Abstract
We prove a rate of convergence for smooth solutions of the Monge-Ampère equation of a stable, monotone and consistent discretization. We consider the Monge-Ampère equation with a small low order perturbation. With such a perturbation, we can prove uniqueness of a solution to the discrete problem and stability of the discrete solution. The discretization considered is then known to converge to the viscosity solution but no rate of convergence was known.
Keywords:
rate of convergence; Monge-Ampère; monotone scheme; smooth solution; Primary: 65N12, Secondary: 65M06 MSC:
Primary: 65N12, Secondary: 65M06
1. Introduction
We obtain a rate of convergence, in the case of smooth convex solutions, of the finite difference schemes introduced in [,] for the elliptic Monge-Ampère equation
Here, for a smooth function u, is the Hessian of u, a symmetric matrix field. We assume that Ω is a bounded convex domain of , can be extended to a function which is convex in Ω and . We consider in this paper a finite difference scheme which is stable, monotone and consistent for the perturbed Monge-Ampère equation
Here is a small parameter and, for the discretizations we consider, both the term and its discretization are often omitted by an abuse of notation. The scheme we consider was introduced in [] but the approach we take also applies to the one introduced in []. The consistency error is O() where h is the spatial resolution and the directional resolution. Our error estimates are in terms of O(). The stability of the schemes for smooth solutions is a direct consequence of our error estimates. See also Remark 3.3. for a proof of stability in the general case, which seems to indicate that the perturbation is needed for the stability of the scheme we consider.
Rate of convergence for smooth solutions were previously established in the context of finite elements [,,] or the standard finite difference method [,]. The rate of convergence proven in this paper for a stable, monotone and consistent scheme is a key component of the theory developed in [] for the convergence of finite difference discretizations to the Aleksandrov solution of the Monge-Ampère equation. It follows from the approach taken therein and the results of this paper, that the discretizations proposed in [,] have approximations which converge to the weak solution, as defined in [], of an approximate problem to Equation (2), even in the general case where Equation (2) does not have a smooth solution. For convergence results in the classical sense, the rate of convergence given here is expected to help establish a rate of convergence for the scheme without a dependence on the smoothness of the solution. Such a result would eliminate the need for convergence to an approximate problem, instead of Equation (2), for the theory developed in []. See also [] for a different approach.
2. Notations and Preliminaries
We make the usual convention of using the letter C for various constants independent of the discretization. We make the assumption that . Let denote the mesh size. We assume without loss of generality that . Put
For , we denote the maximum norm of x by . We will use the notation for the Euclidean norm.
Let denote the set of real valued functions defined on , i.e. the set of mesh functions. For a subset of , and we define
Let v be a continuous function on Ω and let denote the unique element of defined by
We extend the operator canonically to vector fields and matrix fields. For a function g defined on , defines the analogous restriction on .
We are first interested in discrete versions of Equation (1)
where denotes a finite difference discretization of Equation (1).
2.1. Consistency
We recall that the consistency error of the scheme is defined as and that the scheme is consistent if for as . We will assume that the discretization is consistent. For the scheme we consider, the consistency error is given in terms of the usual spatial resolution h and the directional resolution.
To introduce the directional resolution, we first note that the determinant of the Hessian of a smooth function is essentially a second order directional derivative. More precisely, if we let W denote the set of orthogonal bases of , we have []
where denotes the transpose of .
The local directional resolution at x is defined as
We note that as , . The directional resolution is then defined as
The directional resolution can also be defined in terms of the angles between vectors. Here we have followed the approach used in [].
Let such that
We denote by the set of orthogonal bases of such that if and only if . Let be a given a mesh function and let , such that but for some i. We denote by the closest to x point of intersection with of the line through x and . We then define to be the value obtained by quadratic interpolation of , and . Similarly, if but for some i, we define by quadratic interpolation of , and , where now is the closest to x point of intersection with of the line through x and .
We consider the discrete Monge-Ampère operator defined by
The operator is shown in [] to be consistent with . We give here a detailed proof.
By a Taylor series expansion
Thus, using for such that , we have
Moreover, for , and with , a direction vector closest to v, i.e., , we have
By definition of α, . Let us assume that (so that second derivatives of u are locally bounded). Since , for a constant we have . Thus by Equation (5)
Let and let such that
Thus
It follows that
Next, if is a basis of eigenvectors of , we know from [] that . We claim that for each , we can find such that for all i. This follows from the observation that for , is obtained from by an orthogonal transformation. We recall that orthogonal transformations preserve inner products and that for such that , the image of μ by a rotation of angle in a plane spanned by two axis vectors is a vector for which . Thus having found such that , the other vectors are obtained by orthogonal transformations.
It then follows from Equation (6) that
We conclude that
3. Rate of Convergence
The proof of the rate of convergence is an application of the combined fixed point iterative method used in [].
To ensure convergence to a convex solution of Equation (1), it is natural to use a suitable notion of discrete convexity. We require that at an interior grid point x and for such that
with the usual assumption of the value of (resp. ) obtained by quadratic interpolation of , and (resp. , and ) when (resp. ). Here we have denoted by the closest to x point of intersection with of the line through x and (resp. the line through x and ).
As with [], the discrete convexity conditions can be combined with a discretization of the differential operator in a single equation. Recall that and define
Put
In practice the term is not used. It makes the discretization proper as defined below. And guarantees uniqueness of the discrete solution. For consistency, we are forced to consider the perturbed Equation (2).
The discrete Monge-Ampère equation is given by
Let denote the set of points and let denote the cardinality of the set . We note that the discretization takes the form
where for , is a real valued map defined on . For convenience we do not write explicitly the dependence of on x.
A scheme is proper if there is such that for and for all and , implies .
Thus our scheme is proper and the constant can be chosen independently of h.
The scheme is degenerate elliptic since if it is nondecreasing in each of the variables and .
A scheme is Lipschitz continuous if there is such that for all and
We claim that our scheme is Lipschitz continuous with Lipschitz constant . For , we can find such that
If we take . Otherwise we have
for an orthogonal basis . We then take , and . In both cases, .
Since the map is multilinear, is Lipschitz continuous with Lipschitz constant which we can take as for h sufficiently small.
Next we define the mapping
for . We have ([] Theorem 7)
Lemma 3.1. There exists a positive constant such that for all , we have
for where and are positive constants.
The proof of ([] Theorem 7) shows that under the assumption , the constant a takes the form . If necessary, by taking ν smaller we may assume that and . Thus the constant a takes the form
and since , we have .
We can now state the main result of this paper.
Theorem 3.2. For a solution of Equation (9) and for we have
for a constant C which is a scalar multiple of the maximum of the derivatives of u up to order 4 on and d.
We recall that is it is proven in [], see [] for details, that Equation (9) has a solution which is a fixed point of the mapping S. We have
by the consistency of the scheme, the observation that is discrete convex and Equation (2). The constant C is a consistency error constant and depends on the maximum of derivatives of u up to order 4 on and d.
We therefore have
This completes the proof.
Remark 3.3. The approach in the proof of the previous theorem also gives stability of a solution of Equation (9) in the general case, when f is uniformly bounded. We have
Therefore
Acknowledgments
The authors would like to thank the referees for suggestions to improve the paper and Brittney Froese for many useful discussions. The author was partially supported by NSF grant DMS-1319640.
References
- Oberman, A.M. Wide stencil finite difference schemes for the elliptic Monge-Ampère equation and functions of the eigenvalues of the Hessian. Discret. Contin. Dyn. Syst. Ser. B 2008, 10, 221–238. [Google Scholar] [CrossRef]
- Froese, B.; Oberman, A. Convergent finite difference solvers for viscosity solutions of the elliptic Monge-Ampère equation in dimensions two and higher. SIAM J. Numer. Anal. 2011, 49, 1692–1714. [Google Scholar] [CrossRef]
- Böhmer, K. On finite element methods for fully nonlinear elliptic equations of second order. SIAM J. Numer. Anal. 2008, 46, 1212–1249. [Google Scholar] [CrossRef]
- Feng, X.; Neilan, M. Analysis of Galerkin methods for the fully nonlinear Monge-Ampère equation. J. Sci. Comput. 2011, 47, 303–327. [Google Scholar] [CrossRef]
- Brenner, S.C.; Gudi, T.; Neilan, M.; Sung, L.Y. C0 penalty methods for the fully nonlinear Monge-Ampère equation. Math. Comp. 2011, 80, 1979–1995. [Google Scholar] [CrossRef]
- Awanou, G. Iterative Methods for k-Hessian Equations. Available online: http://arxiv.org/abs/1406.5366 (accessed on 21 March 2013).
- Awanou, G. On Standard Finite Difference Discretizations of the Elliptic Monge-Ampère Equation. Available online: http://arxiv.org/abs/1311.2812 (accessed on 21 March 2014).
- Trudinger, N.S.; Wang, X.J. Hessian measures. I. Topol. Methods Nonlinear Anal. 1997, 10, 225–239. [Google Scholar]
- Awanou, G.; Awi, R. Convergence of stable and consistent schemes to the Aleksandrov solution of the Monge-Ampère equation. Acta Appli. 2016. [Google Scholar] [CrossRef]
- Buades, A.; Coll, B.; Lisani, J.L.; Sbert, C. Conditional image diffusion. Inverse Probl. Imaging 2007, 1, 593–608. [Google Scholar]
- Oberman, A.M. Convergent difference schemes for degenerate elliptic and parabolic equations: Hamilton-Jacobi equations and free boundary problems. SIAM J. Numer. Anal. 2006, 44, 879–895. [Google Scholar] [CrossRef]
© 2016 by the author; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons by Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).