pith. sign in

arxiv: 2305.05211 · v3 · submitted 2023-05-09 · 🧮 math.FA · math.DS· math.OC· math.PR

A Lagrangian approach to totally dissipative evolutions in Wasserstein spaces

Giulia Cavagnari , Giuseppe Savar\'e , Giacomo Enrico Sodini This is my paper

Pith reviewed 2026-05-24 08:55 UTC · model grok-4.3

classification 🧮 math.FA math.DSmath.OCmath.PR
keywords Wasserstein spacetotally dissipative operatorsmultivalued probability vector fieldsmaximal dissipative operatorsmean field limitsgradient flowsdisplacement convex functionalsLagrangian approach
0
0 comments X

The pith

Totally dissipative multivalued probability vector fields on Wasserstein space are in one-to-one correspondence with law-invariant dissipative operators in L2 spaces of random variables.

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

The paper shows that totally dissipative MPVFs on the Wasserstein space correspond bijectively to law-invariant dissipative operators in a Hilbert space of random variables, preserving maximality. This correspondence allows importing existence, uniqueness, stability, and approximation results from the theory of maximal dissipative operators in Hilbert spaces to the Wasserstein framework. It also establishes that demicontinuous single-valued fields with metric dissipativity are totally dissipative and provides a maximal extension from discrete measures that characterizes mean-field limits of particle systems. Additionally, it uncovers structural properties for gradient flows of displacement convex functionals.

Core claim

We show that the class of totally dissipative multivalued probability vector fields on the Wasserstein space is in one to one correspondence with law-invariant dissipative operators in L2 of random variables, preserving maximality. This allows importing tools from Hilbert space theory to derive existence, uniqueness, stability, and approximation results for the flow generated by a maximal totally dissipative MPVF and the equivalence of its Eulerian and Lagrangian characterizations. Demicontinuous single-valued probability vector fields satisfying a metric dissipativity condition are totally dissipative, and from a rich set of discrete measures a unique maximal totally dissipative version can

What carries the argument

The one-to-one correspondence between totally dissipative multivalued probability vector fields (MPVFs) on Wasserstein space and law-invariant dissipative operators in L2(Ω, B, P; X) of random variables.

If this is right

  • Existence, uniqueness, stability, and approximation results for the flow generated by a maximal totally dissipative MPVF.
  • Equivalence of Eulerian and Lagrangian characterizations of the flow.
  • Demicontinuous single-valued probability vector fields satisfying metric dissipativity are totally dissipative.
  • Recovery of a unique maximal totally dissipative MPVF from discrete measures, characterizing asymptotic limits of particle systems.
  • New structural properties for gradient flows of displacement convex functionals with a core of discrete measures dense in energy.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The correspondence may enable the use of Hilbert space numerical methods to approximate Wasserstein evolutions.
  • It could connect to the study of mean-field limits in other contexts like interacting particle systems beyond the discrete case.
  • Potential extensions to non-Euclidean or infinite-dimensional base spaces if the law-invariance can be maintained.

Load-bearing premise

The multivalued probability vector field must satisfy the total dissipativity condition along with maximality for the bijection with law-invariant dissipative operators to hold.

What would settle it

A concrete example of a law-invariant dissipative operator in L2 that does not correspond to a totally dissipative MPVF on the Wasserstein space, or vice versa, would disprove the claimed one-to-one correspondence.

read the original abstract

We introduce and study the class of totally dissipative multivalued probability vector fields (MPVF) $\boldsymbol{\mathrm F}$ on the Wasserstein space $(\mathcal{P}_2(\mathsf{X}),W_2)$ of Euclidean or Hilbertian probability measures. We show that such class of MPVFs is in one to one correspondence with law-invariant dissipative operators in a Hilbert space $L^2(\Omega,\mathcal{B},\mathbb{P};\mathsf{X})$ of random variables, preserving a natural maximality property. This allows us to import in the Wasserstein framework many of the powerful tools from the theory of maximal dissipative operators in Hilbert spaces, deriving existence, uniqueness, stability, and approximation results for the flow generated by a maximal totally dissipative MPVF and the equivalence of its Eulerian and Lagrangian characterizations. We will show that demicontinuous single-valued probability vector fields satisfying a metric dissipativity condition are in fact totally dissipative. Starting from a sufficiently rich set of discrete measures, we will also show how to recover a unique maximal totally dissipative version of a MPVF, proving that its flow provides a general mean field characterization of the asymptotic limits of the corresponding family of discrete particle systems.Such an approach also reveals new interesting structural properties for gradient flows of displacement convex functionals with a core of discrete measures dense in energy.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 4 minor

Summary. The paper introduces totally dissipative multivalued probability vector fields (MPVFs) on the Wasserstein space (P₂(X), W₂) and proves a bijective correspondence between this class and law-invariant maximal dissipative operators on the Hilbert space L²(Ω; X) of random variables, with the correspondence preserving maximality. This bijection is used to transfer existence, uniqueness, stability, and approximation results from Hilbert-space theory, to establish equivalence of Eulerian and Lagrangian characterizations of the generated flow, to show that demicontinuous single-valued fields satisfying metric dissipativity are totally dissipative, and to construct maximal totally dissipative MPVFs from discrete measures whose flows characterize mean-field limits of particle systems. New structural properties are derived for gradient flows of displacement-convex functionals whose discrete cores are dense in energy.

Significance. If the central bijection is established without circularity, the work supplies a systematic Lagrangian bridge that imports the full toolkit of maximal dissipative operator theory into Wasserstein geometry. The explicit recovery of maximal extensions from discrete measures and the resulting mean-field characterization constitute concrete, falsifiable contributions with direct applicability to interacting particle systems and gradient flows.

minor comments (4)
  1. [Definition of total dissipativity] The definition of total dissipativity (presumably in §2 or §3) should be stated with an explicit formula or inequality that makes the preservation of maximality under the correspondence immediate to verify; a short remark comparing it to the classical Minty–Browder condition would help readers.
  2. [Main correspondence theorem] In the statement of the main correspondence theorem, the precise construction mapping an MPVF to the induced operator on L² random variables (and the inverse) should be displayed as a numbered display or algorithm; this is the load-bearing step and currently appears only in prose.
  3. [Mean-field limit section] The claim that the flow provides a mean-field characterization of discrete particle systems should include a precise statement of the initial-data convergence assumption (e.g., empirical measures converging in W₂) and the topology in which the limit is taken.
  4. [Notation] Notation for the multivalued field (bold F) and the single-valued case should be introduced once and used consistently; a short table of symbols would eliminate occasional ambiguity between F and its selections.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the careful reading, the accurate summary of our contributions, and the positive assessment leading to a recommendation of minor revision. No specific major comments were raised in the report.

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper establishes a bijective correspondence between totally dissipative MPVFs on the Wasserstein space and law-invariant maximal dissipative operators on L^2, then transfers standard existence/uniqueness results from Hilbert-space theory. No step reduces a claimed prediction or uniqueness result to a fitted input, self-definition, or load-bearing self-citation chain; the central bijection is presented as a theorem whose proof is independent of the target conclusions, and the abstract and reader's summary indicate the derivation remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 1 invented entities

The paper introduces a new classification of vector fields resting on standard background from functional analysis and optimal transport; no free parameters are fitted, and the new entity is the defined class itself with no independent evidence supplied beyond the correspondence.

axioms (2)
  • standard math Standard properties of the 2-Wasserstein metric, Hilbert space inner products, and dissipativity inequalities
    Invoked to define total dissipativity and establish the correspondence with L2 operators.
  • domain assumption Existence of maximal dissipative extensions in Hilbert spaces
    Used to guarantee a unique maximal totally dissipative version from discrete approximations.
invented entities (1)
  • totally dissipative multivalued probability vector field (MPVF) no independent evidence
    purpose: To classify vector fields generating dissipative evolutions on Wasserstein space that admit the Hilbert-space correspondence
    Newly introduced concept whose properties are defined and studied in the paper; no external falsifiable handle is provided.

pith-pipeline@v0.9.0 · 5777 in / 1526 out tokens · 40419 ms · 2026-05-24T08:55:33.286020+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Continuous transformations of probability measures and their transport representations

    math.FA 2026-04 unverdicted novelty 7.0

    Lipschitz continuous transformations F of probability measures w.r.t. Wasserstein distance admit continuous transport maps f(·,μ) such that F(μ) = f(·,μ)_# μ.

Reference graph

Works this paper leans on

51 extracted references · 51 canonical work pages · cited by 1 Pith paper

  1. [1]

    Ambrosio, M

    L. Ambrosio, M. Fornasier, M. Morandotti, and G. Savar´ e . Spatially inhomogeneous evolutionary games. Communications on Pure and Applied Mathematics , 74(7):1353–1402, 2021

    work page 2021

  2. [2]

    Ambrosio, N

    L. Ambrosio, N. Gigli, and G. Savar´ e. Gradient Flows In Metric Spaces and in the Space of Probabili ty Measures. Lectures in Mathematics. ETH Z¨ urich. Birkh¨ auser Basel,2008

    work page 2008

  3. [3]

    H. Attouch. Familles d’op´ erateurs maximaux monotones et mesurabilite. Ann. Mat. Pura Appl. (4) , 120:35– 111, 1979

    work page 1979

  4. [4]

    Viscosity solutions of centralized control problems in measure spaces

    Aussedat, Averil, Jerhaoui, Othmane, and Zidani, Hasna a. Viscosity solutions of centralized control problems in measure spaces. ESAIM: COCV , 30:91, 2024

    work page 2024

  5. [5]

    Averboukh and D

    Y. Averboukh and D. Khlopin. Pontryagin maximum princip le for the deterministic mean field type optimal control problem via the lagrangian approach. Journal of Differential Equations , 430:113205, 2025

    work page 2025

  6. [6]

    Averboukh, A

    Y. Averboukh, A. Marigonda, and M. Quincampoix. Extrema l shift rule and viability property for mean field-type control systems. J. Optim. Theory Appl. , 189(1):244–270, 2021

    work page 2021

  7. [7]

    Y. V. Averboukh. A mean field type differential inclusion w ith upper semicontinuous right-hand side. Vestn. Udmurt. Univ. Mat. Mekh. Komp’yut. Nauki , 32(4):489–501, 2022

    work page 2022

  8. [8]

    Badreddine and H

    Z. Badreddine and H. Frankowska. Solutions to Hamilton- Jacobi equation on a Wasserstein space. Calc. Var. Partial Differential Equations , 61(1):Paper No. 9, 41, 2022

    work page 2022

  9. [9]

    Badreddine and H

    Z. Badreddine and H. Frankowska. Viability and invarian ce of systems on metric spaces. Nonlinear Analysis, 225:113–133, 2022

    work page 2022

  10. [10]

    H. H. Bauschke and X. Wang. The kernel average for two con vex functions and its application to the extension and representation of monotone operators. Trans. Amer. Math. Soc. , 361(11):5947–5965, 2009

    work page 2009

  11. [11]

    H. H. Bauschke and X. Wang. Firmly nonexpansive and Kirs zbraun-Valentine extensions: a constructive approach via monotone operator theory. In Nonlinear analysis and optimization I. Nonlinear analysis , volume 513 of Contemp. Math. , pages 55–64. Amer. Math. Soc., Providence, RI, 2010

    work page 2010

  12. [12]

    B´ enilan

    P. B´ enilan. Solutions int´ egrales d’´ equations d’´ evolution dans un espace de Banach. C. R. Acad. Sci. Paris S´ er. A-B, 274:A47–A50, 1972

    work page 1972

  13. [13]

    Bonnet and H

    B. Bonnet and H. Frankowska. Differential inclusions in Wasserstein spaces: the Cauchy-Lipschitz framework. J. Differential Equations , 271:594–637, 2021

    work page 2021

  14. [14]

    Bonnet and H

    B. Bonnet and H. Frankowska. Necessary optimality cond itions for optimal control problems in Wasserstein spaces. Appl. Math. Optim. , 84(suppl. 2):S1281–S1330, 2021

    work page 2021

  15. [15]

    Bonnet and H

    B. Bonnet and H. Frankowska. Viability and exponential ly stable trajectories for differential inclusions in wasserstein spaces. In 2022 IEEE 61st Conference on Decision and Control (CDC) , pages 5086–5091, 2022

    work page 2022

  16. [16]

    Bonnet-Weill and H

    B. Bonnet-Weill and H. Frankowska. Carath´ eodory theo ry and a priori estimates for continuity inclusions in the space of probability measures. Nonlinear Anal. , 247:Paper No. 113595, 32, 2024

    work page 2024

  17. [17]

    H. Brezis. Operateurs maximaux monotones et semi-groupes de contract ions dans les espaces de Hilbert , volume No. 5 of North-Holland Mathematics Studies . North-Holland Publishing Co., Amsterdam-London,

    work page

  18. [18]

    Notas de Matematica, No. 50. [Mathematical Notes]

    work page

  19. [19]

    H. Brezis. Functional Analysis, Sobolev Spaces and Partial Differentia l Equations. Universitext. Springer New York, 2010

    work page 2010

  20. [20]

    J. A. Ca˜ nizo, J. A. Carrillo, and J. Rosado. A well-pose dness theory in measures for some kinetic models of collective motion. Math. Models Methods Appl. Sci. , 21(3):515–539, 2011

    work page 2011

  21. [21]

    Camilli, G

    F. Camilli, G. Cavagnari, R. De Maio, and B. Piccoli. Sup erposition principle and schemes for measure differential equations. Kinet. Relat. Models , 14(1):89–113, 2021

    work page 2021

  22. [22]

    Capuani and A

    R. Capuani and A. Marigonda. Constrained mean field game s equilibria as fixed point of random lifting of set- valued maps. IF AC-PapersOnLine, 55(30):180–185, 2022. 25th International Symposium on Ma thematical Theory of Networks and Systems MTNS 2022

    work page 2022

  23. [23]

    Cardaliaguet

    P. Cardaliaguet. Notes on mean field games. From P.-L. Li ons’ lectures at College de France, link to the notes

    work page

  24. [24]

    Carmona and F

    R. Carmona and F. Delarue. Probabilistic theory of mean field games with applications. I, volume 83 of Probability Theory and Stochastic Modelling . Springer, Cham, 2018. Mean field FBSDEs, control, and games

    work page 2018

  25. [25]

    Cavagnari, S

    G. Cavagnari, S. Lisini, C. Orrieri, and G. Savar´ e. Lag rangian, Eulerian and Kantorovich formulations of multi-agent optimal control problems: equivalence and g amma-convergence. J. Differential Equations , 322:268–364, 2022

    work page 2022

  26. [26]

    Cavagnari, A

    G. Cavagnari, A. Marigonda, and B. Piccoli. Generalize d dynamic programming principle and sparse mean- field control problems. J. Math. Anal. Appl. , 481(1):123–137, 45, 2020

    work page 2020

  27. [27]

    Cavagnari, A

    G. Cavagnari, A. Marigonda, and M. Quincampoix. Compat ibility of state constraints and dynamics for multiagent control systems. J. Evol. Equ. , 21(4):4491–4537, 2021

    work page 2021

  28. [28]

    Cavagnari, G

    G. Cavagnari, G. Savar´ e, and G. E. Sodini. Dissipative probability vector fields and generation of evolution semigroups in Wasserstein spaces. Probab. Theory Related Fields , 185(3-4):1087—-1182, 2023. 86 GIULIA CA V AGNARI, GIUSEPPE SA V AR ´E, AND GIACOMO ENRICO SODINI

    work page 2023

  29. [29]

    Cavagnari, G

    G. Cavagnari, G. Savar´ e, and G. E. Sodini. Extension of monotone operators and Lipschitz maps invariant for a group of isometries. Canad. J. Math. , 77(1):149–186, 2025

    work page 2025

  30. [30]

    Chizat and F

    L. Chizat and F. Bach. On the global convergence of gradi ent descent for over-parameterized models using optimal transport. Advances in neural information processing systems , pages 3036–3056, 2018

    work page 2018

  31. [31]

    M. R. D’Orsogna, Y.-L. Chuang, A. L. Bertozzi, and L. Cha yes. Self-propelled particles with soft-core inter- actions: Patterns, stability, and collapse. Phys. Rev. Lett. , 96:104302–1/4, 2006

    work page 2006

  32. [32]

    Fornasier, S

    M. Fornasier, S. Lisini, C. Orrieri, and G. Savar´ e. Mea n-field optimal control as gamma-limit of finite agent controls. European J. Appl. Math. , 30(6):1153–1186, 2019

    work page 2019

  33. [33]

    Fornasier, G

    M. Fornasier, G. Savar´ e, and G. E. Sodini. Density of su balgebras of Lipschitz functions in metric Sobolev spaces and applications to Wasserstein Sobolev spaces. J. Funct. Anal. , 285(11):Paper No. 110153, 76, 2023

    work page 2023

  34. [34]

    Gangbo and A

    W. Gangbo and A. Tudorascu. On differentiability in the W asserstein space and well-posedness for Hamilton- Jacobi equations. J. Math. Pures Appl. (9) , 125:119–174, 2019

    work page 2019

  35. [35]

    C. Jimenez. Equivalence between strict viscosity solu tion and viscosity solution in the Wasserstein space and regular extension of the Hamiltonian in L2 P. J. Convex Anal. , 31(2):619–670, 2024

    work page 2024

  36. [36]

    Jimenez, A

    C. Jimenez, A. Marigonda, and M. Quincampoix. Optimal c ontrol of multiagent systems in the Wasserstein space. Calc. Var. Partial Differential Equations , 59(2):Paper No. 58, 45, 2020

    work page 2020

  37. [38]

    Jordan, D

    R. Jordan, D. Kinderlehrer, and F. Otto. The variationa l formulation of the Fokker-Planck equation. SIAM J. Math. Anal. , 29(1):1–17, 1998

    work page 1998

  38. [39]

    P.-L. Lions. Th´ eorie des jeux ` a champs moyen et applic ations. 2007

    work page 2007

  39. [40]

    R. J. McCann. A convexity principle for interacting gas es. Adv. Math. , 128(1):153–179, 1997

    work page 1997

  40. [41]

    Naldi and G

    E. Naldi and G. Savar´ e. Weak topology and Opial propert y in Wasserstein spaces, with applications to gradient flows and proximal point algorithms of geodesicall y convex functionals. Atti Accad. Naz. Lincei Rend. Lincei Mat. Appl. , 32(4):725–750, 2021

    work page 2021

  41. [42]

    Natile and G

    L. Natile and G. Savar´ e. A Wasserstein approach to the o ne-dimensional sticky particle system. SIAM J. Math. Anal. , 41(4):1340–1365, 2009

    work page 2009

  42. [43]

    R. H. Nochetto and G. Savar´ e. Nonlinear evolution gove rned by accretive operators in Banach spaces: error control and applications. Math. Models Methods Appl. Sci. , 16(3):439–477, 2006

    work page 2006

  43. [44]

    R. H. Nochetto, G. Savar´ e, and C. Verdi. A posteriori er ror estimates for variable time-step discretizations of nonlinear evolution equations. Comm. Pure Appl. Math. , 53(5):525–589, 2000

    work page 2000

  44. [45]

    F. Otto. The geometry of dissipative evolution equatio ns: the porous medium equation. Comm. Partial Differential Equations , 26(1-2):101–174, 2001

    work page 2001

  45. [46]

    G. Parker. Some convexity criteria for differentiable f unctions on the 2-Wasserstein space. Bull. Lond. Math. Soc., 56(5):1839–1858, 2024

    work page 2024

  46. [47]

    B. Piccoli. Measure differential inclusions. 2018 IEEE Conference on Decision and Control (CDC) , pages 1323–1328, 2018

    work page 2018

  47. [48]

    B. Piccoli. Measure differential equations. Arch. Ration. Mech. Anal. , 233(3):1289–1317, 2019

    work page 2019

  48. [49]

    B. Piccoli. Control of multi-agent systems: results, o pen problems, and applications. Open Math., 21(1):Paper No. 20220585, 26, 2023

    work page 2023

  49. [50]

    Pogodaev

    N. Pogodaev. Optimal control of continuity equations. NoDEA Nonlinear Differential Equations Appl. , 23(2):Art. 21, 24, 2016

    work page 2016

  50. [51]

    L. Qi. Uniqueness of the maximal extension of a monotone operator. Nonlinear Anal., Theory Methods Appl. , 7:325–332, 1983

    work page 1983

  51. [52]

    Sznitman

    A.-S. Sznitman. Topics in propagation of chaos. In ´Ecole d’ ´Et´ e de Probabilit´ es de Saint-Flour XIX—1989, volume 1464 of Lecture Notes in Math. , pages 165–251. Springer, Berlin, 1991. Giulia Cavagnari: Politecnico di Milano, Dipartimento di M atematica, Piazza Leonardo Da Vinci 32, 20133 Milano (Italy) Email address : giulia.cavagnari@polimi.it Gius...

    work page 1989