Pith. sign in

REVIEW 4 major objections 5 minor 38 references

Existence of Solutions and Selection Problem for Quasi-stationary Contact Mean Field Games

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Pith's one-line read As the discount parameter goes to zero, normalized value functions of quasi-stationary contact mean field games converge to the Peierls barrier anchored at the unique Aubry point, and densities converge to the critical mean field game.

desk verdict First existence and selection result for quasi-stationary first-order contact MFG, with an explicit Peierls barrier formula; the selection theorem is honest about its strong singleton Aubry-set assumption, and one technical lemma needs real fixing, but the paper deserves a serious referee. read the letter →

arxiv 2507.10112 v1 pith:OPUG4KD6 submitted 2025-07-14 math.AP math.DS

classification math.APmath.DS MSC 35Q8937J5149N80
keywords contactmeanfieldgamesquasi-stationaryselectionproblemvanishingdiscountviscositysolutionsweakKAMtheoryPeierlsbarrierAubryset
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

Quasi-stationary mean field games describe a large population whose agents, at each time, solve a stationary Hamilton-Jacobi equation using only the current population distribution, while the distribution itself evolves by the continuity equation. The paper proves existence of weak solutions for the contact version of these games, where the Hamiltonian also depends on the value function itself, without the reversibility assumption used in earlier work. It then studies a discounted approximation: for each $\lambda>0$ the stationary equation is $H(x,\lambda u^\lambda,Du^\lambda)=F(x,m^\lambda(t))+c(m^\lambda(t))$ with $c$ the Mañé critical value, and the question is what happens as $\lambda\to 0$. Under a hypothesis that the critical Hamiltonian has a single Aubry point for every admissible density flow, the paper proves that the normalized value functions $u^\lambda-u^\lambda(\bar x,\cdot)$ converge uniformly to $\bar v-\bar v(\bar x,\cdot)$, the densities $m^\lambda$ converge uniformly, and the limit is identified explicitly as the Peierls barrier difference anchored at that Aubry point. If correct, this solves the vanishing-discount selection problem for quasi-stationary contact mean field games: the limit is unique and computable.

What carries the argument

The argument is carried by weak KAM theory for contact Hamiltonian systems, together with PDE estimates for viscosity solutions. For each admissible measure flow $m(t)$, one introduces the auxiliary Hamiltonian $H(x,0,p)-F(x,m(t))$ and its projected Aubry set $A_m(t)$; assumption (H4) forces $A_m(t)=\{x_m\}$ for every $t$, which makes the critical equation $H(x,0,Dv)=F(x,m)+c(m)$ have a unique viscosity solution up to additive constants. The central objects are the Peierls barrier $h_m(x,y)$, the minimal action between two points at the critical energy, and the flow $\Phi_m$ generated by the Hamiltonian derivative $D_pH(x,u_m,Du_m)$, whose pushforward solves the continuity equation. The proof establishes a uniform modulus of continuity for the normalized solutions $u^\lambda(x,t)-u^\lambda(x_m,t)$ (Proposition 5.4) by a contradiction argument that excludes both $\lambda\to 0$ and $\lambda$ staying away from zero, then uses stability of viscosity solutions and the uniqueness from (H4) to identify the limit as the Peierls barrier difference.

What would settle it

On the one-dimensional torus, take $H(x,u,p)=\beta u+\frac12 p^2-F(x,m)$ with $F(x,m)=\int \kappa(x,y)\,dm(y)$, $\kappa(0,y)\equiv 0$ and $\kappa(x,y)>0$ for $x\neq 0$, so the Aubry set is $\{0\}$ for every $m$; solve the discounted system numerically for small $\lambda$ and compare $u^\lambda(x,t)-u^\lambda(0,t)$ with $h_{\bar\mu(t)}(0,x)-h_{\bar\mu(t)}(0,0)$. Any disagreement beyond discretization error, or any dependence of the numerical limit on the choice of subsequence, would falsify the identification.

Watch

Extended reading notes

Core claim

The central claim is Theorem 2.3. For each $\lambda>0$, let $(u^\lambda,m^\lambda)$ be a weak solution of the discounted quasi-stationary system $H(x,\lambda u^\lambda,Du^\lambda)=F(x,m^\lambda(t))+c(m^\lambda(t))$, where $c(m^\lambda(t))$ is the Mañé critical value of $H(x,0,p)-F(x,m^\lambda(t))$, together with the continuity equation and initial condition $m^\lambda(0)=m_0$. Under assumptions (H1), (H2), (H3'), (H4), (F1)-(F3), (P), the paper proves that, up to subsequences, $u^\lambda-u^\lambda(\bar x,\cdot)$ converges uniformly to $\bar v-\bar v(\bar x,\cdot)$ and $m^\lambda$ converges uniformly in the $d_1$ distance to $\bar\mu$, where $(\bar v,\bar\mu)$ is a weak solution of the critical system $H(x,0,Dv)=F(x,\bar\mu(t))+c(\bar\mu(t))$, $\partial_t \bar\mu - \mathrm{div}(\bar\mu D_pH(x,0,D\bar v))=0$, $\bar\mu(0)=m_0$. The identified limit satisfies $\bar v-\bar v(\bar x,\cdot)=h_{\bar\mu(\cdot)}(x_{\bar\mu},\cdot)-h_{\bar\mu(\cdot)}(x_{\bar\mu},\bar x)$, where $h$ is the Peierls barrier of $H(x,0,p)-F(x,\bar\mu(t))$ and $x_{\bar\mu}$ is its unique Aubry point. Thus the vanishing-discount limit is unique up to additive constants and is completely determined by the Aubry set.

Load-bearing premise

The selection theorem rests on assumption (H4): for every admissible density flow, the set of distinguished points of the critical Hamiltonian (its Aubry set) is exactly one point at every time, which forces the limit value function to be independent of the subsequence.

Editorial extensions

If this is right

  • Every subsequential limit of the discounted pair $(u^\lambda,m^\lambda)$ is the same pair $(\bar v,\bar\mu)$, so the selection problem has a definite answer rather than a family of possible limits.
  • The selected value function can be computed directly from the Peierls barrier of the critical Hamiltonian, without first solving the discounted system.
  • The limiting density $\bar\mu$ is the unique weak solution of the continuity equation driven by $D_pH(x,0,D\bar v)$, so the critical system (qMFG0) is the actual vanishing-discount limit system.
  • The existence theorem provides weak solutions of quasi-stationary contact MFG without assuming the reversibility condition used in earlier stationary contact MFG results.
  • The equicontinuity estimate for $u^\lambda-u^\lambda(\bar x,\cdot)$ is the key mechanism that turns convergence of densities into convergence of value functions.

Reading between the lines

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

  • The appendix's moving-singular-point examples suggest that without (H4), the vanishing-discount limit may fail to be unique; an editor's inference is that a general selection theory would need to track the full Mather measure rather than a single Aubry point.
  • A natural testable consequence: in the examples of Remark 4, where the Aubry point is $x=0$ for every density flow, the theorem predicts $u^\lambda(x,t)-u^\lambda(0,t)$ converges to $h_{\bar\mu(t)}(0,x)-h_{\bar\mu(t)}(0,0)$ for the selected $\bar\mu$, which can be checked numerically.
  • The proof only needs (H3') for the selection theorem, so the mechanism should extend to contact Hamiltonians with merely positive derivative in $u$, as long as the Aubry-set uniqueness (H4) holds; the Lipschitz regularity of $t\mapsto u(x,t)$ is lost but the normalized equicontinuity survives.
  • One could try to relax (H4) to a condition that the Aubry set is a singleton only on the support of the limiting flow, rather than on all admissible flows, since the compactness steps only need uniqueness at the limit measure.
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.

Referee Report

4 major / 5 minor

Summary. The paper studies a first-order quasi-stationary contact mean field game system in which, for each time t, u(·,t) solves a stationary contact Hamilton-Jacobi equation H(x,u,Du)=F(x,m(t)) and m solves a continuity equation driven by D_pH(x,u,Du). Under monotonicity assumptions (H1)–(H3) on H, the author proves existence of a weak solution (Theorem 2.2) by a Schauder fixed point argument on pushforwards of the initial measure m_0 along the characteristic flow of the Hamilton-Jacobi equation. For a λ-perturbed system with critical constant, the paper then studies the limit λ→0. The main selection result, Theorem 2.3, asserts that, under the stronger assumptions (H3’) and (H4), any subsequential limit of (u^λ−u^λ(x̄,·), m^λ) is a weak solution of the critical system (qMFG_0) and that the normalized value function is characterized by the Peierls barrier of the limiting Hamiltonian. The paper closes with an appendix on existence under weaker monotonicity and a discussion of non-uniqueness phenomena.

Significance. If the technical gaps are fixed, this is a meaningful contribution to the weak-KAM/PDE approach to first-order mean field games. The paper extends the quasi-stationary framework of Camilli–Marchi–Mendico to contact Hamiltonians, removes the reversibility assumption (R1) used in earlier stationary contact MFG results, and gives an explicit Peierls-barrier formula for the vanishing-discount limit under a singleton Aubry-set condition. The proof strategy is coherent and the paper is careful to separate the existence theory from the selection mechanism. The main strengths are the self-contained existence construction via the flow of the Hamilton-Jacobi equation and the explicit use of weak KAM quantities (Mañé critical value, Aubry set, Peierls barrier) to describe the limit.

major comments (4)
  1. [§4, Lemma 4.2] The proof of the lower bound |x−y| ≤ C_2 |Φ_m(x,t)−Φ_m(y,t)| is not justified as written. The velocity field is ∂H/∂p(x,u_m(t),Du_m(t)), and while u_m is uniformly semiconcave (Proposition 3.2), its gradient Du_m is not Lipschitz. The displayed estimate on ẋ−ẏ therefore cannot follow merely from “locally Lipschitz continuity of H and Proposition 3.2.” One needs a one-sided Lipschitz estimate on the velocity field coming from semiconcavity, as in the argument of [3, Lemma 4.3], or an alternative BV-regularity argument. This point is load-bearing because Lemma 4.2 is used to prove the density bound in Lemma 4.3, which in turn underpins the fixed point argument for Theorem 2.2 and the compactness of {m^λ} in Theorem 2.3.
  2. [§5, Proof of Theorem 2.3, item (c4)] The proof asserts that Du^λ converges a.e. to Dv̄ “due to the uniform semi-concavity of u^λ.” However, Theorem 2.3 is stated under (H1), (H2), (H3’), and uniform semiconcavity is proved in Proposition 3.2 only under the stronger assumption (H3). Appendix A, under (H3’), establishes only uniform boundedness and equi-Lipschitz continuity, not semiconcavity. Without a proof (or citation) of uniform semiconcavity under (H3’), the a.e. convergence of derivatives and the identification μ̄ = Φ(·,·)♯m_0 in Step (c5) are unsupported. This is a load-bearing gap in the central proof.
  3. [§5, Theorem 2.3 and Corollary 5.5] The paper does not prove uniqueness of the selected measure μ̄. Theorem 2.3 only asserts convergence up to a subsequence for the joint family (u^λ−u^λ(x̄,·), m^λ). Corollary 5.5 shows that once a full limit m of m^λ is fixed, the normalized value function has a unique limit, but different subsequences of {m^λ} could in principle converge to different μ̄. Appendix B explicitly documents non-uniqueness phenomena for (qMFG_0). The theorem and abstract should therefore be worded so that “selection” refers to the subsequential characterization obtained, not to uniqueness of the measure component; otherwise the stated result is stronger than what is proved.
  4. [§2, Assumption (H4), Remark 4, and Appendix B] Assumption (H4) is the true crux of the selection theorem, and it is very restrictive: it requires the projected Aubry set of H(x,0,p)−F(x,m(t)) to be a singleton for every continuous measure path. The paper’s own Appendix B shows that even when the projected Aubry set is a singleton, the singular set of the viscosity solution may move with the parameter, and when (H4) fails the Peierls-barrier characterization in Theorem 2.3 has no well-defined x_m. The statement that “all the examples presented in [11] fit our assumption (H4)” is not accompanied by a proof or reference checking the time-uniformity in (H4). The paper should either provide this verification or temper Remark 4, and it should state explicitly that the selection result is conditional on a condition strictly stronger than (R2).
minor comments (5)
  1. [§4, Lemma 4.2] The inequality for ẋ(r)−ẏ(r) is written with a bare “≤” for vectors; the intended estimate is a one-sided scalar inequality after taking the inner product with x(r)−y(r). Please rewrite this step to avoid confusion.
  2. [§5, Lemma 5.3 and Corollary 5.5] The name “Azelá-Ascoli” should read “Arzelà–Ascoli”.
  3. [§5, Theorem 2.3] The notation x_{μ̄} in the formula v̄−v̄(x̄,·)=h_{μ̄(·)}(x_{μ̄},·)−h_{μ̄(·)}(x_{μ̄},x̄) is not defined before use; since (H4) gives a unique x_m for each measure m, please define x_{μ̄} explicitly in the statement or in the paragraph preceding the theorem.
  4. [§3, Proposition 3.3] In the proof, v_+ and v_− are constructed using the same constant ∥F(·,m(t))−F(·,m(s))∥∞; the comparison step is correct but would benefit from a short explanation of why the sign of ∂H/∂u is used exactly as τ bounds the shift.
  5. [Appendix A, Proposition A.2] The constant D_{t_0} is defined with a lower bound −D_1 diam(M) and an upper bound C_{t_0}+D_1 diam(M); please check that the minimizer γ_2 exists and that the notation h_t^m(x,y) is consistent with the action being minimized over curves with γ(0)=x, γ(t)=y.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the selection limit is obtained from convergence estimates and external weak-KAM uniqueness; the only self-citations are technical regularity lemmas that are not load-bearing for the central claim.

full rationale

The central result (Theorem 2.3) is conditional and derived, not assumed. Under (H4), the critical equation H(x,0,Dv)=F(x,m)+c(m) has, by the external Aubry-Mather theory [34], a unique viscosity solution up to additive constants; the paper proves that any subsequential limit of u^lambda - u^lambda(xbar,.) is such a solution ([36, Theorem 1.1]) and then identifies it with the Peierls barrier h_m(x_m,.)-h_m(x_m,xbar) because both are normalized at xbar (or at x_m). This is a use of uniqueness, not an equivalence with the assumptions by construction. H4 is an explicit, strong assumption (singleton projected Aubry set for every measure path), not a parameter fitted to data; it is a limitation on scope, and the paper's Appendix B itself explains the non-uniqueness that can occur when H4 fails. The self-citations [23,24] supply uniform boundedness, equi-Lipschitz regularity and semi-concavity estimates from a published peer-reviewed paper; these are independent support and do not smuggle in the selection conclusion. The qMFG0 system is indeed the lambda-to-0 limit by construction, but the theorem's content is proving convergence and the Peierls-barrier representation, which is not tautological. A caveat, not a circularity: Theorem 2.3 only asserts convergence up to subsequences for m^lambda, so uniqueness of the selected mu-bar is not established.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The paper introduces no fitted free parameters and no new physical or mathematical entities beyond standard weak KAM objects such as Aubry sets and Peierls barriers, which are taken from the literature. The central claims rest on standard weak KAM and viscosity solution theory, the author's earlier stationary contact MFG estimates, and the paper-specific singleton Aubry set assumption (H4).

assumptions (5)
  • standard math Aubry-Mather theory for contact Hamiltonian systems as developed in [34], including the unique viscosity solution of H(x,u,Du)=c under (H3) and the Mane critical value formula for H(x,0,p)-F(x,m).
    Invoked throughout Section 3 and the appendix to define the solution u_m and the critical value c(m).
  • standard math Ambrosio-Crippa theory of continuity equations with non-smooth (BV) velocity fields, used to define the flow Phi_m and the pushforward solution of the continuity equation.
    Used in Section 4, around equation (4.1) and Proposition 4.4, to justify the flow and uniqueness of weak solutions.
  • standard math Uniform estimates for the stationary contact mean field game from [24], including uniform boundedness of the constants a_m and Lipschitz and semiconcavity of u_m.
    Proposition 3.2 cites [24, Lemma 1, Lemma 5, Proposition 9] for these estimates, which underpin the fixed point argument in Section 4.
  • standard math The convergence and a priori estimates for the fixed-Hamiltonian contact Hamilton-Jacobi equation from [36], in particular [36, Lemma 2.3] and the Peierls barrier characterization.
    Used in Proposition 5.4 and Corollary 5.5 to obtain uniform bounds and the explicit form of the limit selected by the vanishing discount.
  • domain assumption Assumption (H4): for every m in C([0,T];P(M)), the projected Aubry set A_m(t) of H(x,0,p)-F(x,m(t)) is a singleton {x_m} for all t in [0,T].
    This is the key structural hypothesis for Theorem 2.3, introduced just before Remark 3. It guarantees uniqueness of the critical viscosity solution up to constants and drives the explicit Peierls barrier formula.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Existence of Solutions and Selection Problem for Quasi-stationary Contact Mean Field Games." pith.science (2026). https://pith.science/paper/OPUG4KD6

@misc{pith2026250710112,
  author       = {Pith},
  title        = {Pith review of: Existence of Solutions and Selection Problem for Quasi-stationary Contact Mean Field Games},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OPUG4KD6}},
  note         = {Machine review of arXiv:2507.10112}
}
abstract

First, we study the existence of solutions for a class of first order mean field games systems \begin{equation*} \left\{\begin{aligned} &H(x,u,Du)=F(x,m(t)),\quad &&x\in M,\ \forall\ t\in[0,T],\\ &\partial_t m-\text{div}\left(m\dfrac{\partial H}{\partial p}(x,u,Du)\right)=0,\quad &&(x,t)\in M\times(0,T],\\ &m(0)=m_0, \end{aligned}\right. \end{equation*} where the system comprises a stationary Hamilton-Jacobi equation in the contact case and an evolutionary continuity equation. Then, for any fixed $\lambda>0$, let $(u^\lambda,m^\lambda)$ be a solution of the system \begin{equation*} \left\{ \begin{aligned} &H(x,\lambda u^\lambda,Du^\lambda)=F(x,m^\lambda(t))+c(m^\lambda(t)),\quad &&x\in M,\ \forall t\in[0,T],\\ &\partial_t m^\lambda-\text{div}\left(m^\lambda\dfrac{\partial H}{\partial p}(x,\lambda u^\lambda,Du^\lambda)\right)=0,\quad &&(x,t)\in M\times(0,T],\\ &m(0)=m_0, \end{aligned}\right. \end{equation*} where $c(m^\lambda(t))$ is the Ma\~n\'e critical value of the Hamiltonian $H(x,0,p)-F(x,m^\lambda(t))$. We investigate the selection problem for the limit of $(u^\lambda,m^\lambda)$ as $\lambda$ tends to 0.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references · 36 canonical work pages

  1. [11]

    A note on first order quasi-stationary Mean Field Games

    F. Camilli, C. Marchi, and C. Mendico, A note on first order quasi-stationary Mean Field Games, preprint, arXiv:2409.18483, 2024

  2. [24]

    Hu and K

    X. Hu and K. Wang, Existence of solutions to contact mean-field games of first order, Adv. Nonlinear Stud. 22 (2022), 289–307

  3. [1]

    Ambrosio and G

    L. Ambrosio and G. Crippa, Continuity equations and ODE flows with non-smooth velocity, Proc. Roy. Soc. Edinburgh Sect. A144 (2014), 1191–1244

  4. [2]

    Cardaliaguet, Long time average of first order mean field games and weak KAM theory, Dyn

    P. Cardaliaguet, Long time average of first order mean field games and weak KAM theory, Dyn. Games Appl. 3 (2013), 473–488

  5. [3]

    Cardaliaguet, Notes on mean field games, available at http://www.ceremade

    P. Cardaliaguet, Notes on mean field games, available at http://www.ceremade. dauphine.fr/~cardaliaguet/MFG20130420.pdf, 2013

  6. [4]

    Chen, Convergence of solutions of Hamilton–Jacobi equations depending nonlinearly on the unknown function, Adv

    Q. Chen, Convergence of solutions of Hamilton–Jacobi equations depending nonlinearly on the unknown function, Adv. Calc. Var.16 (2023), 45–68

  7. [5]

    Q. Chen, W. Cheng, H. Ishii, and K. Zhao, Vanishing contact structure problem and convergence of the viscosity solutions, Comm. Partial Differential Equations 44 (2019), 801–836

  8. [6]

    Cannarsa, W

    P. Cannarsa, W. Cheng, C. Mendico, and K. Wang, Long-time behavior of first-order mean field games on Euclidean space, Dyn. Games Appl. 10 (2020), 361–390

Show all 38 references
  1. [7]

    Cannarsa, W

    P. Cannarsa, W. Cheng, C. Mendico, and K. Wang, Weak KAM aspects of Hamilton-Jacobi equations with state constraints and applications to long-time behavior of constrained first-order mean field games, J. Dynam. Differ. Equ.(2021)

  2. [8]

    Q. Chen, A. Fathi, M. Zavidovique, and J. Zhang, Convergence of the solutions of the nonlinear discounted Hamilton–Jacobi equation: The central role of Mather measures, J. Math. Pures Appl. 181 (2024), 22–57

  3. [9]

    Contreras, R

    G. Contreras, R. Iturriaga, G. P. Paternain, and M. Paternain, Lagrangian graphs, minimizing measures and Mañé’s critical values, Geom. Funct. Anal. 8 (1998), 788–809

  4. [10]

    Camilli and C

    F. Camilli and C. Marchi, On quasi-stationary mean field games of controls,Appl. Math. Optim. 87 (2023), Paper No. 47, 31 pp

  5. [12]

    Cardaliaguet and A

    P. Cardaliaguet and A. Porretta, Long time behavior of the master equation in mean field game theory, Anal. PDE 12 (2019), 1397–1453. Quasi-stationary MFG 21

  6. [13]

    Cirant and A

    M. Cirant and A. Porretta, Long time behavior and turnpike solutions in mildly non- monotone mean field games, ESAIM Control Optim. Calc. Var.27 (2021), 86

  7. [14]

    Cannarsa and C

    P. Cannarsa and C. Sinestrari,Semiconcave Functions, Hamilton-Jacobi Equations, and Optimal Control, V ol. 58, Springer, 2004

  8. [15]

    Chen and M

    Q. Chen and M. Zhou, Perturbation estimates of weak KAM solutions and minimal invari- ant sets for nearly integrable Hamiltonian systems, Proc. Amer. Math. Soc. 145 (2017), 201–214

  9. [16]

    Davini, A

    A. Davini, A. Fathi, R. Iturriaga, and M. Zavidovique, Convergence of the solutions of the discounted Hamilton–Jacobi equation, Invent. Math. 206 (2016), 29–55

  10. [17]

    Davini, P

    A. Davini, P. Ni, J. Yan, and M. Zavidovique, Convergence/divergence phenomena in the vanishing discount limit of Hamilton-Jacobi equations, preprint, arXiv:2411.13780, 2024

  11. [18]

    Davini and L

    A. Davini and L. Wang, On the vanishing discount problem from the negative direction, Discrete Contin. Dyn. Syst. 41 (2021), 2377–2389

  12. [19]

    Fathi, Weak KAM Theorem and Lagrangian Dynamics, available at http://www.math.u-bordeaux.fr/pthieull/Recherche/KamFaible/ Publications/Fathi2008-01.pdf, 2008

    A. Fathi, Weak KAM Theorem and Lagrangian Dynamics, available at http://www.math.u-bordeaux.fr/pthieull/Recherche/KamFaible/ Publications/Fathi2008-01.pdf, 2008

  13. [20]

    Gomes, Generalized Mather problem and selection principles for viscosity solutions and Mather measures, Adv

    D. Gomes, Generalized Mather problem and selection principles for viscosity solutions and Mather measures, Adv. Calc. Var., 1 (2008), 291-307

  14. [21]

    Gomes, H

    D. Gomes, H. Mitake, and H. Tran, The selection problem for discounted Hamilton-Jacobi equations: some non-convex cases, J. Math. Soc. Japan 70 (2018), 345–364

  15. [22]

    Gomes, H

    D. Gomes, H. Mitake, and K. Terai, The selection problem for some first-order stationary mean-field games, Netw. Heterog. Media15 (2020), 681–710

  16. [23]

    Hu, The asymptotic problem on contact Hamilton–Jacobi equations with state constraints, Commun

    X. Hu, The asymptotic problem on contact Hamilton–Jacobi equations with state constraints, Commun. Nonlinear Sci. Numer. Simul. 143 (2025), Paper No. 108593, 19 pp

  17. [25]

    Iturriaga and K

    R. Iturriaga and K. Wang, A discrete weak KAM method for first-order stationary mean field games, SIAM J. Appl. Dyn. Syst. 22 (2023), 1253–1274

  18. [26]

    Ishii, K

    H. Ishii, K. Wang, L. Wang and J. Yan, Hamilton–Jacobi equations with their Hamiltonians depending Lipschitz continuously on the unknown, Comm. Partial Differential Equations 47 (2021), 417–452

  19. [27]

    Lions, Generalized solutions of Hamilton–Jacobi equations, Pitman, Boston, 1982

    P.-L. Lions, Generalized solutions of Hamilton–Jacobi equations, Pitman, Boston, 1982

  20. [28]

    Lasry and P.-L

    J.-M. Lasry and P.-L. Lions, Mean field games, Jpn. J. Math. 2 (2007), 229–260. 22 Xiaotian Hu

  21. [29]

    Mañé, Lagrangian flows: the dynamics of globally minimizing orbits, Bol

    R. Mañé, Lagrangian flows: the dynamics of globally minimizing orbits, Bol. Soc. Bras. Mat. 28 (1997), 141–153

  22. [30]

    Mouzouni, On quasi-stationary mean field games models, Appl

    C. Mouzouni, On quasi-stationary mean field games models, Appl. Math. Optim. 81 (2020), 655–684

  23. [31]

    Mitake and K

    H. Mitake and K. Terai, On weak solutions to first-order discount mean field games, Minimax Theory Appl. 8 (2023), 139–170

  24. [32]

    X. Su, L. Wang and J. Yan, Weak KAM theory for Hamilton-Jacobi equations depending on unknown functions, Discrete Contin. Dyn. Syst. 36 (2016), 6487-6522

  25. [33]

    Wang and J

    K. Wang and J. Yan, Ergodic problems for contact Hamilton-Jacobi equations, preprint, arXiv:2107.11554, 2021

  26. [34]

    K. Wang, L. Wang, and J. Yan, Aubry-Mather theory for contact Hamiltonian systems, Commun. Math. Phys. 366 (2019), 981–1023

  27. [35]

    K. Wang, L. Wang, and J. Yan, Variational principle for contact Hamiltonian systems and its applications, J. Math. Pures Appl. 123 (2019), 167–200

  28. [36]

    Y . Wang, J. Yan, and J. Zhang, Convergence of viscosity solutions of generalized contact Hamilton-Jacobi equations, Arch. Ration. Mech. Anal. 241 (2021), 885–902

  29. [37]

    Zavidovique, Convergence of solutions for some degenerate discounted Hamil- ton–Jacobi equations, Anal

    M. Zavidovique, Convergence of solutions for some degenerate discounted Hamil- ton–Jacobi equations, Anal. PDE 15 (2022), 1287–1311. Appendix A Existence result under (H3’) In the Appendix, we show that the existence result for (qM F Gλ) under (H1), (H2) and the weaker monoton...

  30. [38]

    1 shows the variation of singular points with ε

    The following Fig. 1 shows the variation of singular points with ε. Fig. 1 We have ∥Duε1 − Duε2∥∞ ≥ max 1 2 <x≤min{xε1 ,xε2 } p −2Vε1(x) + p −2Vε2(x) > 2. An analogous behaviour can be observed in the contact case, where the Hamiltonians take the form u + Hϵ(x, p). 28 Xiaotian...

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.