Jalal M. Fadili,Nicolas Forcadel,Rita Zantout
Abstract:In this paper, we consider a non-local approximation of the time-dependent Eikonal equation defined on a Riemannian manifold. We show that the local and the non-local problems are well-posed in the sense of viscosity solutions and we prove regularity properties of these solutions in time and space. If the kernel is properly scaled, we then derive error bounds between the solution to the non-local problem and the one to the local problem, both in continuous-time and Forward Euler discretization. Finally, we apply these results to a sequence of random weighted graphs with n vertices. In particular, we establish that the solution to the problem on graphs converges almost surely uniformly to the viscosity solution of the local problem as the kernel scale parameter decreases at an appropriate rate when the number of vertices grows and the time step vanishes.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is the non - local approximation problem of the time - dependent Eikonal equation defined on Riemannian manifolds. Specifically, the author studies the continuous limit problem of the Eikonal equation defined on weighted geometric graphs when the number of vertices tends to infinity. In addition, the paper also explores the existence, uniqueness and regularity of these solutions, and derives the error bounds between local and non - local problems.
### Main research problems
1. **Existence of the continuous limit**:
- Given the scale parameter $\varepsilon_n$ depending on the number of vertices $n$, when $n \to \infty$, does the solution of the non - local Eikonal equation on the graph have a continuous limit?
- In what sense does this continuous limit exist?
2. **Convergence rate**:
- When $n \to \infty$ and the time step $\Delta t \to 0$, at what rate does the solution of the non - local problem converge to the solution of the local problem?
3. **Main quantities of the error bound**:
- What are the main quantities that affect the error bound? How do these quantities affect the size of the error bound?
### Research background
The Eikonal equation is an important class of partial differential equations and has wide applications in fields such as image processing, machine learning, and data analysis. In particular, when dealing with high - dimensional data, the Eikonal equation can be used to calculate the shortest path or distance. However, it is often very difficult to directly solve the Eikonal equation in high - dimensional space. Therefore, approximating the continuous problem through non - local approximation on graphs has become an effective means.
### Main contributions of the paper
1. **Existence and uniqueness**:
- It is proved that the local Eikonal equation (P) and the non - local Eikonal equation (Pε) are well - posed in the sense of viscosity solutions, that is, the existence and uniqueness of solutions.
2. **Regularity**:
- The regularity properties of the solution in time and space are studied.
3. **Error bound**:
- The error bound between the local problem (P) and the non - local problem (Pε) is derived.
4. **Discretization consistency**:
- The forward Euler method is used to discretize (Pε) in time, and the error bound between the discretized problem (PFDε) and the continuous problem (P) is provided.
5. **Application to a sequence of random geometric graphs**:
- The above results are applied to a sequence of random geometric graphs with $n$ vertices, and it is proved that when the kernel scale parameter decreases at an appropriate rate, the solution of the discretized problem converges almost uniformly to the viscosity solution of the local problem.
### Mathematical model
The non - local Eikonal equation considered in the paper has the following form:
\[
\begin{cases}
\frac{\partial f^\varepsilon(x,t)}{\partial t}+\left\|\nabla_{\eta_\varepsilon} f^\varepsilon(x,t)\right\|_\infty = P(x), & (x,t)\in(\tilde{M}\setminus\Gamma)\times[0, T], \\
f^\varepsilon(x,t)=f_0^\varepsilon(x), & (x,t)\in(\partial\Gamma\times[0, T])\cup\tilde{M}\times\{0\},
\end{cases}
\]
where:
- $\tilde{M}$ is a subset of the manifold $M$,
- $\Gamma$ is the set of boundary points,
- $P$ is the potential function,
- $f_0^\varepsilon$ is the initial condition,
- $\nabla_{\eta_\varepsilon}$ is a non - local operator, defined as:
\[
\left\|\nabla_{\eta_\varepsilon} f^\varepsil