betaFeedback

WorksMathematics

Moment approximations in optimal transport: geometric estimates and a semiclassical limit

Abstract

We study finite moment approximations of linear optimization problems over measures. Under suitable assumptions, we establish the existence of discrete minimizers with a controlled number of support points, duality and ΓΓ-convergence as the moment approximation is refined. This framework accommodates several transport problems, in which marginal constraints are replaced by finitely many moment constraints. For moment functions localized at scale rnr_n, we prove an estimate controlling W∞W_\infty by WpW_p up to an additive error of order rnr_n, which also applies to discrete measures. For repulsive costs, we show that minimizers remain separated from the diagonal, uniformly in the number of moments, and we derive consequences for dual potentials. Finally, under a moment-correction assumption, we prove that the fermionic minimum energies converge in the semiclassical limit to the moment-constrained Coulomb transport value for a fixed family of moment functions.

Introduction

The numerical approximation of optimal transport problems becomes particularly difficult when several marginals are prescribed. A common approach consists in approximating each marginal by a discrete measure and solving the resulting finite-dimensional problem. If each of MM marginals is supported on nn points, the transport plan has up to nMn^M possible configurations. This difficulty motivates approximations that allow the support of the transport plan to be sparse [2, 1].

In [2], the authors study one such approximation, which consists in replacing the marginal constraints by finitely many moment constraints. The prescribed marginals enter the problem through their integrals against a chosen family of test functions, called moment functions. Under suitable assumptions, the relaxed problem admits a discrete minimizer whose number of support points is controlled by the number of constraints. Both the positions of these points and their weights can then be treated as variables in the approximation.

The choice of moment functions is consequently an essential part of the approximation. The relaxed constraints allow measures whose marginals differ from the prescribed ones, and it is necessary to understand what information about those marginals is retained. In this paper, we study finite moment approximations of linear optimization problems over measures within a common framework, and we investigate which geometric information on the marginals is retained by the relaxation. To fix ideas, we first describe the approximation in the Monge–Kantorovich case.

Let XX and YY be Polish spaces, and let μ∈P(X)\mu\in\mathcal{P}(X) and ν∈P(Y)\nu\in\mathcal{P}(Y). Given a lower semicontinuous cost c:X×Y→R∪{+∞}c:X \times Y \to\mathbb{R} \cup\{+\infty\}, bounded from below, the Monge–Kantorovich problem consists in minimizing

C(π)=∫X×Yc(x,y) dπ(x,y)C(\pi) = \int_{X \times Y} c(x,y)\,\mathrm{d}\pi(x,y)

among probability measures π\pi whose first and second marginals are μ\mu and ν\nu, respectively. We denote the set of all such measures by Π(μ,ν)\Pi(\mu,\nu). Equivalently, these marginal constraints can be expressed using test functions:

∫X×Yφ(x) dπ(x,y)=∫Xφ dμfor every φ∈Cb(X),\int_{X \times Y} \varphi(x)\,\mathrm{d}\pi(x,y) = \int_X \varphi\,\mathrm{d}\mu\quad\text{for every } \varphi\in C_b(X),
∫X×Yψ(y) dπ(x,y)=∫Yψ dνfor every ψ∈Cb(Y).\int_{X \times Y} \psi(y)\,\mathrm{d}\pi(x,y) = \int_Y \psi\,\mathrm{d}\nu\quad\text{for every } \psi\in C_b(Y).

The moment relaxation consists in requiring these identities only for selected finite families of test functions. For each n∈Nn \in\mathbb{N}, let In⊂Cb(X)\mathcal{I}^n \subset C_b(X) and Jn⊂Cb(Y)\mathcal{J}^n \subset C_b(Y) be finite families of moment functions. We denote by Πn(μ,ν)\Pi^n(\mu,\nu) the set of probability measures satisfying the preceding identities for every φ∈In\varphi\in\mathcal{I}^n and ψ∈Jn\psi\in\mathcal{J}^n. The moment-constrained Monge–Kantorovich problem is then

inf⁡π∈Πn(μ,ν)C(π).(1)\inf_{\pi\in\Pi^n(\mu,\nu)} C(\pi). \tag*{(1)}

Since Π(μ,ν)⊆Πn(μ,ν)\Pi(\mu,\nu) \subseteq\Pi^n(\mu,\nu), the infimum of the relaxed problem is no larger than that of the original problem; when the moment functions are refined so as to fully characterize μ\mu and ν\nu, the relaxed values converge to the original one (see Theorem 2.13).

Even when the relaxation preserves the optimal value, it may admit minimizers with very different supports, as the following example shows.

Example 1.1. Let X=Y=[0,1]X = Y = [0,1], consider the attractive cost c(x,y)=∣x−y∣2c(x,y) = \lvert x-y\rvert^2, and take

μ=ν=12δ1/4+12δ3/4.\mu= \nu= \frac{1}{2}\delta_{1/4} + \frac{1}{2}\delta_{3/4}.

The unique minimizer of the original problem is

π⋆=12δ(1/4,1/4)+12δ(3/4,3/4).\pi^\star= \frac{1}{2}\delta_{(1/4,1/4)} + \frac{1}{2}\delta_{(3/4,3/4)}.

For I3=J3={x↦xk:k=1,2,3}\mathcal{I}^3 = \mathcal{J}^3 = \{x \mapsto x^k : k = 1,2,3\}, a minimizer of the moment-constrained problem is

π3=18δ(0,0)+34δ(1/2,1/2)+18δ(1,1),C(π3)=0.\pi^3 = \frac{1}{8}\delta_{(0,0)} + \frac{3}{4}\delta_{(1/2,1/2)} + \frac{1}{8}\delta_{(1,1)}, \qquad C(\pi^3) = 0.

Both its marginals are

18δ0+34δ1/2+18δ1,\frac{1}{8}\delta_0 + \frac{3}{4}\delta_{1/2} + \frac{1}{8}\delta_1,

which have the same first three moments as μ\mu. Nevertheless, their support is disjoint from that of μ\mu, and supp⁡(π3)∩supp⁡(π⋆)=∅\operatorname{supp}(\pi^3) \cap\operatorname{supp}(\pi^\star) = \varnothing.

Example 1.1 shows that the optimal value alone does not detect how far the relaxed marginals are from the prescribed ones: the moment functions have to be chosen so as to retain geometric information. We now present our main results in a simplified form, together with their consequences; precise statements and assumptions are given in the subsequent sections. Throughout, for a cost cc we write C(π)=∫c dπC(\pi) = \int c\,\mathrm{d}\pi.

Geometric estimates

We say that the moment functions φ1,…,φn\varphi_1,\ldots,\varphi_n are localized at scale rnr_n on the support of μ\mu if they form a partition of unity on supp⁡(μ)\operatorname{supp}(\mu) made of functions supported in balls of radius rnr_n. Any probability measure having the same integrals as μ\mu against φ1,…,φn\varphi_1,\ldots,\varphi_n is called a relaxed marginal of μ\mu.

Theorem A (Geometric estimates). Assume that the moment functions are localized at scale rnr_n. Then every relaxed marginal μn\mu_n of μ\mu satisfies W∞(μ,μn)≤2rnW_\infty(\mu,\mu_n) \leq2r_n. Moreover, if μ\mu has a density bounded from below on a bounded Lipschitz domain Ω⊂Rd\Omega\subset\mathbb{R}^d and ν\nu is concentrated on Ω‾\overline{\Omega}, then, for all relaxed marginals μn\mu_n of μ\mu and νn\nu_n of ν\nu,

W∞(μn,νn)≤4rn+CWp(μn,νn)pp+d.W_\infty(\mu_n,\nu_n) \leq4r_n + C W_p(\mu_n,\nu_n)^{\frac{p}{p+d}}.

The first estimate rules out the behavior of Example 1.1: with localized moment functions, the relaxation cannot move any portion of mass farther than 2rn2r_n, so that the relaxed marginals converge to the prescribed ones in the strongest Wasserstein sense as rn→0r_n \to0, and not only narrowly. The second estimate compares an averaged distance with a uniform one. If μn\mu_n and νn\nu_n are close on average, namely Wp(μn,νn)W_p(\mu_n,\nu_n) is small, then they can be coupled without moving any point farther than CWp(μn,νn)p/(p+d)C W_p(\mu_n,\nu_n)^{p/(p+d)}, up to the resolution 4rn4r_n of the moment functions. Estimates of this type are classical when one of the measures has a density bounded from below [7, 20]. Here, the constants only depend on the prescribed marginal μ\mu, so that the estimate applies to the relaxed marginals, which are typically discrete. See Lemma 3.2, Theorem 3.4 and Corollary 3.5 for the precise statements.

Repulsive costs

An important motivation for these sparse discretizations comes from multi-marginal optimal transport problems with Coulomb cost, which arise in the strictly correlated electrons limit of density functional theory [9, 12] and for which moment-constrained approximations were introduced in [2, 1, 16]. For such costs, it is natural to ask whether the support of a minimizer of the moment relaxation remains at a positive distance from the diagonal, where the cost is singular.

We consider M≥2M \ge2 marginals all equal to ρ∈P(X)\rho\in\mathcal{P}(X) and costs of the form c(x1,…,xM)=∑i<jh(d(xi,xj))c(x_1,\ldots,x_M) = \sum_{i<j} h(d(x_i,x_j)), where hh is decreasing and blows up at 0, such as the Coulomb cost h(r)=1/rh(r) = 1/r on R3\mathbb{R}^3. The moment relaxation only prescribes the integrals of the marginals against finitely many moment functions, which need not be localized.

Theorem B (Separation from the diagonal). Assume that ρ\rho does not concentrate too much mass on small balls. Then there exists α⋆>0\alpha^\star> 0, independent of the number nn of moment functions, such that every minimizer of the moment-constrained problem gives no mass to configurations in which two points are closer than α⋆\alpha^\star.

This has two consequences. First, since minimizers never see the singularity, the cost can be replaced by a bounded continuous truncation without changing the minimizers (Theorem 4.6); this justifies the use of bounded costs in numerical computations. Second, when XX is compact and the moment functions are localized, the dual problem admits a symmetric maximizer in the span of the moment functions, with explicit bounds on its coefficients (Theorem 4.8), in the spirit of the bounds on Kantorovich potentials for the Coulomb cost [11, 8]. See Theorem 4.3 and Corollary 4.5 for the precise statements.

A semiclassical limit

Our third main result connects the moment-constrained Coulomb problem with its quantum counterpart. In density functional theory, the multi-marginal problem with Coulomb cost arises as the semiclassical limit of the Lieb functional [12, 5, 24, 13]: when the kinetic energy is multiplied by a small parameter ε\varepsilon, the minimal energy of MM fermions with a prescribed density converges to the optimal transport value. We consider the same quantum problem, where the density constraint is replaced by finitely many moment constraints on the MM-particle law, as in [16].

Theorem C (Semiclassical limit). Fix n∈Nn \in\mathbb{N} and let cc be the Coulomb cost. Under a technical assumption ensuring that the moment constraints can be corrected, the minimal energy

inf⁡Γ{ε2Tr⁡(−ΔΓ)+C(πΓ)},\inf_{\Gamma}\left\{\frac{\varepsilon}{2}\operatorname{Tr}(-\Delta\Gamma)+\mathcal{C}(\pi_\Gamma)\right\},

where the infimum is taken over the fermionic states Γ\Gamma whose MM-particle law πΓ\pi_\Gamma satisfies the moment constraints, converges, as ε→0\varepsilon\to0, to the minimum of C\mathcal{C} under the same moment constraints. Moreover, the laws of almost minimizing states converge, up to subsequences, to minimizers of the moment-constrained transport problem.

In other words, imposing finitely many moments and taking the semiclassical limit are compatible operations: the moment-constrained transport problem studied in the previous results is exactly the strictly correlated limit of the moment-constrained quantum problem. This justifies the use of moment-constrained optimal transport as an approximation of the quantum problem in the strongly correlated regime. The result follows from a Γ\Gamma-convergence statement (Theorem 5.7). The construction of recovery sequences relies on three steps: we approximate a transport plan by finitely supported plans with finite cost, which are therefore separated from the diagonal; we replace each atom by a smooth fermionic wave packet, which keeps the Coulomb energy under control; finally, we restore the moment constraints exactly through convex combinations with finitely many fixed states, which is where the technical assumption (Assumption 5.1) enters.

References

  1. [1]Aurélien Alfonsi, Rafaël Coyaud, and Virginie Ehrlacher. Constrained overdamped Langevin dynamics for symmetric multimarginal optimal transportation. Mathematical Models and Methods in Applied Sciences, 32(2):403–455, 2022.arxiv.org/abs/2102.03091
  2. [2]Aurélien Alfonsi, Rafaël Coyaud, Virginie Ehrlacher, and Damiano Lombardi. Approximation of optimal transport problems with marginal moments constraints. Mathematics of Computation, 90(328):689–737, 2021.arxiv.org/abs/1905.05663
  3. [3]Christian Bayer and Josef Teichmann. The proof of Tchakaloff’s theorem. Proceedings of the American Mathematical Society, 134(10):3035–3040, 2006.arxiv.org/abs/math/0502473
  4. [4]Mathias Beiglböck, Pierre Henry-Labordère, and Friedrich Penkner. Model-independent bounds for option prices—a mass transport approach. Finance and Stochastics, 17(3):477–501, 2013.
  5. [5]Ugo Bindini and Luigi De Pascale. Optimal transport with Coulomb cost and the semiclassical limit of density functional theory. Journal de l’École polytechnique — Mathématiques, 4:909–934, 2017.DOI
  6. [6]Jonathan M. Borwein and Adrian S. Lewis. Partially finite convex programming, part I: Quasi relative interiors and duality theory. Mathematical Programming, 57(1–3):15–48, 1992.DOI
  7. [7]Guy Bouchitté, Chloé Jimenez, and Mahadevan Rajesh. A new L∞ estimate in optimal mass transport. Proceedings of the American Mathematical Society, 135(11):3525–3536, 2007.DOI
  8. [8]Giuseppe Buttazzo, Thierry Champion, and Luigi De Pascale. Continuity and estimates for multi-marginal optimal transportation problems with singular costs. Applied Mathematics & Optimization, 78(1):185–200, 2018.
  9. [9]Giuseppe Buttazzo, Luigi De Pascale, and Paola Gori-Giorgi. Optimal-transport formulation of electronic density-functional theory. Physical Review A, 85(6):062502, 2012.DOI
  10. [10]Guillaume Carlier, Hugo Malamut, and Maxime Sylvestre. Weak optimal transport with moment constraints: constraint qualification, dual attainment and entropic regularization. arXiv preprint arXiv:2511.16211, 2025.arxiv.org/abs/2511.16211
  11. [11]Maria Colombo, Simone Di Marino, and Federico Stra. Continuity of multimarginal optimal transport with repulsive cost. SIAM Journal on Mathematical Analysis, 51(4):2903–2926, 2019.DOI
  12. [12]Codina Cotar, Gero Friesecke, and Claudia Klüppelberg. Density functional theory and optimal transportation with Coulomb cost. Communications on Pure and Applied Mathematics, 66(4):548–599, 2013.arxiv.org/abs/1104.0603
  13. [13]Codina Cotar, Gero Friesecke, and Claudia Klüppelberg. Smoothing of transport plans with fixed marginals and rigorous semiclassical limit of the Hohenberg–Kohn functional. Archive for Rational Mechanics and Analysis, 228(3):891–922, 2018.DOI
  14. [14]Gianni Dal Maso. An Introduction to Γ-Convergence, volume 8 of Progress in Nonlinear Differential Equations and Their Applications. Birkhäuser, Boston, 1993.DOI
  15. [15]Philipp J. di Dio and Konrad Schmüdgen. The multidimensional truncated moment problem: The moment cone. Journal of Mathematical Analysis and Applications, 511(1):126066, 2022.
  16. [16]Virginie Ehrlacher and Luca Nenna. Sparse approximation of the Lieb functional with moment constraints. SIAM Journal on Mathematical Analysis, 58(3):2865–2893, 2026.arxiv.org/abs/2306.00806
  17. [17]Gaoyue Guo and Jan Obłój. Computational methods for martingale optimal transport problems. The Annals of Applied Probability, 29(6):3311–3347, 2019.
  18. [18]Nouredinne Igbida and Van Thanh Nguyen. Optimal partial transport problem with Lagrangian costs. ESAIM: Mathematical Modelling and Numerical Analysis, 52(5):2109–2132, 2018.DOI
  19. [19]Matt Jacobs, Wonjun Lee, and Flavien Léger. The back-and-forth method for Wasserstein gradient flows. ESAIM: Control, Optimisation and Calculus of Variations, 27:28, 2021.arxiv.org/abs/2011.08151
  20. [20]Heikki Jylhä and Tapio Rajala. L∞ estimates in optimal mass transportation. Journal of Functional Analysis, 270(11):4297–4321, 2016.DOI
  21. [21]Alexander V. Kolesnikov, Aleksandra Novikova, and Kirill Sokolov. Duality for a martingale transport problem with moment constraints. arXiv preprint arXiv:2609.03187, 2026.arxiv.org/abs/2609.03187
  22. [22]Jonathan Korman and Robert J. McCann. Optimal transportation with capacity constraints. Transactions of the American Mathematical Society, 367(3):1501–1521, 2015.arxiv.org/abs/1201.6404
  23. [23]Jonathan Korman, Robert J. McCann, and Christian Seis. Dual potentials for capacity constrained optimal transport. Calculus of Variations and Partial Differential Equations, 54(1):573–584, 2015.arxiv.org/abs/1307.7774
  24. [24]Mathieu Lewin. Semi-classical limit of the Levy–Lieb functional in density functional theory. Comptes Rendus Mathématique, 356(4):449–455, 2018.arxiv.org/abs/1706.02199
  25. [25]John Lott and Cédric Villani. Ricci curvature for metric-measure spaces via optimal transport. Annals of Mathematics, 169(3):903–991, 2009.
  26. [26]Iosif Pinelis. On the extreme points of moments sets. Mathematical Methods of Operations Research, 83(3):325–349, 2016.arxiv.org/abs/1204.0249
  27. [27]Alessandro Pinzi and Giuseppe Savaré. Totally convex functions, L²-optimal transport for laws of random measures, and solution to the Monge problem. arXiv preprint arXiv:2509.01768, 2025.arxiv.org/abs/2509.01768
  28. [28]Walter Rudin. Functional Analysis. International Series in Pure and Applied Mathematics. McGraw-Hill, Inc., New York, NY, second edition, 1991.
  29. [29]Konrad Schmüdgen. The moment problem, volume 277 of Graduate Texts in Mathematics. Springer, Cham, 2017.DOI
  30. [30]Cédric Villani. Topics in Optimal Transportation, volume 58 of Graduate Studies in Mathematics. American Mathematical Society, Providence, RI, 2003.DOI
  31. [31]Gerhard Winkler. Extreme points of moment sets. Mathematics of Operations Research, 13(4):581–587, 1988.DOI
  32. [32]Danila A. Zaev. On the Monge–Kantorovich problem with additional linear constraints. Mathematical Notes, 98(5–6):725–741, 2015.arxiv.org/abs/1404.4962

Paper details

vlnc:W_ea4esqd4
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 2

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