Regularization of Inverse Problems

In the last two decades, the field of inverse problems has certainly been one of the fastest growing areas in applied mathematics. This growth has largely been driven by the needs of applications both in other sciences and in industry. In Chapter 1, we will give a short overview over some classes of inverse problems of practical interest. Like everything in this book, this overview is far from being complete and quite subjective. As will be shown, inverse problems typically lead to mathematical models that are not well-posed in the sense of Hadamard, i.e., to ill-posed problems. This means especially that their solution is unstable under data perturbations. Numerical meth ods that can cope with this problem are the so-called regularization methods. This book is devoted to the mathematical theory of regularization methods. For linear problems, this theory can be considered to be relatively complete and will be de scribed in Chapters 2 - 8. For nonlinear problems, the theory is so far developed to a much lesser extent. We give an account of some of the currently available results, as far as they might be of lasting value, in Chapters 10 and 11. Although the main emphasis of the book is on a functional analytic treatment in the context of operator equations, we include, for linear problems, also some information on numerical aspects in Chapter 9.

This book is devoted to the mathematical theory of regularization methods and gives an account of the currently available results about regularization methods for linear and nonlinear ill-posed problems.

1. Introduction: Examples of Inverse Problems.- 1.1. Differentiation as an Inverse Problem.- 1.2. Radon Inversion (X-Ray Tomography).- 1.3. Examples of Inverse Problems in Physics.- 1.4. Inverse Problems in Signal and Image Processing.- 1.5. Inverse Problems in Heat Conduction.- 1.6. Parameter Identification.- 1.7. Inverse Scattering.- 2. Ill-Posed Linear Operator Equations.- 2.1. The Moore-Penrose Generalized Inverse.- 2.2. Compact Linear Operators: Singular Value Expansion.- 2.3. Spectral Theory and Functional Calculus.- 3. Regularization Operators.- 3.1. Definition and Basic Results.- 3.2. Order Optimality.- 3.3. Regularization by Projection.- 4. Continuous Regularization Methods.- 4.1. A-priori Parameter Choice Rules.- 4.2. Saturation and Converse Results.- 4.3. The Discrepancy Principle.- 4.4. Improved A-posteriori Rules.- 4.5. Heuristic Parameter Choice Rules.- 4.6. Mollifier Methods.- 5. Tikhonov Regularization.- 5.1. The Classical Theory.- 5.2. Regularization with Projection.- 5.3. Maximum Entropy Regularization.- 5.4. Convex Constraints.- 6. Iterative Regularization Methods.- 6.1. Landweber Iteration.- 6.2. Accelerated Landweber Methods.- 6.3. The ?-Methods.- 7. The Conjugate Gradient Method.- 7.1. Basic Properties.- 7.2. Stability and Convergence.- 7.3. The Discrepancy Principle.- 7.4. The Number of Iterations.- 8. Regularization With Differential Operators.- 8.1. Weighted Generalized Inverses.- 8.2. Regularization with Seminorms.- 8.3. Examples.- 8.4. Hilbert Scales.- 8.5. Regularization in Hilbert Scales.- 9. Numerical Realization.- 9.1. Derivation of the Discrete Problem.- 9.2. Reduction to Standard Form.- 9.3. Implementation of Tikhonov Regularization.- 9.4. Updating the Regularization Parameter.- 9.5. Implementation of Iterative Methods.- 10. TikhonovRegularization of Nonlinear Problems.- 10.1. Introduction.- 10.2. Convergence Analysis.- 10.3. A-posteriori Parameter Choice Rules.- 10.4. Regularization in Hilbert Scales.- 10.5. Applications.- 10.6. Convergence of Maximum Entropy Regularization.- 11. Iterative Methods for Nonlinear Problems.- 11.1. The Nonlinear Landweber Iteration.- 11.2. Newton Type Methods.- A. Appendix.- A.1. Weighted Polynomial Minimization Problems.- A.2. Orthogonal Polynomials.- A.3. Christoffel Functions.