Short-Time Existence for a General Backward–Forward Parabolic System Arising from Mean-Field Games


We study the local in time existence of a regular solution of a nonlinear parabolic backward–forward system arising from the theory of mean-field games (briefly MFG). The proof is based on a contraction argument in a suitable space that takes account of the peculiar structure of the system, which involves also a coupling at the final horizon. We apply the result to obtain existence to very general MFG models, including also congestion problems.

This is a preview of subscription content, log in to check access.


  1. 1.

    Achdou Y, Porretta A (2018) Mean field games with congestion. Ann Inst H Poincaré Anal Non Linéaire 35:443–480

    MathSciNet  Article  Google Scholar 

  2. 2.

    Ambrose DM (2018) Strong solutions for time-dependent mean field games with non-separable Hamiltonians. J Math Pures Appl 9(113):141–154

    MathSciNet  Article  Google Scholar 

  3. 3.

    Ambrose DM (2018) Existence theory for non-separable mean field games in Sobolev spaces. preprint arXiv:1807.02223

  4. 4.

    Bardi M, Cirant M (2019) Uniqueness of solutions in MFG with several populations and Neumann conditions. In: Cardaliaguet P, Porretta A, Salvarani F (eds) PDE models for multi-agent phenomena. Springer INdAM series, pp 1–20

  5. 5.

    Bardi M, Fischer M (2018) On non-uniqueness and uniqueness of solutions in finite-horizon mean field games. ESAIM Control Optim Calc Var.

    Article  Google Scholar 

  6. 6.

    Cardaliaguet P (2013) Notes on mean field games. Accessed 10 Apr 2019

  7. 7.

    Cirant M (2019) On the existence of oscillating solutions in non-monotone mean-field games. J Differ Equ.

    MathSciNet  Article  MATH  Google Scholar 

  8. 8.

    Cirant M, Tonon D (2018) Time-dependent focusingmean-field games: the sub-critical case. J Dyn Differ Equ 31(1):49–79

    Article  Google Scholar 

  9. 9.

    Evangelista D, Gomes D (2018) On the existence of solutions for stationary mean-field games with congestion. J Dyn Differ Equ. 30(4):1365–1388

    MathSciNet  Article  Google Scholar 

  10. 10.

    Gianni R (1995) Global existence of a classical solution for a large class of free boundary problems. NoDEA Nonlinear Differ Equ Appl 2(3):291–321

    MathSciNet  Article  Google Scholar 

  11. 11.

    Gomes DA, Voskanyan VK (2015) Short-time existence of solutions for mean filed games with congestion. J Lond Math Soc 92(3):778–799

    MathSciNet  Article  Google Scholar 

  12. 12.

    Gomes D, Nurbekyan L, Prazeres M (2016) Explicit solutions of one-dimensional, first-order, stationary mean-field games with congestion. In: 2016 IEEE 55th conference on decision and control, CDC 2016, pp 4534–4539

  13. 13.

    Gomes D, Mitake H (2015) Existence for stationary mean-field games with congestion and quadratic Hamiltonians. NoDEA Nonlinear Differ Equ Appl 22(6):1897–1910

    MathSciNet  Article  Google Scholar 

  14. 14.

    Gomes D, Pimentel E, Sanchez-Morgado H (2016) Time dependent mean-field games in the superquadratic case. ESAIM Control Optim Calc Var 22(2):562–580

    MathSciNet  Article  Google Scholar 

  15. 15.

    Gomes D, Pimentel E, Sanchez-Morgado H (2015) Time-dependent mean-field games in the sub-quadratic case. Commun Partial Differ Equ 40(1):40–76

    Article  Google Scholar 

  16. 16.

    Gomes D, Pimentel E, Voskanyan V (2016) Regularity theory for mean-field game systems. Springer, New York

    Google Scholar 

  17. 17.

    Graber PJ (2015) Weak solutions for mean field games with congestion. preprint arXiv: 1503.04733

  18. 18.

    Graber PJ, Bensoussan A (2018) Existence and uniqueness of solutions for Bertrand and Cournot mean field games. Appl Math Optim 77(1):47–71

    MathSciNet  Article  Google Scholar 

  19. 19.

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

    MathSciNet  MATH  Google Scholar 

  20. 20.

    Ladyzhenskaya OA, Solonnikov VA, Uraltseva NN (1968) Linear and quasilinear equations of parabolic type. Translations of Math Mon. Providence, p 23

  21. 21.

    Lions P-L (2007–2011) College de france course on mean-field games.

  22. 22.

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

    MathSciNet  Article  Google Scholar 

  23. 23.

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

    MathSciNet  Article  Google Scholar 

  24. 24.

    Lasry JM, Lions PL (2007) Mean field games. Jpn J Math (N.S.) 2:229–260

    MathSciNet  Article  Google Scholar 

Download references


Marco Cirant and Paola Mannucci are members of GNAMPA-INdAM, and were partially supported by the research project of the University of Padova “Mean-Field Games and Nonlinear PDEs” and by the Fondazione CaRiPaRo Project “Nonlinear Partial Differential Equations: Asymptotic Problems and Mean-Field Games”.

Author information



Corresponding author

Correspondence to Marco Cirant.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.



In this final appendix, we prove Proposition 2.6 of Sect. 2.

Proof of Proposition 2.6

We write the proof for the existence of a solution in the class \(C^{2+\alpha , 1+\alpha /2}(Q_T)\) and for the estimate (2.8). In a similar way one obtains the existence in \(W^{2,1}_q(Q_T)\) and the proof of (2.9).

Recall that the problem on \(\mathbb {T}^N\times [0,T]\) is equivalent to the same problem with 1-periodic data in the x-variable in \({\mathbb {R}}^N\times [0,T]\), namely with all the data satisfying \(w(x+z,t)=w(x,t)\) for all \(z\in {{I\!\!Z}}^N\). As far as the existence of a smooth solution of problem (2.7) is concerned, it is sufficient to apply Theorem 5.1 p.320 of [20]. Since the solution of such a Cauchy problem is unique, it must be periodic in the x-variable. Now we prove estimate (2.8). Let \(R_1^N:=[-1,1]^N\) and \(R_2^N := [-2,2]^N\). Clearly

$$\begin{aligned}{}[0,1]^N \subset R_1^N\subset R_2^N \subset {\mathbb {R}}^N \end{aligned}$$

and \(dist(R_1^N, C(R_2^N))=1\).

We take advantage of local parabolic estimates, which allow us to get an a priori estimate regardless of the lateral boundary conditions which are unknown for us.

In particular, using the local estimate (10.5) p. 352 of [20] with \(\Omega '=R_1^N\) and \(\Omega '':={R_2^N}\), (note that in our case \(S''\) is empty) we have

$$\begin{aligned} |u|^{(\alpha +2)}_{R_1^N\times [0,T^*]}\le C_1\left( |f|^{(\alpha )}_{R_2^N\times [0,T^*]}+ |u_0|^{(\alpha +2)}_{R_2^N}\right) + C_2 |u|_{R_2^N\times [0,T^*]}, \end{aligned}$$

where \(T^*<T\), \(C_1\) and \(C_2\) depend on N, T, and the modulus of Hölder continuity of the coefficients of the operator. It is now crucial to observe that Hölder norms on \(R_1^N\times [0,T^*]\), \(R_2^N\times [0,T^*]\) and \(\mathbb T^N\times [0,T^*]\) coincide by periodicity of u, f, \(u_0\) in the x-variable and the inclusions (A.1). Hence,

$$\begin{aligned} |u|^{(\alpha +2)}_{\mathbb T^N\times [0,T^*]}\le C_1\left( |f|^{(\alpha )}_{\mathbb T^N\times [0,T^*]}+ |u_0|^{(\alpha +2)}_{\mathbb T^N}\right) + C_2(|u_0|_{\mathbb T^N}+ T^*|u_t|_{\mathbb T^N\times [0,T^*]}). \end{aligned}$$

Taking \(T^*\) sufficiently small we can write

$$\begin{aligned} |u|^{(\alpha +2)}_{\mathbb T^N\times [0,T^*]}\le C_3\left( |f|^{(\alpha )}_{\mathbb T^N\times [0,T^*]}+ |u_0|^{(\alpha +2)}_{\mathbb T^N}\right) , \end{aligned}$$

where \(C_3\) depends on the coefficients of the equation, on N, T, \(\alpha \) and \(T^*\) does not depend on \(u_0\). We can iterate the estimate (A.4) to cover all the interval [0, T] in \([\frac{T}{T^*}]+1\) steps, thus obtaining (2.8).

The proof of (2.9) is completely analogous. One has to exploit the local estimate in \(W^{2,1}_p\) of [20], Eq. (10.12), p. 355. Note that since \(R_1^N\) and \(R_2^N\) consist of finite copies of \([0,1]^N\), norms on \(W_p^{2m, m}(R_i^N \times (0,T))\), \(i=1,2\), are multiples (depending on N) of \(W_p^{2m, m}(\mathbb T^N \times (0, T))\). \(\square \)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Cirant, M., Gianni, R. & Mannucci, P. Short-Time Existence for a General Backward–Forward Parabolic System Arising from Mean-Field Games. Dyn Games Appl 10, 100–119 (2020).

Download citation


  • Parabolic equations
  • Backward–forward system
  • Mean-field games
  • Hamilton–Jacobi
  • Fokker–Planck
  • Congestion problems

Mathematics Subject Classification

  • 35K40
  • 35K61
  • 49N90