Article Contents
Article Contents

# A convergent Lagrangian discretization for $p$-Wasserstein and flux-limited diffusion equations

• *Corresponding author

This research was supported by the German Research Foundation (DFG), Collaborative Research Center SFB-TR 109

• We study a Lagrangian numerical scheme for solving a nonlinear drift diffusion equations of the form $\partial_t u = \partial_x(u \cdot ({\sf c}^*)^\prime[\partial_x \mathit{h}^\prime(u)+ \mathit{v}^\prime])$, like Fokker-Plank and $q$-Laplace equations, on an interval. This scheme will consist of a spatio-temporal discretization founded on the formulation of the equation in terms of inverse distribution functions. It is based on the gradient flow structure of the equation with respect to optimal transport distances for a family of costs that are in some sense $p$-Wasserstein like. Additionally we will show that, under a regularity assumption on the initial data, this also includes a family of discontinuous, flux-limiting cost inducing equations like Rosenau's relativistic heat equation. We show that this discretization inherits various properties from the continuous flow, like entropy monotonicity, mass preservation, a minimum/maximum principle and flux-limitation in the case of the corresponding cost. Convergence in the limit of vanishing mesh size will be proven as the main result. Finally we will present numerical experiments including a numerical convergence analysis.

Mathematics Subject Classification: Primary: 35K30, 35Q99, 65M12; Secondary: 35B40.

 Citation:

• Figure 1.  Experiment: p-Wasserstein cost, linear diffusion. Left: Approximate densities $u(t, \cdot)$ at $t = 0.01\cdot 10^k$, $k = 0, 0.12, 0.24, \ldots, \log_{10}(200)$, initial mass uniformly distributed on $[-0.3, 0.3]$. Right: the corresponding characteristics

Figure 2.  Experiment: p-Wasserstein cost, linear diffusion. Left: Approximate densities $u(t, \cdot)$ at $t = 0.01\cdot 10^k$, $k = 0, 0.12, 0.24, \ldots, \log_{10}(200)$, initial mass uniformly distributed on $[-3, -2.4]$. Right: the corresponding characteristics

Figure 3.  Experiment: relativistic cost, linear diffusion. Left: Approximate densities $u(t, \cdot)$ for $t = 0.01\cdot 10^k$, $k = 0, 0.12, 0.24, \ldots, \log_{10}(200)$, initial mass uniformly distributed on $[-0.3, 0.3]$. Right: the corresponding characteristics (dashed: speed of light)

Figure 4.  Convergence analysis: relativistic cost, linear diffusion. $L^1$-error of the inverse distribution function in dependence of the grid size (left), and in dependence of the time step (right)

Figure 5.  Experiment: $q$-Laplace ($p = \frac43, m = \frac53$). Left: Approximate densities $u(t,\cdot)$ for $t = 0.01\cdot 10^k$, $k = 0,0.12,0.24,\ldots,\log_{10}(200)$, initial mass uniformly distributed on $[-0.3,0.3]$. Right: the corresponding characteristics(dashed: speed of light)

•  [1] M. Agueh et al., Existence of solutions to degenerate parabolic equations via the MongeKantorovich theory, Adv. Differ. Equ., 10 (2005), 309–360. [2] L. Ambrosio, N. Gigli and G. Savaré, Gradient Flows: in Metric Spaces and in the Space of Probability Measures, Springer Science & Business Media, 2008. [3] J. D. Benamou and Y. Brenier, A computational fluid mechanics solution to the Monge-Kantorovich mass transfer problem, Numer. Math., 84 (2000), 375-393.  doi: 10.1007/s002110050002. [4] C. Budd, G. Collins, W. Huang and R. Russell, Self–similar numerical solutions of the porous–medium equation using moving mesh methods, Philos. Trans. R. Soc. London Ser. A Math. Phys. Eng., 357 (1999), 1047-1077.  doi: 10.1098/rsta.1999.0364. [5] M. Burger, J. A. Carrillo and M. T. Wolfram et al., A mixed finite element method for nonlinear diffusion equations, Kinet. Relat. Models, 3 (2010). doi: 10.3934/krm.2010.3.59. [6] J. Carrillo, B. D$\ddot{u}$ring, D. Matthes and D. S. McCormick, A Lagrangian scheme for the solution of nonlinear diffusion equations using moving simplex meshes, J. Sci. Comput., 75 (2018), 1463-1499.  doi: 10.1007/s10915-017-0594-5. [7] J. A. Carrillo and J. S. Moll, Numerical simulation of diffusive and aggregation phenomena in nonlinear continuity equations by evolving diffeomorphisms, SIAM J. Sci. Comput., 31 (2009), 4305-4329.  doi: 10.1137/080739574. [8] F. Cavalli and G. Naldi, A wasserstein approach to the numerical solution of the one-dimensional cahn-hilliard equation, Kinet. Relat. Models, 3 (2010), 123-142.  doi: 10.3934/krm.2010.3.123. [9] V. De Cicco, N. Fusco and A. Verde, On l1-lower semicontinuity in bv, J. Convex Anal., 12 (2005), 173-185. [10] B. Düring, D. Matthes and J. P. Milišic, A gradient flow scheme for nonlinear fourth order equations, Discrete Contin. Dyn. Syst. Ser. B, 14 (2010), 935-959.  doi: 10.3934/dcdsb.2010.14.935. [11] L. C. Evans, Partial Differential Equations, Springer, 2013. [12] E. Giusti and G. H. Williams, Minimal Surfaces and Functions of Bounded Variation, Vol. 2, Springer, 1984. [13] L. Gosse and G. Toscani, Identification of asymptotic decay to self-similarity for one-dimensional filtration equations, SIAM J. Numer. Anal., 43 (2006), 2590-2606.  doi: 10.1137/040608672. [14] R. Jordan, D. Kinderlehrer and F. Otto, The variational formulation of the fokker–planck equation, SIAM J. Math. Anal., 29 (1998), 1-17.  doi: 10.1137/S0036141096303359. [15] D. Kinderlehrer and N. J. Walkington, Approximation of parabolic equations using the wasserstein metric, ESAIM Math. Model. Numer. Anal., 33 (1999), 837-852.  doi: 10.1051/m2an:1999166. [16] M. Leven, Gradientenfluß-basierte Diskretisierung Parabolischer Gleichungen, Ph.D thesis, Inst. für Angew. Math. der Univ., 2002. [17] R. MacCamy and E. Socolovsky, A numerical procedure for the porous media equation, Comput. Math. Appl., 11 (1985), 315-219.  doi: 10.1016/0898-1221(85)90156-7. [18] D. Matthes and H. Osberger, Convergence of a variational lagrangian scheme for a nonlinear drift diffusion equation, ESAIM Math. Model. Numer. Anal., 48 (2014), 697-726.  doi: 10.1051/m2an/2013126. [19] R. J. McCann and M. Puel, Constructing a relativistic heat flow by transport time steps, Ann. Inst. Henri Poincare (C) Non Linear Anal., 26 (2009), 2539-2580.  doi: 10.1016/j.anihpc.2009.06.006. [20] T. Roessler, Discretizing the Porous Medium Equation Based on Its Gradient Flow Structure: A Consistency Paradox, Universit$\ddot{a}$t Bonn, SFB 611, Singul$\ddot{a}$re Ph$\ddot{a}$nomene und Skalierung in, Technical Report, 2004. [21] P. Rosenau, Tempered diffusion: A transport process with propagating fronts and inertial delay, Phys. Rev. A, 46 (1992), R7371. [22] R. Rossi and G. Savaré, Tightness, integral equicontinuity and compactness for evolution problems in banach spaces, Ann. Scuola Norm. Super. Pisa-Cl. Sci. Ser. V, 2 (2003), 395. [23] G. Russo, Deterministic diffusion of particles, Commun. Pure Appl. Math., 43 (1990), 697-733.  doi: 10.1002/cpa.3160430602. [24] F. Santambrogio, Optimal transport for applied mathematicians, Birkäuser, NY, 99–102. doi: 10.1007/978-3-319-20828-2. [25] C. Villani, Topics in Optimal Transportation, Vol. 58, American Mathematical Soc., 2003. doi: 10.1090/gsm/058. [26] M. Westdickenberg and J. Wilkening, Variational particle schemes for the porous medium equation and for the system of isentropic euler equations, ESAIM Math. Model. Numer. Anal., 44 (2010), 133-166.  doi: 10.1051/m2an/2009043.

Figures(5)