Newton s method for inverse obstacle scattering meets the method of least squares (Q4466429)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 2072187
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | Newton s method for inverse obstacle scattering meets the method of least squares |
scientific article; zbMATH DE number 2072187 |
Statements
Newton s method for inverse obstacle scattering meets the method of least squares (English)
0 references
10 June 2004
0 references
inverse scattering problems
0 references
nonlinear ill-posed problems
0 references
Helmholtz equation
0 references
Newton's method
0 references
regularization
0 references
integral equation
0 references
numerical example
0 references
inverse obstacle scattering
0 references
0.8442736268043518
0 references
0.8329119086265564
0 references
0.8276336789131165
0 references
Two approaches for the numerical solution of the inverse scattering problem are considered. In a first approach this problem is written as a nonlinear ill-posed operator equation \( F(\partial D) = u_\infty \). Here \( \partial D \) denotes the boundary of the unknown scatterer \( D \subset \mathbb{R}^3 \), and \( u_\infty \) denotes the far-field pattern of the scattered wave \( u^{\text{s}} \) generated by an incident wave \( u^{\text{i}} \). The author reviews the basic mapping properties of the operator \( F \) like nonlinearity, differentiability and ill-posedness. Additionally, results on regularized Newton methods and its numerical implementation are presented. In the second part of this paper, a new approach for the inverse scattering problem is considered. Here, this problem is written as a nonlinear equation \( G(\partial D) = 0 \), where the operator \( G \) is given by \( G: \Gamma \mapsto u_{| \Gamma} \), i.e., it maps a closed \( C^2 \) surface \( \Gamma \) onto the trace of the total field \( u = u^{\text{i}} + u^{\text{s}} \). It is shown that this mapping is well-defined for surfaces \( \Gamma \) in a neighborhood of \( \partial D \). Moreover it is shown that an evaluation of \( G(\Gamma) \) requires only the determination of a density function \( \varphi_\Gamma \) defined on \( \Gamma \) as well as the evaluation of potentials, the solution of the direct scattering problem is not needed. Here, the density function \( \varphi_\Gamma \) is obtained by solving an integral equation of the first kind considered on the surface \( \Gamma \). Differentiability of the operator \( G \) and the implementation of Newton's method for solving \( G(\partial D) = 0 \) are considered, and finally a numerical example is presented. The second approach is related to an existing method for the inverse obstacle scattering problem where the equation \( G(\partial D) = 0 \) is solved in a least squares sense. This connection is indicated by the author in the title of his paper.
0 references