Lipschitz continuity of the cut time for globally hyperbolic spacetimes

Alec Metsch

REVIEW 4 major objections 3 minor 58 references

Lipschitz continuity of the cut time for globally hyperbolic spacetimes

T0 review · 4 major / 3 minor · reviewed 2026-08-02 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The cut time is locally Lipschitz in globally hyperbolic spacetimes

desk verdict Credible extension of Itoh–Tanaka to Lorentzian geometry, but the main estimate has an explicitly non-rigorous gap that needs closing before I would rely on it. read the letter →

arxiv 2607.14374 v1 pith:2LXZUEVU submitted 2026-07-15 math.DG math-phmath.MP

Alec Metsch This is my paper · ORCID

classification math.DGmath-phmath.MP MSC 53C5058C0749Q99

keywords cutlocustimefocalizationgloballyhyperbolicspacetimeLipschitzcontinuityLorentziangeometryHausdorffdimensionconjugatepoints

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper proves that two central functions in Lorentzian geometry, the cut time and the focalization time, are locally Lipschitz continuous, improving the previously known continuity results. For a fixed starting point in a globally hyperbolic spacetime, the focalization time is Lipschitz on its entire finite domain, including null directions. The cut time is Lipschitz in a neighborhood of every timelike direction whose geodesic is defined up to and including the cut time, with an explicit bound that blows up like the inverse square of the velocity near the null cone. Under an additional condition on the cut point, the Lipschitz property extends to null directions as well. A direct consequence is that the cut locus of a point has Hausdorff dimension at most n-1, matching the known Riemannian bound.

What carries the argument

The central objects are the Lorentzian index form, evaluated on piecewise smooth vector fields orthogonal to a causal geodesic, and the corresponding Jacobi fields that encode how nearby geodesics deviate. Lemma 3.1 shows the index form is Lipschitz in the initial velocity, and Theorem 3.3 constructs vector fields with positive index near a conjugate time. For the cut time, the key geometric input is the cut-point dichotomy—that a cut point is either a first conjugate point or the endpoint of a second distinct maximizing geodesic. The proof also relies on the unit-hyperboloid estimate that connects the hyperbolic angle between two future unit timelike directions to the Euclidean distance bet

What would settle it

Construct a globally hyperbolic spacetime with a future null geodesic from a point x that has a conjugate point before any cut point yet still maximizes Lorentzian length beyond the conjugate point. If such an example exists, Theorem A.10 is false, which would remove the cut-point dichotomy for null geodesics and invalidate the proof of the null-cone Lipschitz extension, while leaving the timelike theorem intact.

Watch

Extended reading notes

Core claim

The paper establishes the estimate |ρ(v)-ρ(w)| ≤ C|v-w|_x / min{|v|_g^2, |w|_g^2} for the cut time ρ, valid for future-directed timelike vectors v,w in a neighborhood of a given vector whose geodesic exists at the cut time. In particular, ρ is locally Lipschitz near such a timelike vector. The proof adapts a classical Riemannian strategy: it first shows the focalization time λ is locally Lipschitz everywhere by transferring control of the Lorentzian index form to nearby initial velocities; then it controls the derivative of ρ along line segments in the tangent space using the second maximizing geodesic delivered by the cut-point dichotomy. The denominator min{|v|^2,|w|^2} is inherent to the

Load-bearing premise

The proof relies on the theorem that a null geodesic containing a conjugate point is not maximizing, which is imported from lecture notes without proof; if that theorem is false, the cut-point dichotomy and hence the null-direction extension could fail.

Editorial extensions

If this is right

For any fixed point in a globally hyperbolic spacetime, the focalization time is locally Lipschitz on the open set where it is finite, including along null directions.
The cut time is locally Lipschitz in a neighborhood of every timelike direction whose geodesic is defined up to and including the cut time, with a uniform local Lipschitz constant.
If the cut point of a null direction occurs before its first conjugate point, then the cut time is locally Lipschitz in a full neighborhood of that null direction.
The cut locus of a point has Hausdorff dimension at most n-1, so it cannot fill an open set.
The cut time is locally semiconcave near timelike directions when the cut precedes the conjugate point, a property with independent consequences for the regularity of the time separation function.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The blow-up factor 1/min{|v|^2, |w|^2} is unlikely to be a technical artefact; it suggests that the cut time may be Hölder continuous with a specific exponent in directions approaching the null cone, a claim one could test explicitly in model spacetimes such as the Lorentzian cylinder.
If the imported theorem that a null geodesic containing a conjugate point is not maximizing turned out to fail, the argument for the null-cone part would collapse, while the timelike part would stand; this is a concrete open point worth checking.
The Hausdorff dimension bound opens the way to studying the rectifiability or smooth stratification of the cut locus in globally hyperbolic spacetimes, and to extending the argument to spacelike submanifolds, which the paper states as a natural next step.
A direct numerical check of the Lipschitz constant near the null cone in an explicit spacetime would sharpen the quantitative estimate and could reveal whether the constant C can be chosen independent of the base point in certain classes of spacetimes.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Short version: this is a genuine extension of Itoh–Tanaka, not a repackaging, and the timelike core is probably correct; but the written proof leans on a lemma that the author says is not completely rigorous, and the null-direction results depend on an unproved imported theorem. I'd send it to a referee, not desk-reject it, but I wouldn't cite the current version for the main theorem.

What is actually new: first Lipschitz regularity for the cut time and focalization time in globally hyperbolic spacetimes, plus a clean Hausdorff-codimension corollary. The Lorentzian setting is not a translation: the non-openness of the cut-time domain, possible geodesic incompleteness, and the blow-up near the null cone all require new work. The author is upfront about which parts are adapted and which are new, and the exposition is generally careful.

The main soft spot is Lemma 4.2(iii). The statement is needed to control the quadratic error in the Jacobi-field estimate, and the proof says 'we shall keep the argument intuitive, while being not completely rigorous.' The Taylor expansion of phi composed with the exponential map gives a scalar remainder controlled by phi, but (4.4)–(4.5) require vector bounds uniform over a bounded set of test functions and over the compact set of pairs (v,w). That uniformity is not supplied. It is probably fixable—one can choose phi as a bounded linear functional with controlled Hessian—but as written it is a real gap, not a typo. It affects every timelike direction in Theorem 1.4, because the quadratic estimate is what absorbs the sinh^2(phi/2) factor.

The second gap is Theorem A.10, imported from lecture notes, which says a null geodesic containing a conjugate point is not maximizing. This is used for continuity of rho in null directions and for Corollary 1.6. The timelike core does not need it, but the advertised null-direction statements do. The self-citation to [14] is not a problem, since [13] is cited alongside it.

For a reader: this is aimed at Lorentzian geometers and anyone working on regularity of distance functions. I would bring it to a reading group, and I would want a careful referee on Lemma 4.2(iii). If the author closes that gap, I would cite it.

Referee Report

4 major / 3 minor

Summary. This paper studies the cut time ρ and focalization time λ for a fixed point x in a globally hyperbolic spacetime. It claims: λ is locally Lipschitz on its relatively open domain (Thm 1.3); ρ is locally Lipschitz near timelike vectors with the quantitative estimate |ρ(v)−ρ(w)| ≤ C |v−w|_x / min{|v|^2_g, |w|^2_g} (Thm 1.4), with a refined result near null directions when ρ<λ (Cor 1.6); and dim_H Cut(x) ≤ n−1 (Cor 1.5). The method adapts Itoh–Tanaka's Riemannian argument, using index-form estimates, Jacobi fields, and Lorentzian subgradients. Appendices collect background on conjugate/cut points.

Significance. If complete, this is the first Lipschitz regularity result for the Lorentzian cut time and a natural extension of Itoh–Tanaka and Li–Nirenberg. The quantitative blow-up of the Lipschitz constant near the null cone is new and interesting. The paper is clearly written and mostly self-contained, with explicit constants and a systematic handling of compactness (Lemmas A.14, A.17). However, the central estimate Lemma 4.2(iii) is explicitly non-rigorous, and the null part depends on an unproved theorem from lecture notes. Thus the results are plausible but not yet established.

major comments (4)

[§4.1, Lemma 4.2(iii), Eqs. (4.4)–(4.5)] This lemma is load-bearing: Lemma 4.1 uses it at (4.15)–(4.16) to replace e1(w) by e1 − ˙X/|v| and to control ⟨Y, ˙X(ρ)⟩ after swapping covariant derivatives. The proof is explicitly non-rigorous ('we shall keep the argument intuitive, while being not completely rigorous'). The Taylor expansion (4.7)–(4.8) with an arbitrary test function φ does not establish the uniform vector-valued estimates (4.4)–(4.5): one needs a uniform remainder for a bounded family of C² functions and uniform control over the compact set from Lemma A.17, including the dependence on |v|_g as v approaches the null cone. Without this, the factor sinh²(φ/2) in (4.15)–(4.16) cannot be absorbed and Theorem 1.4 is not proved. A direct Taylor expansion of F(t,v)=exp_x(tv) should be written out.
[Appendix A, Theorem A.10 / proof of Theorem A.15] Theorem A.10 (a null geodesic containing a conjugate point is not maximizing) is imported from lecture notes [2] without proof, and the text explicitly says the argument is omitted. This theorem is used in Theorem A.15, which in turn is used in Lemma 4.7 to produce the second distinct maximizing geodesic w(s0). It is also needed for Corollary 1.6 and the continuity of ρ at null vectors. Since the paper otherwise provides a self-contained survey, please either prove this theorem or replace [2] by a refereed published reference with a proof; currently the null-direction part of the main conclusions is unsupported.
[§3, proof of Theorem 1.3] The proof establishes (3.1) only for v,w∈U∩int(C_x), i.e. for pairs of timelike vectors. The theorem asserts local Lipschitz continuity on the whole relatively open domain, which includes null directions when v* is null. The phrase 'combining the continuity of λ with a connectedness argument' is not carried out. To justify the null case one needs an explicit limiting argument (e.g., approximate each causal vector by timelike vectors and pass to the limit using Lemma A.12), or a variant of Lemma 3.4 that does not require the denominator |v|_g to be nonzero. This is likely repairable, but it is not in the text.
[§4, proof of Theorem 1.4 after (4.3)] In the 'otherwise' case, C̃ is defined as C/ min{|v0|²_g,|v1|²_g} and then used as a Lipschitz constant for λ on U. As written, this constant depends on the pair (v0,v1) that was only fixed at the start of the argument; it is not a uniform constant on U. The proof should either fix a smaller neighbourhood with a uniform lower bound on |v|_g (when v* is timelike) or explicitly state that for each pair one applies the bound with the relevant minimum and then combines the inequalities. Please also check the consistency of the squared denominators in (4.1) and (4.2).

minor comments (3)

[Throughout] Typos: 'Semiconvavity' in Proposition 1.2; 'Uniform estmates' in the heading of Lemma 4.2; 'costs no generality' in Appendix B should be 'costs no loss of generality'; in the proof of Theorem A.15 the curve is sometimes called γ and sometimes c.
[Lemma A.4(i)] In the displayed Jacobi equation, the curvature term is written with E_j(t,v,w), although the parallel frame is E_j(t,v); please correct.
[Lemma 4.2(ii)] The notation exp(y, −t ˙c_v(ρ(v))) is not defined; the map should be exp_y(tu). The derivative computation leading to the contradiction is very terse and should be written out, especially the role of the fact ρ(v_k)=ρ(w_k).

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the cut-time Lipschitz estimate is derived from independent local estimates and external Riemannian benchmarks, not from the target regularity.

full rationale

The derivation chain is self-contained and non-circular. Theorem 1.4 is reduced to Lemma 4.1, which is proved from the exact identity (4.15) and the uniform estimates of Lemma 4.2 (compactness + Taylor bounds on Jacobi fields) and Lemma 4.4 (proved in Appendix C). No fitted parameter is used as a prediction, and no quantity is defined in terms of the theorem it purports to prove. The only author-overlapping citation, [14] for local semiconvexity of the time separation function, is paired with the independent external reference [13] (McCann), so it is not load-bearing; moreover, Proposition 1.2 is only used as a convenience, with the paper explicitly noting in footnote 3 that only local Lipschitz continuity of ρ is needed, which it attributes to the first variation formula. The null-conjugate non-maximality theorem (A.10) is imported from external lecture notes [2], not from the authors' own work, so no uniqueness theorem is imported from the same authors. The paper itself flags two gaps: Lemma 4.2(iii) is kept 'intuitive, while being not completely rigorous', and Theorem A.10 is only proved in the timelike case, with the null case deferred to [2]. These are rigor gaps, not circularity: the Taylor estimate in Lemma 4.2(iii) is an independent bound derived from the smooth exponential map, not an assumption of the target Lipschitz regularity. The central claim does not reduce to its inputs by construction, so the appropriate circularity score is 0.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

No empirical or fitted parameters appear: constants C and C_i are existential bounds produced by compactness and continuity, not numbers fitted to data. The central claim rests on standard Lorentzian geometry—global hyperbolicity, the time-separation semiconvexity theorem, and the classical cut-point/conjugate-point dichotomy, including a null maximality result the paper imports rather than proves.

assumptions (3)

domain assumption The time separation function d(x,·) is locally semiconvex on I^+(x), with subgradient −˙c(b)/|˙c(b)|_g at the endpoint of a maximizing geodesic (§2, (2.1)–(2.2)).
Used to define sub-gradients and to prove semiconcavity of ρ (Proposition 1.2) and Lemma 4.7. Cited to [13] (McCann) and [14] (same author); the latter is a self-citation but [13] gives independent support.
domain assumption Global hyperbolicity gives maximizing geodesics between causally related points and compactness of sets of maximizing geodesics (Lemma A.14).
Used throughout, e.g. in Lemmas A.16–A.17 and Lemma 4.2; standard from [1] and [16].
domain assumption A causal geodesic that contains a conjugate point is not maximizing (Theorem A.10), including the null case.
Used in Theorem A.15 to characterize cut points and hence in Lemma 4.7 and Corollary 1.6. The paper proves only the timelike case and cites lecture notes [2] for the null case.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Lipschitz continuity of the cut time for globally hyperbolic spacetimes." pith.science (2026). https://pith.science/paper/2LXZUEVU

@misc{pith2026260714374,
  author       = {Pith},
  title        = {Pith review of: Lipschitz continuity of the cut time for globally hyperbolic spacetimes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2LXZUEVU}},
  note         = {Machine review of arXiv:2607.14374}
}

abstract

We prove that, for a fixed point in a globally hyperbolic spacetime, the focalization time is locally Lipschitz continuous on the open subset of the future causal cone where it is finite. We also show that the cut time is locally Lipschitz continuous in a neighborhood of any timelike tangent vector whose associated geodesic is defined at least up to (and including) its cut time. Furthermore, we derive quantitative estimates for the Lipschitz constant near the null cone and provide a criterion ensuring that the Lipschitz property extends to null directions. As a consequence, we show that the cut locus of a point has Hausdorff codimension at least $1$. These results extend classical Lipschitz continuity results for complete Riemannian manifolds due to Itoh-Tanaka and Li-Nirenberg. Our approach follows the method of Itoh-Tanaka, suitably adapted to the Lorentzian setting.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

58 extracted references · 15 canonical work pages

[14]

Hamiltonian systems: dynamics, analysis, applications , SERIES =

Fathi, Albert , TITLE =. Hamiltonian systems: dynamics, analysis, applications , SERIES =. 2024 , ISBN =

2024
[13]

Israel J

Fathi, Albert and Figalli, Alessio , TITLE =. Israel J. Math. , FJOURNAL =. 2010 , PAGES =. doi:10.1007/s11856-010-0001-5 , URL =

doi:10.1007/s11856-010-0001-5 2010
[2]

Christian Bär , title =
[1]

2005 , PAGES =

Ambrosio, Luigi and Gigli, Nicola and Savar\'e, Giuseppe , TITLE =. 2005 , PAGES =

2005
[3]

and S\'anchez, Miguel , TITLE =

Bernal, Antonio N. and S\'anchez, Miguel , TITLE =. Classical Quantum Gravity , FJOURNAL =. 2007 , NUMBER =. doi:10.1088/0264-9381/24/3/N01 , URL =

doi:10.1088/0264-9381/24/3/n01 2007
[4]

Optimal Transport: Old and New

Villani, Cedric. Optimal Transport: Old and New
[5]

Bernard, Patrick , TITLE =. Ann. Sci. \'Ecole Norm. Sup. (4) , FJOURNAL =. 2007 , NUMBER =. doi:10.1016/j.ansens.2007.01.004 , URL =

doi:10.1016/j.ansens.2007.01.004 2007
[6]

Fathi, Albert and Figalli, Alessio and Rifford, Ludovic , TITLE =. Comm. Pure Appl. Math. , FJOURNAL =. 2009 , NUMBER =. doi:10.1002/cpa.20250 , URL =

work page doi:10.1002/cpa.20250 2009

Show all 58 references

[7]

Cannarsa, Piermarco and Cheng, Wei and Fathi, Albert , TITLE =. Publ. Math. Inst. Hautes \'Etudes Sci. , FJOURNAL =. 2021 , PAGES =. doi:10.1007/s10240-021-00125-5 , URL =

2021 doi
[8]

Kell, Martin and Suhr, Stefan , TITLE =. Ann. Inst. H. Poincar\'e. 2020 , NUMBER =. doi:10.1016/j.anihpc.2019.09.005 , URL =

2020 doi
[9]

Riemannian Geometry

Do Carmo, Manfredo P. Riemannian Geometry
[10]

M\"unster J

Suhr, Stefan , TITLE =. M\"unster J. Math. , FJOURNAL =. 2018 , NUMBER =. doi:10.17879/87109580432 , URL =

2018 doi
[11]

Bernard, Patrick and Suhr, Stefan , TITLE =. Comm. Math. Phys. , FJOURNAL =. 2018 , NUMBER =. doi:10.1007/s00220-018-3127-7 , URL =

2018 doi
[12]

The weak Kam theorem in Lagrangian dynamics

Fathi, Albert. The weak Kam theorem in Lagrangian dynamics
[15]

ESAIM Control Optim

Figalli, Alessio and Gigli, Nicola , TITLE =. ESAIM Control Optim. Calc. Var. , FJOURNAL =. 2011 , NUMBER =. doi:10.1051/cocv/2010011 , URL =

2011
[16]

and Ehrlich, Paul E

Beem, John K. and Ehrlich, Paul E. and Easley, Kevin L. , TITLE =. 1996 , PAGES =

1996
[17]

Semi-Riemannian geometry with applications to relativity: Volume 103

O'Neill, Barrett. Semi-Riemannian geometry with applications to relativity: Volume 103
[18]

Modelling and optimisation of flows on networks , SERIES =

Ambrosio, Luigi and Gigli, Nicola , TITLE =. Modelling and optimisation of flows on networks , SERIES =. 2013 , ISBN =. doi:10.1007/978-3-642-32160-3\_1 , URL =

2013 doi
[19]

Optimal transportation and applications (

Ambrosio, Luigi and Pratelli, Aldo , TITLE =. Optimal transportation and applications (. 2003 , ISBN =. doi:10.1007/978-3-540-44857-0\_5 , URL =

2003 doi
[20]

Robert J. McCann. Polar factorization of maps on Riemannian manifolds. GAFA Geom. Funct. Anal
[21]

Globally hyperbolic spacetimes can be defined as `causal' instead of `strongly causal'

Bernal, Antonio N and S \'a nchez, Miguel. Globally hyperbolic spacetimes can be defined as `causal' instead of `strongly causal'. Class. Quantum Gravity
[22]

, TITLE =

McCann, Robert J. , TITLE =. Camb. J. Math. , FJOURNAL =. 2020 , NUMBER =. doi:10.4310/CJM.2020.v8.n3.a4 , URL =

2020 doi
[23]

Living Reviews in Relativity , year=

Lorentzian causality theory , author=. Living Reviews in Relativity , year=
[24]

Monge, G. , url=. M. 1781 , publisher=
[25]

, TITLE =

Kantorovitch, L. , TITLE =. C. R. (Doklady) Acad. Sci. URSS (N.S.) , FJOURNAL =. 1942 , PAGES =

1942
[26]

Kantorovich, L. V. , TITLE =. Zap. Nauchn. Sem. S.-Peterburg. Otdel. Mat. Inst. Steklov. (POMI) , FJOURNAL =. 2004 , PAGES =. doi:10.1007/s10958-006-0050-9 , URL =

2004 doi
[27]

1982 , PAGES =

Klingenberg, Wilhelm , TITLE =. 1982 , PAGES =

1982
[28]

Communications on Pure and Applied Mathematics , volume =

Brenier, Yann , title =. Communications on Pure and Applied Mathematics , volume =. doi:https://doi.org/10.1002/cpa.3160440402 , url =. https://onlinelibrary.wiley.com/doi/pdf/10.1002/cpa.3160440402 , abstract =

doi
[29]

Optimal transportation and applications (

Brenier, Yann , TITLE =. Optimal transportation and applications (. 2003 , ISBN =

2003
[30]

Calculus of Variations and Partial Differential Equations , year=

The optimal mass transport problem for relativistic costs , author=. Calculus of Variations and Partial Differential Equations , year=
[31]

Kantorovich potentials and continuity of total cost for relativistic cost functions , journal =

Jerome Bertrand and Aldo Pratelli and Marjolaine Puel , keywords =. Kantorovich potentials and continuity of total cost for relativistic cost functions , journal =. 2018 , issn =. doi:https://doi.org/10.1016/j.matpur.2017.09.005 , url =

2018 doi
[32]

[2021] 2021 , PAGES =

Figalli, Alessio and Glaudo, Federico , TITLE =. [2021] 2021 , PAGES =. doi:10.4171/ETB/22 , URL =

2021 doi
[33]

2016 , eprint=

On the continuity of the total cost in the mass transport problem with relativistic cost functions , author=. 2016 , eprint=

2016
[34]

arXiv preprint arXiv:2412.01012 , year=

Optimal transport and regularity of weak Kantorovich potentials on a globally hyperbolic spacetime , author=. arXiv preprint arXiv:2412.01012 , year=. 2412.01012 , archivePrefix=

arXiv
[35]

Alec Metsch , journal=. C_. 2025 , eprint=

2025
[36]

arXiv preprint arXiv:2507.22737 , year=

On the locus of multiple maximizing geodesics on a globally hyperbolic spacetime , author=. arXiv preprint arXiv:2507.22737 , year=. 2507.22737 , archivePrefix=

arXiv
[37]

Mondino, Andrea and Suhr, Stefan , TITLE =. J. Eur. Math. Soc. (JEMS) , FJOURNAL =. 2023 , NUMBER =. doi:10.4171/jems/1188 , URL =

2023 doi
[38]

Cavalletti, Fabio and Mondino, Andrea , TITLE =. Camb. J. Math. , FJOURNAL =. 2024 , NUMBER =. doi:10.4310/cjm.2024.v12.n2.a3 , URL =

2024 doi
[39]

Acta Mathematica , number =

Karl-Theodor Sturm , title =. Acta Mathematica , number =. 2006 , doi =

2006
[40]

Lott, John and Villani, C\'edric , TITLE =. Ann. of Math. (2) , FJOURNAL =. 2009 , NUMBER =. doi:10.4007/annals.2009.169.903 , URL =

2009 doi
[41]

Causality for nonlocal phenomena , JOURNAL =

Eckstein, Micha. Causality for nonlocal phenomena , JOURNAL =. 2017 , NUMBER =. doi:10.1007/s00023-017-0566-1 , URL =

2017 doi
[42]

2003 , PAGES =

Villani, C\'edric , TITLE =. 2003 , PAGES =. doi:10.1090/gsm/058 , URL =

2003 doi
[43]

Erbar, Matthias and Kuwada, Kazumasa and Sturm, Karl-Theodor , TITLE =. Invent. Math. , FJOURNAL =. 2015 , NUMBER =. doi:10.1007/s00222-014-0563-7 , URL =

2015 doi
[44]

Hawking, S. W. and Penrose, R. , TITLE =. Proc. Roy. Soc. London Ser. A , FJOURNAL =. 1970 , PAGES =. doi:10.1098/rspa.1970.0021 , URL =

1970
[45]

Kunzinger, Michael and S\"amann, Clemens , TITLE =. Ann. Global Anal. Geom. , FJOURNAL =. 2018 , NUMBER =. doi:10.1007/s10455-018-9633-1 , URL =

2018 doi
[46]

A reconstruction of the initial conditions of the Universe by optimal mass transportation , volume =

Frisch, Uriel and Matarrese, Sabino and Mohayaee, Roya and Sobolevski, Andrei , year =. A reconstruction of the initial conditions of the Universe by optimal mass transportation , volume =. Nature , doi =
[47]

and Frisch, U

Brenier, Y. and Frisch, U. and Hénon, M. and Loeper, G. and Matarrese, S. and Mohayaee, R. and Sobolevskiĭ, A. , title =. Monthly Notices of the Royal Astronomical Society , volume =. 2003 , month =. doi:10.1046/j.1365-2966.2003.07106.x , url =

2003
[48]

and Rifford, L

Figalli, A. and Rifford, L. and Villani, C. , TITLE =. Differential Geom. Appl. , FJOURNAL =. 2011 , NUMBER =. doi:10.1016/j.difgeo.2011.02.002 , URL =

2011 doi
[49]

Itoh, Jin-ichi and Tanaka, Minoru , TITLE =. Trans. Amer. Math. Soc. , FJOURNAL =. 2001 , NUMBER =. doi:10.1090/S0002-9947-00-02564-2 , URL =

2001 doi
[50]

Li, Yanyan and Nirenberg, Louis , TITLE =. Comm. Pure Appl. Math. , FJOURNAL =. 2005 , NUMBER =. doi:10.1002/cpa.20051 , URL =

2005 doi
[51]

ESAIM Control Optim

Castelpietra, Marco and Rifford, Ludovic , TITLE =. ESAIM Control Optim. Calc. Var. , FJOURNAL =. 2010 , NUMBER =. doi:10.1051/cocv/2009020 , URL =

2010
[52]

Whitehead, J. H. C. , TITLE =. Ann. of Math. (2) , FJOURNAL =. 1935 , NUMBER =. doi:10.2307/1968651 , URL =

1935 doi
[53]

Gluck, Herman and Singer, David , TITLE =. Ann. of Math. (2) , FJOURNAL =. 1978 , NUMBER =. doi:10.2307/1971170 , URL =

1978 doi
[54]

, TITLE =

Hebda, James J. , TITLE =. J. Differential Geom. , FJOURNAL =. 1994 , NUMBER =

1994
[55]

Itoh, Jin-ichi , TITLE =. J. Differential Geom. , FJOURNAL =. 1996 , NUMBER =

1996
[56]

and Schmuckenschl\"ager, Michael , TITLE =

Cordero-Erausquin, Dario and McCann, Robert J. and Schmuckenschl\"ager, Michael , TITLE =. Invent. Math. , FJOURNAL =. 2001 , NUMBER =. doi:10.1007/s002220100160 , URL =

2001 doi
[57]

Sakai, Takashi , TITLE =. Math. Ann. , FJOURNAL =. 1978 , NUMBER =. doi:10.1007/BF01405010 , URL =

1978 doi
[58]

Tsukuba J

Takeuchi, Masaru , TITLE =. Tsukuba J. Math. , FJOURNAL =. 1978 , PAGES =. doi:10.21099/tkbjm/1496158504 , URL =

1978

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.

[1] [14]

Hamiltonian systems: dynamics, analysis, applications , SERIES =

Fathi, Albert , TITLE =. Hamiltonian systems: dynamics, analysis, applications , SERIES =. 2024 , ISBN =

2024

[2] [13]

Israel J

Fathi, Albert and Figalli, Alessio , TITLE =. Israel J. Math. , FJOURNAL =. 2010 , PAGES =. doi:10.1007/s11856-010-0001-5 , URL =

doi:10.1007/s11856-010-0001-5 2010

[3] [2]

Christian Bär , title =

[4] [1]

2005 , PAGES =

Ambrosio, Luigi and Gigli, Nicola and Savar\'e, Giuseppe , TITLE =. 2005 , PAGES =

2005

[5] [3]

and S\'anchez, Miguel , TITLE =

Bernal, Antonio N. and S\'anchez, Miguel , TITLE =. Classical Quantum Gravity , FJOURNAL =. 2007 , NUMBER =. doi:10.1088/0264-9381/24/3/N01 , URL =

doi:10.1088/0264-9381/24/3/n01 2007

[6] [4]

Optimal Transport: Old and New

Villani, Cedric. Optimal Transport: Old and New

[7] [5]

Bernard, Patrick , TITLE =. Ann. Sci. \'Ecole Norm. Sup. (4) , FJOURNAL =. 2007 , NUMBER =. doi:10.1016/j.ansens.2007.01.004 , URL =

doi:10.1016/j.ansens.2007.01.004 2007

[8] [6]

Fathi, Albert and Figalli, Alessio and Rifford, Ludovic , TITLE =. Comm. Pure Appl. Math. , FJOURNAL =. 2009 , NUMBER =. doi:10.1002/cpa.20250 , URL =

work page doi:10.1002/cpa.20250 2009