Abstract
Nonlinear inverse problems are much more difficult to solve than linear ones and the corresponding theory is far less developed. Each particular problem may demand a specific regularization. Nevertheless, one can formulate and analyze basic versions of nonlinear Tikhonov regularization and nonlinear iterative methods, which will be done in Sects. 4.1 and 4.6, respectively. These basic versions serve as starting points for regularizations adapted to our nonlinear model problems (Sects. 4.2, 4.5, and 4.6). Nonlinear Tikhonov regularization leads to nonlinear least squares problems. Sections 4.3 and 4.4 present algorithms to solve these numerically.
Notes
- 1.
If there is no \(\hat{x} \in D\) with \(F(\hat{x}) = y\) due to discretization errors, then hopefully there will at least exist some \(\bar{y} \approx y\) such that a solution \(\hat{x} \in D\) of \(F(x) =\bar{ y}\) exists. If we know δ > 0 such that \(\|\bar{y} - y^{\delta }\|_{2} \leq \delta\), we can tacitly add discretization to measurement errors, interpreting \(\bar{y}\) as “true” right hand side.
- 2.
To arrange multi-indexed components into a vector, a certain index ordering has to be fixed. In case of rectangular index grids, we will always use a row-wise ordering. For example, the indices α ∈ G n for the grid G n defined below, will always be ordered to form the sequence
$$\displaystyle{(-n,-n),\ldots, (-n,n),\quad \ldots,\quad (n,-n),\ldots, (n,n).}$$ - 3.
This is generally true even if rank(J) = n does not hold, provided only that ∇Z(x i) ≠ 0, see [Bjö96], p. 343.
References
A. Bamberger, G. Chavent, and P. Lailly. Etude mathématique et numérique d’un problème inverse pour l’équation des ondes à une dimension. Rapport LABORIA nr. 226, IRIA, 1977.
M. A. Branch, T. F. Coleman, and Y. Li. A Subspace, Interior, and Conjugate Gradient Method for Large-Scale Bound-Constrained Minimization Problems. SIAM J. Sci. Comput., 21(1):1–23, 1999.
Å. Björck. Numerical Methods for Least Squares Problems. SIAM, 1996.
G. Chavent. Identification of Functional Parameters in Partial Differential Equations. In R. E. Goodson and M. Polis, editors, Identification of Parameters in Distributed Systems, pages 31–48. The American Society of Mechanical Engineering, 1974.
G. Chavent. Nonlinear Least Squares for Inverse Problems. Springer, 2009.
H. W. Engl, M. Hanke, and A. Neubauer. Regularization of Inverse Problems. Kluwer Academic Publishers, 1996.
H. W. Engl, K. Kunisch, and A. Neubauer. Convergence rates for Tikhonov regularisation of non-linear ill-posed problems. Inverse Problems, 5:523–540, 1989.
M. Hanke. A regularizing Levenberg-Marquardt scheme, with applications to inverse groundwater filtration problems. Inverse Problems, 13:79–95, 1997.
M. Hanke. Regularizing properties of a truncated Newton-CG algorithm for nonlinear inverse problems. Numer. Funct. Anal. and Optimiz., 18:971–993, 1997.
A. Kirsch. An Introduction to the Mathematical Theory of Inverse Problems. Springer, 1996.
P. Lailly. The inverse problem in 1-D reflection seismics. In R. Cassinis, editor, The Solution of the Inverse Problem in Geophysical Interpretation, pages 103–140. Plenum Press, 1980.
D. Luenberger and Y. Ye. Linear and Nonlinear Programming. Springer, 2008.
J. J. Moré. The Levenberg-Marquardt Algorithm: Implementation and Theory. In G. A. Watson, editor, Numerical Analysis. Proceedings Biennial Conference Dundee 1977, Lecture Notes in Mathematics, volume 630, pages 105–116. Springer, 1978.
R. Ramlau. Morozov’s discrepancy principle for Tikhonov regularization of nonlinear operators. Numer. Funct. Anal. Optimization, 23:147–172, 2002.
L. I. Rudin, S. Osher, and E. Fatemi. Nonlinear total variation based noise removal algorithms. Physica D, 60:259–268, 1992.
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 2016 Springer International Publishing AG
About this chapter
Cite this chapter
Richter, M. (2016). Regularization of Nonlinear Inverse Problems. In: Inverse Problems . Lecture Notes in Geosystems Mathematics and Computing. Birkhäuser, Cham. https://doi.org/10.1007/978-3-319-48384-9_4
Download citation
DOI: https://doi.org/10.1007/978-3-319-48384-9_4
Published:
Publisher Name: Birkhäuser, Cham
Print ISBN: 978-3-319-48383-2
Online ISBN: 978-3-319-48384-9
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)