Back to search

University of British Columbia

Numerical strategies for the solution of inverse problems

Abstract

dc:description

This thesis deals with the numerical solutions of linear and nonlinear inverse problems. The goal of this thesis is to review and develop new techniques for solving such problems. In so doing, the computational tools for solving inverse problems are comprehensively studied. The thesis can be divided into two parts. In the first part, linear inverse theory is dealt with. Methods to estimate noise and efficiently invert large and full matrixes are reviewed and developed. Emphasis is given to Generalized Cross Validation (GCV) for noise estimation, and to Krylov space methods for efficient methods to invert large systems. This part is summarized by applying and comparing the methods developed on linear inverse problems which arise in gravity and tomography. In the second part of this thesis, extensive use of the linear algebra and the noise estimation methods which were developed in the first part of the thesis is made. A review of the current methods to carry out nonlinear inverse problems is given. A test example is constructed to demonstrate that these methods may fail. Next, a new algorithm for solving nonlinear inverse problems is developed. The algorithm is based on the ability to differentiate between correlated errors which comes from the linearization, and non-correlated noise which comes from the measurement. Based on these two types of noise, a regularization procedure which has two parts is developed. The first part is made of global regularization, to deal with the measurement noise, and the second part is made from a local regularization, to deal with the nonlinearity. The thesis demonstrates that GCV can be used in order to determine the measurement noise, and the Damped Gauss- Newton method can be used in order to deal with the local nonlinear terms. Another aspect of nonlinear inverse theory which is developed in this work concerns approximate sensitivities. A new formulation is suggested for the approximate sensitivities and bounds are calculated using this formulation. This part is summarized by applying the techniques to the nonlinear gravity problem and to the magnetotelluric problem.

Degree

thesis:*
Name thesis:degree_name
Doctor of Philosophy - PhD
Level thesis:degree_level
doctoral
Discipline thesis:degree_discipline
Geophysics
Grantor dc:publisher
University of British Columbia
Year dc:date
1997

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Haber, Eldad

Rights

dc:rights
Statement dc:rights
  • For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use.
Language dc:language
eng

Identifiers

dc:identifier.*
Handle dc:identifier
http://hdl.handle.net/2429/8641

Chain of custody

source
Harvested from
University of British Columbia
Base URL
circle.library.ubc.ca/oai/request
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
related terms
citation

Haber, Eldad. Numerical strategies for the solution of inverse problems. doctoral thesis, University of British Columbia, 1997. http://hdl.handle.net/2429/8641