Abstract:We consider a unique continuation problem where the Dirichlet trace of the solution is known to have finite dimension. We prove Lipschitz stability of the unique continuation problem and design a finite element method that exploits the finite dimensionality to enhance stability. Optimal a priori and a posteriori error estimates are shown for the method. The extension to problems where the trace is not in a finite dimensional space, but can be approximated to high accuracy using finite dimensional functions is discussed. Finally, the theory is illustrated in some numerical examples.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is the numerical stability and accuracy problems in the **Unique Continuation (UC)**, especially when the Dirichlet trace of the solution is known to be finite - dimensional. Specifically, the authors focus on how to use the finite element method (FEM) to achieve the numerical solution of this unique continuation problem and ensure its good stability and high accuracy.
### Specific description of the problem
1. **Background of the unique continuation problem**:
- The unique continuation problem refers to inferring the function values in a larger region \(\Omega\) from the given function values in a certain region \(\omega\). This problem has wide applications in fields such as data assimilation, inverse problems, and control.
- Although the unique continuation problem is well - posed in theory, it is very unstable in numerical calculations and usually has only conditional logarithmic stability.
2. **Assumption of finite - dimensional trace**:
- If we know that the Dirichlet trace \(u|_{\partial\Omega}\) of the solution belongs to a known finite - dimensional space \(V_N\), then the stability of the problem can be improved to Lipschitz stability, making the numerical solution more feasible.
3. **Objective**:
- The authors' objective is to design a finite element method that can use this finite - dimensional information to enhance the stability and accuracy of the numerical solution. Specifically, they hope to prove the a priori and a posteriori error estimates of this method and demonstrate its effectiveness in practical applications.
### Main contributions of the paper
- **Lipschitz stability estimate**: By introducing the assumption of finite - dimensional trace, the authors proved the Lipschitz stability of the unique continuation problem.
- **Design of the finite element method**: Based on the above - mentioned stability estimate, a new finite element method was designed, which can effectively use finite - dimensional information to improve the accuracy of the numerical solution.
- **Error estimate**: The optimal a priori and a posteriori error estimates of this method were proved, ensuring the convergence of the solution.
- **Numerical experiments**: Through a series of numerical experiments, the validity of the theoretical results was verified, and the advantages of this method in dealing with the unique continuation problem were demonstrated.
### Summary of mathematical formulas
- **Lipschitz stability estimate**:
\[
\|u\|_{H^1(\Omega)} \lesssim \|u\|_{L^2(\omega)} + \|Qu\|_{H^{1/2}(\partial \Omega)} + \|\Delta u\|_{H^{-1}(\Omega)}, \quad u \in H^1(\Omega)
\]
where \(P\) is the operator projected onto \(V_N\), and \(Q = I - P\).
- **Error estimate**:
\[
\|u - u_h\|_{H^1(\Omega)} \leq Ch^k \|D^{k + 1}u\|_{L^2(\Omega)}
\]
where \(h\) is the mesh parameter and \(k\) is the polynomial order of the finite element space.
Through these works, the authors successfully solved the numerical stability and accuracy problems in the unique continuation problem and provided an effective numerical method to deal with such problems.