pith. sign in

arxiv: 2506.10509 · v2 · submitted 2025-06-12 · 🧮 math.NA · cs.NA

A semi-Lagrangian scheme for First-Order Mean Field Games based on monotone operators

Elisabetta Carlini , Valentina Coscetti This is my paper

Pith reviewed 2026-05-19 09:56 UTC · model grok-4.3

classification 🧮 math.NA cs.NA
keywords semi-Lagrangian schememean field gamesmonotone operatorsconvergence analysisnumerical methodsfirst-order systemslearning value algorithmpolicy iteration
0
0 comments X

The pith

A semi-Lagrangian scheme for first-order Mean Field Games converges to weak solutions by exploiting monotonicity of the operators.

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

The paper develops a numerical scheme to approximate solutions of first-order time-dependent non-local Mean Field Games. It builds the discretization so that a monotonicity property of the operators carries through to the discrete level. This property is then used to prove that the scheme converges to a weak solution of the continuous system. Readers care because the method comes with both a convergence guarantee and practical algorithms to compute the discrete solutions, including an accelerated variant.

Core claim

The authors construct a semi-Lagrangian scheme for first-order, time-dependent, and non-local Mean Field Games. The convergence of the scheme to a weak solution of the system is analyzed by exploiting a key monotonicity property. To solve the resulting discrete problem, they implement a Learning Value Algorithm, prove its convergence, and propose an acceleration strategy based on a Policy iteration method. Numerical experiments validate the effectiveness of the proposed schemes.

What carries the argument

The semi-Lagrangian scheme based on monotone operators, which discretizes the MFG system while preserving the monotonicity needed for the convergence proof.

If this is right

  • The discrete problem can be solved with a Learning Value Algorithm whose convergence is established.
  • A Policy iteration acceleration improves computational performance while preserving the underlying convergence.
  • The scheme applies to time-dependent non-local interactions and produces numerical solutions that match the expected weak limits.
  • The monotonicity-based analysis provides a template for proving convergence of similar discretizations.

Where Pith is reading between the lines

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

  • The same monotonicity argument could be tested on MFG systems with different non-local coupling terms to see how far the convergence guarantee extends.
  • Pairing the scheme with adaptive mesh refinement might reduce the cost of high-dimensional examples without losing the proven convergence.
  • The acceleration technique suggests that policy iteration can be grafted onto other monotone-operator discretizations for faster solves.

Load-bearing premise

The key monotonicity property of the operators holds for the non-local first-order mean field game system under consideration.

What would settle it

A concrete first-order non-local MFG for which the computed semi-Lagrangian solutions fail to approach any weak solution of the continuous system as the time step and spatial mesh are refined.

Figures

Figures reproduced from arXiv: 2506.10509 by Elisabetta Carlini, Valentina Coscetti.

Figure 1
Figure 1. Figure 1: Example 5.2. Initial density for d = 1 (left) and for d = 2 (center and right). 5.2.1 Case d = 1 with state-independent quadratic Hamiltonian In this example, we consider the above problem with quadratic Hamiltonian H(x, p) = |p| 2/2 and dimension d = 1. In the following simulations, we set T = 4 and ∆x = 2.5 · 10−2 . We 23 [PITH_FULL_IMAGE:figures/full_fig_p023_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Example 5.2.1. Time-evolution of the density for [PITH_FULL_IMAGE:figures/full_fig_p024_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Example 5.2.2. Time-evolution of the density, [PITH_FULL_IMAGE:figures/full_fig_p025_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Example 5.2.2. Time-evolution of the density, [PITH_FULL_IMAGE:figures/full_fig_p025_4.png] view at source ↗
read the original abstract

We construct a semi-Lagrangian scheme for first-order, time-dependent, and non-local Mean Field Games. The convergence of the scheme to a weak solution of the system is analyzed by exploiting a key monotonicity property. To solve the resulting discrete problem, we implement a Learning Value Algorithm, prove its convergence, and propose an acceleration strategy based on a Policy iteration method. Finally, we present numerical experiments that validate the effectiveness of the proposed schemes and show that the accelerated version significantly improves performance.

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

1 major / 2 minor

Summary. The manuscript constructs a semi-Lagrangian scheme for first-order, time-dependent, non-local Mean Field Games. Convergence of the scheme to a weak solution is analyzed by exploiting a monotonicity property of the discrete operators. The resulting discrete system is solved via a Learning Value Algorithm whose convergence is proven, together with an acceleration strategy based on Policy Iteration. Numerical experiments are presented to validate the schemes and demonstrate performance gains from acceleration.

Significance. If the central convergence claim holds, the work supplies a monotone discretization for non-local first-order MFGs, a setting that appears in applications such as crowd motion and mean-field control. The combination of monotonicity-based analysis, a provably convergent solver, and an accelerated variant offers both theoretical and practical contributions to numerical methods for these systems.

major comments (1)
  1. [§4] §4 (Convergence analysis): The proof that the scheme converges to a weak solution rests on the discrete operators being monotone, including after discretization of the non-local coupling term. No explicit verification lemma or set of hypotheses on the interaction kernel is supplied to confirm that monotonicity is preserved for general non-local terms; if the kernel lacks sufficient positivity or monotonicity-preserving structure, the passage to the limit via monotonicity may fail. This assumption is load-bearing for the main theorem.
minor comments (2)
  1. [Abstract] The abstract states that the accelerated version 'significantly improves performance' without specifying the quantitative metrics (iteration count, CPU time, or residual reduction).
  2. [§3] Notation for the local Hamiltonian versus the non-local integral term could be made more distinct when the discrete operators are first introduced.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the careful reading and constructive feedback on our manuscript. We have addressed the major comment on the convergence analysis by strengthening the presentation of the monotonicity properties.

read point-by-point responses

  1. Referee: [§4] §4 (Convergence analysis): The proof that the scheme converges to a weak solution rests on the discrete operators being monotone, including after discretization of the non-local coupling term. No explicit verification lemma or set of hypotheses on the interaction kernel is supplied to confirm that monotonicity is preserved for general non-local terms; if the kernel lacks sufficient positivity or monotonicity-preserving structure, the passage to the limit via monotonicity may fail. This assumption is load-bearing for the main theorem.

    Authors: We agree that an explicit verification strengthens the argument. In the revised manuscript we add Lemma 4.3, which states that if the interaction kernel K is non-negative, continuous, and Lipschitz continuous in its second argument (standard assumptions for non-local MFGs arising in crowd motion), then the semi-Lagrangian discretization of the non-local term preserves monotonicity. The proof of the lemma follows directly from the positivity of the discrete weights and the monotonicity of the coupling function; it is inserted immediately before the main convergence theorem. This clarifies the hypotheses under which the passage to the limit holds while covering the kernels used in our numerical examples. revision: yes

Circularity Check

0 steps flagged

Monotonicity property is an external structural assumption; derivation chain remains self-contained

full rationale

The paper constructs a semi-Lagrangian scheme for non-local first-order MFGs and analyzes convergence by exploiting a monotonicity property of the discrete operators. This property is presented as holding for the underlying continuous system and is used to pass to the limit, without being defined in terms of the scheme itself or fitted from its outputs. No self-citations, ansatzes, or uniqueness theorems from prior author work are invoked as load-bearing steps that reduce the central claim to a tautology. The Learning Value Algorithm and Policy iteration acceleration are analyzed separately with their own convergence proofs. The overall derivation does not reduce to its inputs by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claims rest on the existence and applicability of a monotonicity property for the operators in the MFG system; no free parameters or invented entities are mentioned in the abstract.

axioms (1)
  • domain assumption The operators satisfy a key monotonicity property that enables convergence of the semi-Lagrangian scheme to a weak solution.
    Invoked explicitly for the convergence analysis of the discrete scheme.

pith-pipeline@v0.9.0 · 5605 in / 1316 out tokens · 38030 ms · 2026-05-19T09:56:00.851599+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

40 extracted references · 40 canonical work pages

  1. [1]

    Achdou and I

    Y. Achdou and I. Capuzzo-Dolcetta. Mean field games: numerical methods.SIAM J. Numer. Anal., 48(3):1136–1162, 2010

    work page 2010

  2. [2]

    Achdou, P

    Y. Achdou, P. Cardaliaguet, F. c. Delarue, A. Porretta, and F. Santambrogio.Mean field games, volume 2281 ofLecture Notes in Mathematics. Springer, Cham; Centro Internazionale Matematico Estivo (C.I.M.E.), Florence, 2020. Notes from the CIME School held in Cetraro, June 2019, Edited by Cardaliaguet and Porretta, Fondazione CIME/CIME Foundation Subseries

    work page 2020

  3. [3]

    Achdou and A

    Y. Achdou and A. Porretta. Convergence of a finite difference scheme to weak solutions of the system of partial differential equations arising in mean field games.SIAM J. Numer. Anal., 54(1):161–186, 2016

    work page 2016

  4. [4]

    Almulla, R

    N. Almulla, R. Ferreira, and D. Gomes. Two numerical approaches to stationary mean-field games. Dyn. Games Appl., 7(4):657–682, 2017

    work page 2017

  5. [5]

    Ambrosio, N

    L. Ambrosio, N. Gigli, and G. Savaré.Gradient flows in metric spaces and in the space of probability measures. Lectures in Mathematics ETH Zürich. Birkhäuser Verlag, Basel, second edition, 2008

    work page 2008

  6. [6]

    Angiuli, J.-P

    A. Angiuli, J.-P. Fouque, and M. Laurière. Unified reinforcement Q-learning for mean field game and control problems.Math. Control Signals Systems, 34(2):217–271, 2022

    work page 2022

  7. [7]

    Ashrafyan and D

    Y. Ashrafyan and D. Gomes. A Fully-Discrete Semi-Lagrangian Scheme for a Price Formation MFG Model.Dyn. Games Appl., 15(2):503–533, 2025

    work page 2025

  8. [8]

    Bardi and I

    M. Bardi and I. Capuzzo-Dolcetta.Optimal control and viscosity solutions of Hamilton- Jacobi-Bellman equations. Systems & Control: Foundations & Applications. Birkhäuser Boston, Inc., Boston, MA, 1997. With appendices by Maurizio Falcone and Pierpaolo Soravia

    work page 1997

  9. [9]

    J. F. Bonnans, K. Liu, and L. Pfeiffer. Error estimates of a theta-scheme for second-order mean field games.ESAIM Math. Model. Numer. Anal., 57(4):2493–2528, 2023

    work page 2023

  10. [10]

    Cacace, F

    S. Cacace, F. Camilli, and A. Goffi. A policy iteration method for mean field games.ESAIM Control Optim. Calc. Var., 27:Paper No. 85, 19, 2021

    work page 2021

  11. [11]

    Calzola, E

    E. Calzola, E. Carlini, and F. J. Silva. A high-order scheme for mean field games.J. Comput. Appl. Math., 445:Paper No. 115769, 19, 2024. 27

    work page 2024

  12. [12]

    Camilli and F

    F. Camilli and F. Silva. A semi-discrete approximation for a first order mean field game problem. Netw. Heterog. Media, 7(2):263–277, 2012

    work page 2012

  13. [13]

    Cardaliaguet.Notes on Mean Field Games: From P.-L

    P. Cardaliaguet.Notes on Mean Field Games: From P.-L. Lions’ lectures at Collège de France. 2010

    work page 2010

  14. [14]

    Cardaliaguet and S

    P. Cardaliaguet and S. Hadikhanloo. Learning in mean field games: the fictitious play. ESAIM Control Optim. Calc. Var., 23(2):569–591, 2017

    work page 2017

  15. [15]

    Carlini and F

    E. Carlini and F. J. Silva. A fully discrete semi-Lagrangian scheme for a first order mean field game problem.SIAM J. Numer. Anal., 52(1):45–67, 2014

    work page 2014

  16. [16]

    Carlini and F

    E. Carlini and F. J. Silva. A semi-lagrangian scheme for a degenerate second order mean field game system.Discrete and Continuous Dynamical Systems, 35(9):4269–4292, 2015

    work page 2015

  17. [17]

    Carlini and F

    E. Carlini and F. J. Silva. On the discretization of some nonlinear Fokker-Planck-Kolmogorov equations and applications.SIAM J. Numer. Anal., 56(4):2148–2177, 2018

    work page 2018

  18. [18]

    Carlini, F

    E. Carlini, F. J. Silva, and A. Zorkot. A Lagrange-Galerkin scheme for first order mean field game systems.SIAM J. Numer. Anal., 62(1):167–198, 2024

    work page 2024

  19. [19]

    Carmona and M

    R. Carmona and M. Laurière. Convergence analysis of machine learning algorithms for the numerical solution of mean field control and games I: The ergodic case.SIAM J. Numer. Anal., 59(3):1455–1485, 2021

    work page 2021

  20. [20]

    Carmona and M

    R. Carmona and M. Laurière. Convergence analysis of machine learning algorithms for the numerical solution of mean field control and games: II—The finite horizon case.Ann. Appl. Probab., 32(6):4065–4105, 2022

    work page 2022

  21. [21]

    Chowdhury, O

    I. Chowdhury, O. Ersland, and E. Jakobsen. On numerical approximations of fractional and nonlocal mean field games.Found Comput Math, 23:1381–1431, 2023

    work page 2023

  22. [22]

    Falcone and R

    M. Falcone and R. Ferretti. Semi-Lagrangian approximation schemes for linear and Hamilton-Jacobi equations. Society for Industrial and Applied Mathematics (SIAM), Philadelphia, PA, 2014

    work page 2014

  23. [23]

    Ferreira and D

    R. Ferreira and D. Gomes. Existence of weak solutions to stationary mean-field games through variational inequalities.SIAM J. Math. Anal., 50(6):5969–6006, 2018

    work page 2018

  24. [24]

    Gianatti and F

    J. Gianatti and F. J. Silva. Approximation of deterministic mean field games with control- affine dynamics.Found. Comput. Math., 24(6):2017–2061, 2024

    work page 2017

  25. [25]

    Gianatti, F

    J. Gianatti, F. J. Silva, and A. Zorkot. Approximation of deterministic mean field games under polynomial growth conditions on the data.J. Dyn. Games, 11(2):131–149, 2024

    work page 2024

  26. [26]

    O. Guéant. Mean field games equations with quadratic Hamiltonian: a specific approach. Math. Models Methods Appl. Sci., 22(9):1250022, 37, 2012

    work page 2012

  27. [27]

    Hadikhanloo and F

    S. Hadikhanloo and F. J. Silva. Finite mean field games: fictitious play and convergence to a first order continuous mean field game.J. Math. Pures Appl. (9), 132:369–397, 2019

    work page 2019

  28. [28]

    Huang, R

    M. Huang, R. P. Malhamé, and P. E. Caines. Large population stochastic dynamic games: closed-loop McKean-Vlasov systems and the Nash certainty equivalence principle.Commun. Inf. Syst., 6(3):221–251, 2006. 28

    work page 2006

  29. [29]

    Lasry and P.-L

    J.-M. Lasry and P.-L. Lions. Jeux à champ moyen. I. Le cas stationnaire.C. R. Math. Acad. Sci. Paris, 343(9):619–625, 2006

    work page 2006

  30. [30]

    Lasry and P.-L

    J.-M. Lasry and P.-L. Lions. Jeux à champ moyen. II. Horizon fini et contrôle optimal.C. R. Math. Acad. Sci. Paris, 343(10):679–684, 2006

    work page 2006

  31. [31]

    Lasry and P.-L

    J.-M. Lasry and P.-L. Lions. Mean field games.Jpn. J. Math., 2(1):229–260, 2007

    work page 2007

  32. [32]

    Laurière

    M. Laurière. Numerical methods for mean field games and mean field type control. InMean field games, volume 78 ofProc. Sympos. Appl. Math., pages 221–282. Amer. Math. Soc., Providence, RI, 2021

    work page 2021

  33. [33]

    P.-L. Lions. Cours au college de france.Available at www. college-de-france. fr, 2007

    work page 2007

  34. [34]

    Liu and L

    K. Liu and L. Pfeiffer. A mesh-independent method for second-order potential mean field games. IMA J. Numer. Anal., 45(2):1226–1266, 2025

    work page 2025

  35. [35]

    S. Liu, M. Jacobs, W. Li, L. Nurbekyan, and S. J. Osher. Computational methods for first- order nonlocal mean field games with applications.SIAM J. Numer. Anal., 59(5):2639–2668, 2021

    work page 2021

  36. [36]

    Nurbekyan and J

    L. Nurbekyan and J. Saúde. Fourier approximation methods for first-order nonlocal mean- field games.Port. Math., 75(3-4):367–396, 2018

    work page 2018

  37. [37]

    Y. A. P. Osborne and I. Smears. Analysis and numerical approximation of stationary second- order mean field game partial differential inclusions.SIAM J. Numer. Anal., 62(1):138–166, 2024

    work page 2024

  38. [38]

    Y. A. P. Osborne and I. Smears. Finite element approximation of time-dependent mean field games with nondifferentiable Hamiltonians.Numer. Math., 157(1):165–211, 2025

    work page 2025

  39. [39]

    Y. A. P. Osborne and I. Smears. Rates of convergence of finite element approximations of second-order mean field games with nondifferentiable hamiltonians, 2025

    work page 2025

  40. [40]

    Quarteroni and A

    A. Quarteroni and A. Valli. Numerical approximation of partial differential equations, volume 23 ofSpringer Series in Computational Mathematics. Springer-Verlag, Berlin, 1994. 29

    work page 1994