic Optimization

Abstract

We study how the local geometry of a linear stochastic optimization problem determines its statistical difficulty, measured by expected decision regret. The optimal local rate is governed by how fast the support function of the feasible set grows away from its linearization at the true mean; depending on the instance, it can be n−1/2n^{-1/2}, n−1n^{-1}, n−2n^{-2}, or a non-power rate, and sample average approximation (SAA) attains these rates in a Gaussian mean model. At smooth instances, SAA also attains the classical sharp local minimax constant. We then allow the data distribution to vary in a Kullback–Leibler (KL) ball of radius proportional to 1/n1/n around a reference distribution PP. The exact local minimax limit of this benchmark, which may depend on PP, coincides with that of a Gaussian model whose local parameter set is the covariance ellipsoid of PP. A simple coordinatewise truncated SAA rule, which does not use PP, attains the minimax rate, and at smooth instances its constant matches the minimax constant as the neighborhood radius grows. Ordinary SAA, in contrast, can have strictly suboptimal worst-case expected regret when the local rate is fast relative to the tails of PP, because rare tail events then dominate the expectation. The results delineate when SAA is sharply optimal, rate-optimal, or statistically suboptimal, and quantify the local price of distributional robustness.

Introduction

Sample average approximation (SAA) replaces an unknown expectation by its empirical counterpart and then solves the resulting optimization problem. It is model-free, com putationally transparent, and supported by a mature theory of consistency, asymptotic distributions, and finite-sample guarantees; see, for example, Shapiro [31] and the monograph treatment in Shapiro et al. [33]. A more demanding question is whether SAA is statistically optimal for the particular optimization instance and uncertainty model at hand. Standard worst-case rates do not answer this question. The attainable regret can vary sharply with the local geometry of the feasible region, and the optimality of the empirical mean can additionally depend on how distributional perturbations interact with the tails of the data.

We study these issues in the canonical linear stochastic optimization problem

max⁡w∈ΩwTμ,(1)\max_{w \in\Omega} \mathbf{w}^{\mathsf{T}}\boldsymbol{\mu}, \tag*{(1)}

where Ω⊂Rd\Omega\subset\mathbb{R}^{d} is compact and convex and μ\boldsymbol{\mu} is unknown. Its optimal value is the support function

f∗(z)=sup⁡w∈ΩwTz,f_{*}(z) = \sup_{w \in\Omega} \mathbf{w}^{\mathsf{T}}z,

and we write w∗(z)\mathbf{w}_{*}(z) for a selected maximizer. Given nn independent observations, x1,…,xn\boldsymbol{x}_{1}, \ldots, \boldsymbol{x}_{n}, with mean μ\boldsymbol{\mu}, ordinary SAA is the plug-in policy w∗(xˉ)\mathbf{w}_{*}(\bar{\boldsymbol{x}}) where xˉ\bar{\boldsymbol{x}} is the sample mean. We evaluate a data-dependent policy through decision regret and compare it with all measurable sample-based policies.

Our analysis considers two local statistical experiments. We first study Gaussian observations with covariance Σ\Sigma and local mean perturbations μ+t/n\boldsymbol{\mu} + \boldsymbol{t}/\sqrt{n}. We then allow the entire data-generating distribution to vary in a Kullback–Leibler neighborhood of a reference distribution PP,

Qn,MKL(P)={Q≪P:EQ∥x∥2<∞, DKL(Q∥P)≤M22n}.\mathcal{Q}_{n,M}^{\mathrm{KL}}(P) = \left\{Q \ll P : \mathbb{E}_{Q}\lVert x\rVert_{2} < \infty,\ D_{\mathrm{KL}}(Q\mathbin{\|}P) \le\frac{M^{2}}{2n}\right\}.

The second experiment is locally distributionally robust: it permits changes in the mean, covariance, and tails, subject only to a shrinking entropy budget. In both experiments, the central deterministic object is the local support-function remainder

rμ(h)=f∗(μ+h)−f∗(μ)−w∗(μ)Th.r_{\boldsymbol{\mu}}(h) = f_{*}(\boldsymbol{\mu} + h) - f_{*}(\boldsymbol{\mu}) - \mathbf{w}_{*}(\boldsymbol{\mu})^{\mathsf{T}}h.

Statistical uncertainty determines the scale and directions in which μ\boldsymbol{\mu} can be learned; the local growth of rμr_{\boldsymbol{\mu}} converts that uncertainty into regret.

This viewpoint yields both an optimality result and a limitation of SAA. In the Gaussian experiment, SAA adapts to the local geometry and attains the local minimax rate: the regret is of order n−1n^{-1} at ordinary smooth instances and n−1/2n^{-1/2} at nonsmooth instances, and degenerate curvature can produce faster rates such as n−2n^{-2} or non-power rates. On KL neighborhoods of radius 1/n1/n, the exact local minimax benchmark is governed by the same support-function geometry, and a coordinatewise truncated version of SAA attains the minimax rate. Ordinary SAA, however, need not attain even the rate: rare but KL-admissible tail perturbations can dominate its expected regret when the local geometric rate is faster than the available tail control for the empirical mean. Thus there is no universal answer to whether SAA is optimal. It can be sharply optimal, merely rate-optimal, or strictly suboptimal, depending jointly on local geometry and distributional tails.

Main contributions

More specifically, our results can be summarized as follows.

General geometry and exact Gaussian local risk. If

rμ(εt)=a(ε)ψ(t)+o(a(ε))r_{\mu}(\varepsilon t) = a(\varepsilon)\psi(t) + o(a(\varepsilon))

uniformly on bounded sets, the Gaussian local minimax regret, normalized by a(n−1/2)a(n^{-1/2}), converges to the Bayes Jensen-gap functional VM,Σ(ψ)V_{M,\Sigma}(\psi) in (26). When ψ\psi is nonaffine on the local parameter ball, this constant is positive and SAA attains the rate. This covers nonsmooth, quadratic, degenerate, and non-power geometry in a common decision-theoretic representation. One fixed strongly convex feasible set exhibits rates n−1/2n^{-1/2}, n−1n^{-1}, and n−2n^{-2} at different centers; a separate compact convex set gives 1/(nlog⁡n)1/(\sqrt{n}\log n). Finite-sample upper and lower bounds in terms of a symmetric local optimization modulus dispense with a local expansion altogether.

Exact local minimax risk on KL neighborhoods. For the distributional neighborhood Qn,MKL(P)Q^{\mathrm{KL}}_{n,M}(P), we establish the exact local minimax limit

Rn,MKL(P)a(n−1/2)⟶VM,Id(ψ∘Σ1/2),\frac{R^{\mathrm{KL}}_{n,M}(P)}{a(n^{-1/2})} \longrightarrow V_{M,\mathrm{Id}}\left(\psi\circ\Sigma^{1/2}\right),

where VM,IdV_{M,\mathrm{Id}} is given by (26) and Σ\Sigma is the covariance matrix of xx under PP, under a local exponential-moment condition; the matching lower bound requires only a finite second moment and nonsingular covariance. The rule attaining the exact constant uses knowledge of PP and MM; the truncated mean introduced below does not, and attains the rate. The same limit arises in a matched Gaussian experiment with local alternatives μ+Σ1/2u/n\mu+\Sigma^{1/2}u/\sqrt{n}, so the KL neighborhood is asymptotically summarized by the covariance ellipsoid of the reference distribution while the loss is summarized by the local geometry ψ\psi. Whenever the limiting local modulus is nonaffine, a coordinatewise truncated empirical mean, followed by optimization, is rate-minimax. At smooth instances, its large-local-ball constant reduces to tr⁡(Σ∇2f∗(μ))/2\operatorname{tr}(\Sigma\nabla^{2}f_{*}(\mu))/2, which is also the local minimax constant. In this sense, local distributional robustness need not change first-order efficiency, although it changes the class of distributions against which an estimator must perform uniformly.

When ordinary SAA is suboptimal. We compare ordinary SAA with the truncated plug-in policy uniformly over the KL neighborhood. A geometry-weighted tail criterion supplies sufficient conditions for rate-optimality and for equality of the leading maximal regrets of ordinary and truncated SAA. For power-type geometry a(ε)=εαa(\varepsilon)=\varepsilon^\alpha and a reference distribution satisfying an exponential-power moment condition EPexp⁡(c∥x∥2ζ)<∞\mathbb{E}_{P}\exp(c\lVert x\rVert_{2}^{\zeta})<\infty, ordinary SAA has maximal regret bounded by

O(n−α/2+n−ζ).O(n^{-\alpha/2}+n^{-\zeta}).

It is rate-optimal when α≤2ζ\alpha\leq2\zeta, and it has the same leading maximal-regret constant as truncated SAA when α<2ζ\alpha<2\zeta. Matching examples establish rate sharpness over the stated class. In particular, for Laplace tails (ζ=1\zeta=1) and quartic local geometry (α=4\alpha=4), the local minimax and truncated-SAA risks are of order n−2n^{-2}, whereas ordinary SAA has worst-case regret of order n−1n^{-1}. The failure is not visible from the Gaussian weak limit of the sample mean; it arises from lack of uniform integrability under rare distributional perturbations.

Smooth efficiency as a special case. At a C2C^{2} instance, a(ε)=ε2a(\varepsilon)=\varepsilon^{2} and ψ(t)=tTHμt/2\psi(t)=t^{\mathsf{T}}H_{\mu}t/2, where Hμ=∇2f∗(μ)H_{\mu}=\nabla^{2}f_{*}(\mu). The expanding-ball Gaussian minimax constant is tr⁡(ΣHμ)/2\operatorname{tr}(\Sigma H_{\mu})/2, attained by SAA without knowledge of Σ\Sigma. Under the boundary regularity conditions in § 4, this equals

12∥μ∥22∑i=1d−1uiTΣuiκi(μ).\frac{1}{2\lVert\mu\rVert_{2}^{2}}\sum_{i=1}^{d-1}\frac{\mathbf{u}_{i}^{\mathsf{T}}\Sigma\mathbf{u}_{i}}{\kappa_{i}(\mu)}.

where ui\mathbf{u}_{i} and κi(μ)\kappa_{i}(\mu) are the principal tangent directions and curvatures. Thus tangent noise is weighted by inverse curvature. Truncated SAA has the same smooth efficiency constant on KL neighborhoods after the local radius tends to infinity.

Relation to existing literature

Classical SAA asymptotics. The statistical theory of SAA establishes consistency, asymptotic distributions of optimal values and solutions, and sample-size guarantees under increasingly general stochastic-programming models; representative references include Shapiro [31, 32]; Shapiro et al. [33]. Our focus differs in two respects. First, the target is regret rather than estimation error of the optimizer or optimal value. Second, we compare SAA with the local minimax risk over all measurable policies. This asks an efficiency question rather than only deriving the sampling distribution of SAA. The support-function structure of (1) makes it possible to identify sharp constants, characterize nonquadratic rates, and distinguish weak convergence from convergence of expected regret under distributional perturbations.

Local minimax and instance-optimal stochastic optimization. The closest conceptual antecedent is Duchi and Ruan (2021), who develop a Hájek–Le Cam-type local minimax theory for stochastic convex optimization and relate statistical difficulty to tilt stability, together with asymptotically optimal online algorithms. Our linear setting permits a different level of geometric resolution. We obtain an exact Bayes–minimax representation in terms of the support-function remainder, sharp constants expressed through principal curvatures, finite-sample modulus bounds, and a theory of degenerate and non-power rates. We also extend the local analysis from a Gaussian mean experiment to population-centered KL neighborhoods and identify when the empirical-mean implementation ceases to attain the benchmark. The smooth trace constant serves as the classical benchmark; the additional conclusions are the exact fixed-radius Jensen-gap limit for general local support growth, the corresponding full-KL minimax limit, and a separation of regret rates caused by rare tails.

A recent and complementary contribution is Jiang and Pananjady [20], who study finite-sample instance optimality for unconstrained smooth strongly convex stochastic optimization with additive and multiplicative oracle noise. They show that SAA and averaged stochastic approximation can be substantially suboptimal at realistic sample sizes and construct an instance-optimal variance-reduced method. Their loss, oracle model, and finite-sample objective differ from ours. Together, the two perspectives emphasize that the optimality of SAA is experiment-, loss-, and instance-dependent.

Distributionally robust optimization and shrinking divergence neighborhoods. KL and more general divergence neighborhoods have been used both to quantify model sensitivity and to construct statistically calibrated robust procedures. Lam [21] derives infinitesimal expansions for worst-case performance over shrinking KL neighborhoods. Lam [22] and Duchi et al. [7] connect empirical divergence-based DRO with empirical likelihood, asymptotically calibrated confidence bounds, and variance regularization, while He and Lam [17] develop higher-order expansions and Bartlett corrections for divergence-based DRO. From a large-deviations perspective, Van Parys et al. [37] show that relative-entropy DRO is optimal for controlling out-of-sample disappointment. Wasserstein and other data-driven ambiguity sets, which are centered at the empirical distribution, are studied in Mohajerin Esfahani and Kuhn [29]; Bertsimas et al. [3]; Gao and Kleywegt [12]. Complementary work asks whether robust or optimistic reformulations can improve the out-of-sample behavior of SAA in particular model classes; see Anderson and Philpott [2] and Gotoh et al. [13].

Our KL neighborhood is centered at a fixed reference distribution rather than the empirical distribution, and our target is different. We characterize the minimax decision regret over all measurable policies, not the value of a particular robust optimization formulation or the coverage of an uncertainty set. The radius 1/n1/n generates a local asymptotic experiment in which covariance and support geometry determine the minimax benchmark. The comparison between ordinary and truncated SAA then exposes an additional phenomenon: two procedures can have the same Gaussian weak limit but different maximal expected regret because KL-admissible rare events affect uniform integrability. This is the mechanism behind the suboptimality of ordinary SAA at sufficiently fast geometric rates.

Predict-then-optimize and decision-focused learning. For linear downstream optimization, the smart predict-then-optimize framework of Elmachtoub and Grigas [9] evaluates predictions through induced decision regret rather than prediction error. Generalization analyses such as El Balghiti et al. [8] highlight the role of margins around parameter values that lead to nonunique decisions and introduce geometric strength conditions on the feasible set. Our ambiguity set plays a related geometric role, but our question is different: we fix a local statistical experiment and characterize its minimax regret, including exact constants and the effect of distributional tails, rather than bounding generalization error over prediction classes.

Nonsmooth functional and norm estimation. A separate literature studies minimax estimation of nonsmooth functionals and norms in Gaussian models. The connection is direct: for p⋆=p/(p−1)p^\star= p/(p-1) and Ω=Bp⋆d={w:∥w∥p⋆≤1}\Omega= B_{p^\star}^d = \{w : \|w\|_{p^\star} \le1\}, we have f⋆(μ)=∥μ∥pf_\star(\mu) = \|\mu\|_p, so estimating an ℓp\ell_p-norm is estimating the optimal value of the same stochastic optimization problem. Relevant work includes Lepski et al. [27] on LrL_r-norms in nonparametric Gaussian models, Cai and Low [4] on the ℓ1\ell_1 functional in the Gaussian sequence model, and the extensions to general LrL_r-norms (Han et al., [16]) and to ℓγ\ell_\gamma-functionals over sparse normal means (Collier et al., [5]).

Our target is different: we evaluate a data-dependent decision through regret rather than estimating f⋆(μ)f_\star(\mu) or w⋆(μ)w_\star(\mu). Regret removes the affine first-order component of the value functional, so the local remainder becomes the governing object, and value estimation and decision regret can have different local orders. At a differentiable nonzero point, ∥μ∥p\|\mu\|_p is a regular scalar functional with root-nn estimation error, whereas decision regret is typically second order. At degenerate points the contrast is larger: for p>2p > 2 and μ=re1\mu= r e_1, the value remains first-order regular in the radial direction, while the transverse decision regret has order n−p/2n^{-p/2}. These faster-order instances are also where tail stabilization can become necessary under KL robustness; see Examples 4.10 and 6.5.

Empirical optimization, model structure, and ETO/IEO. A closely related line of work led by Lam asks when empirical optimization can be statistically improved. Lam [24] proves, under suitable smoothness with respect to the data-generating distribution, that the first-order optimality-gap behavior of alternative data-driven formulations is second-order stochastically dominated by empirical optimization. Our smooth, nondegenerate results are consistent with this efficiency principle: ordinary SAA is sharply optimal in the quadratic regime of the Gaussian experiment. Our KL counterexamples delineate a different regime. When the leading quadratic curvature vanishes and a faster regret rate is attainable, rare tail perturbations can make the raw empirical mean suboptimal even though a truncated empirical optimizer remains rate-minimax.

Other work identifies mechanisms that improve on SAA through additional structure. Lam and Wang [24] show that first-order improvements can arise when side information is correctly specified and geometrically effective. The comparison of SAA with estimate-then-optimize (ETO) and integrated-estimation-optimization (IEO) is similar in spirit: Elmachtoub et al. [11] show that under correct parametric specification ETO asymptotically dominates IEO in the sense of stochastic dominance of regret and, when the comparison applies, can also dominate SAA, whereas under misspecification the ordering can reverse. Elmachtoub et al. [10] isolate the impact of misspecification, and Lan et al. [25] study local misspecification through contiguity and derive an explicit bias–variance tradeoff among SAA, ETO, and IEO. In these works the gain comes from model structure that reduces effective estimation variance. The mechanism in our KL results is different: tail stabilization can improve maximal expected regret without altering the Gaussian local information. Consequently, no unconditional ranking among data-driven procedures should be expected; which procedure is preferable depends on whether model structure or tail behavior is the binding constraint.

Organization

§ 2 contrasts the global Gaussian-width benchmark with a local quadratic-growth bound. § 3 develops the sharp Gaussian local minimax theory, including the Bayes representation, smooth efficiency bound, general local scales, and a finite-sample optimization modulus. § 4 translates these statistical results into boundary curvature, ambiguity, and contact geometry, and illustrates the resulting range of rates. § 5 first introduces a matched ellipsoidal Gaussian benchmark, then establishes the exact local minimax risk over shrinking KL neighborhoods and quantifies the price of robustness. § 6 compares ordinary and truncated SAA and gives sharp conditions for optimality or suboptimality. § 7 illustrates the Gaussian rates and the geometry–tail transition numerically. § 8 concludes. All proofs, intermediate technical results, and detailed calculations are collected in the Online Supplement. Table 2 in § 6 summarizes the benchmarks and procedure-specific risks used throughout.

QuantityStatistical experimentLocal parameter setRule may use
Rn,MΣ(μ)R_{n,M}^{\Sigma}(\mu)Gaussian, fixed covarianceμ+B2(M)/n\mu+ B_2(M)/\sqrt{n}benchmark center and fixed Σ\Sigma
Rn,Mell,Σ(μ)R_{n,M}^{\mathrm{ell},\Sigma}(\mu)Matched Gaussianμ+Σ1/2B2(M)/n\mu+ \Sigma^{1/2}B_2(M)/\sqrt{n}benchmark center and fixed Σ\Sigma
Rn,MKL(P)R_{n,M}^{\mathrm{KL}}(P)KL neighborhoodDKL(Q∥P)≤M2/(2n)D_{\mathrm{KL}}(Q\mathbin\Vert P) \le M^2/(2n)reference law PP, but not QQ
Wn,Mmean,Wn,MtrW_{n,M}^{\mathrm{mean}}, W_{n,M}^{\mathrm{tr}}KL maximal riskssame KL neighborhoodprescribed ordinary or truncated plug-in rule

Table 2. Statistical benchmarks and procedure-specific maximal risks used in the paper.

Global Complexity and Local Geometric Fast Rates

Problem setup

Let Ω⊂Rd\Omega\subset\mathbb{R}^{d} be a compact convex set containing 0\mathbf{0}, and define its support function by

f∗(z):=max⁡w∈Ωw⊤z,z∈Rd.(2)f_{*}(z) := \max_{w \in\Omega} w^{\top}z,\qquad z \in\mathbb{R}^{d}. \tag*{(2)}

For each zz, let

W∗(z):=arg⁡max⁡w∈Ωw⊤z,\mathcal{W}_{*}(z) := \arg\max_{w \in\Omega} w^{\top}z,

The measurable maximum theorem gives a Borel measurable selection; fix one and denote it by w∗(z)∈W∗(z)w_{*}(z) \in\mathcal{W}_{*}(z) [1 Theorem 18.19]. The ambiguity set is

A:={z∈Rd:W∗(z) is not a singleton}.(3)A := \{z \in\mathbb{R}^{d}: \mathcal{W}_{*}(z) \text{ is not a singleton}\}. \tag*{(3)}

References

  1. [1]Charalambos D. Aliprantis and Kim C. Border. Infinite Dimensional Analysis: A Hitch-hiker’s Guide. Springer, Berlin, third edition, 2006.
  2. [2]Edward Anderson and Andy Philpott. Improving sample average approximation using distributional robustness. INFORMS Journal on Optimization, 4(1):90–124, 2022. doi: 10.1287/ijoo.2021.0061.DOI
  3. [3]Dimitris Bertsimas, Vishal Gupta, and Nathan Kallus. Data-driven robust optimization. Mathematical Programming, 167(2):235–292, 2018.
  4. [4]T. Tony Cai and Mark G. Low. Testing composite hypotheses, Hermite polynomials and optimal estimation of a nonsmooth functional. The Annals of Statistics, 39(2):1012–1041, 2011. doi: 10.1214/10-AOS849.DOI
  5. [5]Olivier Collier, Laëtitia Comminges, and Alexandre B. Tsybakov. On estimation of nonsmooth functionals of sparse normal means. Bernoulli, 26(3):1989–2020, 2020. doi: 10.3150/19-BEJ1180.DOI
  6. [7]John C. Duchi, Peter W. Glynn, and Hongseok Namkoong. Statistics of robust optimization: A generalized empirical likelihood approach. Mathematics of Operations Research, 46(3):946–969, 2021. doi: 10.1287/moor.2020.1085.DOI
  7. [8]Othman El Balghiti, Adam N. Elmachtoub, Paul Grigas, and Ambuj Tewari. Generalization bounds in the predict-then-optimize framework. Mathematics of Operations Research, 48(4):2043–2065, 2023. doi: 10.1287/moor.2022.1330.DOI
  8. [9]Adam N. Elmachtoub and Paul Grigas. Smart “predict, then optimize”. Management Science, 68(1):9–26, 2022. doi: 10.1287/mnsc.2020.3922.DOI
  9. [10]Adam N. Elmachtoub, Henry Lam, Haixiang Lan, and Haofeng Zhang. Dissecting the impact of model misspecification in data-driven optimization. In Proceedings of the 28th International Conference on Artificial Intelligence and Statistics, volume 258 of Proceedings of Machine Learning Research, pages 1594–1602. PMLR, 2025a.arxiv.org/abs/2503.00626
  10. [11]Adam N. Elmachtoub, Henry Lam, Haofeng Zhang, and Yunfan Zhao. Estimate-then-optimize versus integrated-estimation-optimization versus sample average approximation: A stochastic dominance perspective, 2025b. arXiv:2304.06833, version 4.arxiv.org/abs/2304.06833
  11. [12]Rui Gao and Anton Kleywegt. Distributionally robust stochastic optimization with Wasserstein distance. Mathematics of Operations Research, 48(2):603–655, 2023.arxiv.org/abs/1604.02199
  12. [13]Jun-ya Gotoh, Michael Jong Kim, and Andrew E. B. Lim. Technical note—a data-driven approach to beating SAA out of sample. Operations Research, 73(2):829–841, 2025. doi: 10.1287/opre.2021.0393.
  13. [14]Vishal Gupta and Paat Rusmevichientong. Small-data, large-scale linear optimization with uncertain objectives. Management Science, 67(1):220–241, 2021.DOI
  14. [15]Jaroslav Hájek. Local asymptotic minimax and admissibility in estimation. In Proceedings of the Sixth Berkeley Symposium on Mathematical Statistics and Probability, Volume 1, pages 175–194. University of California Press, 1972.DOI
  15. [16]Yanjun Han, Jiantao Jiao, Rajarshi Mukherjee, and Tsachy Weissman. On estimation of Lᵣ-norms in Gaussian white noise models. Probability Theory and Related Fields, 177:1243–1294, 2020. doi: 10.1007/s00440-020-00982-x.DOI
  16. [17]Shengyi He and Henry Lam. Higher-order expansion and Bartlett correctability of distributionally robust optimization. Mathematics of Operations Research, 51(2):1538–1584, 2026. doi: 10.1287/moor.2023.0191.DOI
  17. [18]Ralph Howard and Daniel Hug. Smooth convex bodies with proportional projection functions. Israel Journal of Mathematics, 159:317–341, 2007.DOI
  18. [19]W. James and Charles Stein. Estimation with quadratic loss. In Proceedings of the Fourth Berkeley Symposium on Mathematical Statistics and Probability, Volume 1, pages 361–379. University of California Press, 1961.
  19. [20]Liwei Jiang and Ashwin Pananjady. Instance-optimal stochastic convex optimization: Can we improve upon sample-average and robust stochastic approximation?, 2026. arXiv:2603.25657.arxiv.org/abs/2603.25657
  20. [21]Henry Lam. Robust sensitivity analysis for stochastic systems. Mathematics of Operations Research, 41(4):1248–1275, 2016. doi: 10.1287/moor.2015.0776.DOI
  21. [22]Henry Lam. Recovering best statistical guarantees via the empirical divergence-based distributionally robust optimization. Operations Research, 67(4):1090–1105, 2019. doi: 10.1287/opre.2018.1786.DOI
  22. [24]Henry Lam and Tianyu Wang. Achieving first-order statistical improvements in data-driven optimization: From no-free-lunch to amplified decision perturbation, 2026. arXiv:2608.04312.arxiv.org/abs/2608.04312
  23. [25]Haixiang Lan, Luofeng Liao, Adam N. Elmachtoub, Christian Kroer, Henry Lam, and Haofeng Zhang. The bias-variance tradeoff in data-driven optimization: A local misspecification perspective. In Advances in Neural Information Processing Systems, volume 38, 2025. doi: 10.52202/085713-2403.
  24. [26]Lucien Le Cam. Limits of experiments. In Proceedings of the Sixth Berkeley Symposium on Mathematical Statistics and Probability, Volume 1, pages 245–261. University of California Press, 1972.DOI
  25. [27]Oleg Lepski, Arkadi Nemirovski, and Vladimir Spokoiny. On estimation of the Lᵣ norm of a regression function. Probability Theory and Related Fields, 113(2):221–253, 1999.DOI
  26. [28]Friedrich Liese and Klaus-J. Miescke. Statistical Decision Theory: Estimation, Testing, and Selection. Springer, New York, 2008. doi: 10.1007/978-0-387-73194-0.DOI
  27. [29]Peyman Mohajerin Esfahani and Daniel Kuhn. Data-driven distributionally robust optimization using the Wasserstein metric: Performance guarantees and tractable reformulations. Mathematical Programming, 171(1–2):115–166, 2018.arxiv.org/abs/1505.05116
  28. [30]Haskell P. Rosenthal. On the subspaces of Lᵖ (p > 2) spanned by sequences of independent random variables. Israel Journal of Mathematics, 8:273–303, 1970.DOI
  29. [31]Alexander Shapiro. Asymptotic behavior of optimal solutions in stochastic programming. Mathematics of Operations Research, 18(4):829–845, 1993. doi: 10.1287/moor.18.4.829.DOI
  30. [32]Alexander Shapiro. Asymptotics of minimax stochastic programs. Statistics & Probability Letters, 78(2):150–157, 2008. doi: 10.1016/j.spl.2007.05.012.DOI
  31. [33]Alexander Shapiro, Darinka Dentcheva, and Andrzej Ruszczyński. Lectures on Stochastic Programming: Modeling and Theory. MOS-SIAM Series on Optimization. Society for Industrial and Applied Mathematics and Mathematical Optimization Society, third edition, 2021. doi: 10.1137/1.9781611976595.
  32. [34]James E. Smith and Robert L. Winkler. The optimizer’s curse: Skepticism and postdecision surprise in decision analysis. Management Science, 52(3):311–322, 2006.DOI
  33. [35]Charles Stein. Inadmissibility of the usual estimator for the mean of a multivariate normal distribution. In Proceedings of the Third Berkeley Symposium on Mathematical Statistics and Probability, Volume 1, pages 197–206. University of California Press, 1956.DOI
  34. [36]A. W. van der Vaart. Asymptotic Statistics. Cambridge University Press, Cambridge, 1998.
  35. [37]Bart P. G. Van Parys, Peyman Mohajerin Esfahani, and Daniel Kuhn. From data to decisions: Distributionally robust optimization is optimal. Management Science, 67(6):3387–3402, 2021. doi: 10.1287/mnsc.2020.3678.DOI
  36. [38]Bengt von Bahr and Carl-Gustav Esseen. Inequalities for the rth absolute moment of a sum of random variables, 1 ≤ r ≤ 2. The Annals of Mathematical Statistics, 36(1):299–303, 1965.DOI

Paper details

vlnc:W_7sda42fk
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 6

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents