Using homotopy to invert geophysical data
Academic Article
Overview
Identity
Additional Document Info
Other
View All
Overview
abstract
Homotopy is a powerful tool for solving nonlinear equations. It is used here to solve smalldimensional geophysical inverse problems by locating the solutions of the governing normal equations. An EulerNewton numerical continuation scheme is used to map trajectories in model space that start from a prescribed solution to a trivial set of equations and terminate at a solution to the inverse problem. The trajectories often map out a continuum of equivalent solutions that are caused by model equivalences or overparameterization. This allows exploration of the solution space topology. The homotopy method, in this application, is relatively insensitive to the choice of starting model. Several examples based on synthetic controlledsource electromagnetic (CSEM) responses are shown to illustrate the method. An inversion of actual CSEM data from the Canadian Shield is also provided.
