The Tikhonov Regularization Method in Hilbert Scales for Determining the Unknown Source for the Modified Helmholtz Equation

Lei You, Zhi Li, Jianbang Huang, Du Ai-huaJournal of Applied Mathematics and Physics 2016 · DOI

Abstract

In this paper, we consider an unknown source problem for the modified Helmholtz equation. The Tikhonov regularization method in Hilbert scales is extended to deal with ill-posedness of the problem. An a priori strategy and an a posteriori choice rule have been present to obtain the regularization parameter and corresponding error estimates have been obtained. The smoothness parameter and the a priori bound of exact solution are not needed for the a posteriori choice rule. Numerical results are presented to show the stability and effectiveness of the method.

Introduction

The modified Helmholtz equation (MHE) is a fundamental model used in diverse scientific fields, including Debye-Huckel theory in chemistry, implicit heat equation schemes, and the linearization of the Poisson-Boltzmann equation in biophysics. A significant challenge in these applications is the inverse source problem: reconstructing an unknown source term (the "input") from noisy measurements of the field (the "output") at a specific boundary.

This problem is notoriously ill-posed. In such systems, tiny errors or "noise" in the measured data are amplified exponentially when attempting to calculate the source. Without a mathematical "stabilizer" or regularization, the reconstructed source would appear as meaningless static rather than a physical signal. This paper presents an improved Tikhonov regularization method using "Hilbert scales" to solve this instability, introducing a parameter choice rule that achieves high accuracy without requiring prior knowledge of the source's smoothness.

Methodology

The researchers employ Tikhonov Regularization in Hilbert Scales. While standard Tikhonov regularization balances data fitting with a simple penalty term to keep the solution "small," the Hilbert scales approach uses a norm that accounts for different levels of smoothness (derivatives). This allows for much more precise control over the reconstruction.

The paper evaluates two strategies for choosing the "regularization parameter" ($\alpha$), which determines how much to trust the data versus the penalty:

  1. A Priori: Based on the known noise level and the assumed smoothness of the source.
  2. A Posteriori (Discrepancy Principle): A more practical method that chooses $\alpha$ based on the residual error. Crucially, this method does not require knowing the smoothness of the exact solution beforehand, making it significantly more useful for real-world applications where the source is truly unknown.

To validate these methods, the authors conducted synthetic numerical simulations. They defined a known "exact" source, calculated the resulting field, added artificial random noise to simulate measurement error, and then used their algorithm to reconstruct the original source.

Results and Discussion

The numerical tests demonstrate that the method remains stable even as the discretization (the number of points $N$) increases. A key insight is the method's ability to overcome the "smoothness gap"; the authors prove that using the discrepancy principle in Hilbert scales allows the algorithm to "self-tune" to the correct smoothness level.

Table 1: Relative errors for various parameters $p$ and $N$ with $k=2$.

As shown in Table 1, the relative errors remain consistently low across various noise levels ($\delta_1$) and discretization levels ($N$). The authors proved that their method achieves an optimal convergence rate of $O(\delta^{p/(p+2)})$, meaning that as measurements get cleaner, the reconstruction improves at a mathematically predictable and efficient rate.

In comparative tests, the new method (M1) significantly outperformed previously established simplified Tikhonov methods (M2). For example, at a 10% noise level ($10^{-1}$), the relative error for M1 was roughly 0.07, while M2 was 0.11. At lower noise levels ($10^{-3}$), M1 achieved a relative error as low as 0.0072, whereas M2 remained at a much higher error of 0.0612.

Conclusion

This research provides a robust framework for solving inverse source problems in the modified Helmholtz equation. By utilizing Hilbert scales and an a posteriori parameter choice rule, the authors have developed a method that is both stable and highly accurate without requiring unrealistic prior knowledge of the source's characteristics. Future work may involve extending this method to three-dimensional engineering problems and complex, irregular geometries.

Summary updated

Abstract

In this paper, we consider an unknown source problem for the modified Helmholtz equation. The Tikhonov regularization method in Hilbert scales is extended to deal with ill-posedness of the problem. An a priori strategy and an a posteriori choice rule have been present to obtain the regularization parameter and corresponding error estimates have been obtained. The smoothness parameter and the a priori bound of exact solution are not needed for the a posteriori choice rule. Numerical results are presented to show the stability and effectiveness of the method.
BibTeX
@article{you2016tikhonov,
  author = {Lei You and Zhi Li and Jianbang Huang and DU Ai-hua},
  title = {{The Tikhonov Regularization Method in Hilbert Scales for Determining the Unknown Source for the Modified Helmholtz Equation}},
  journal = {Journal of Applied Mathematics and Physics},
  year = {2016},
  doi = {10.4236/jamp.2016.41017},
}
Figure