Collaborative Research: Computational techniques for nonlinear joint inversion

Project: Research project

Description

Scope of Work ASU: Novel regularization algorithms for joint nonlinear inverse problems Ill-posed or ill-conditioned inverse problems arise when measurements or observations are used to infer physical properties of a physical system for which the model and/or the data provide inconsistent information or are incomplete. A given problem can be made well-posed by adding constraints or regularization terms, but the manner in which the model is augmented by such a priori information has a significant impact on results. Although ill-posed linear inverse problems are well-studied, efficient and effective estimation of regularization parameters and operators is often overlooked in the solution of nonlinear inverse problems. PI Renaut from ASU is collaborating with Mathematician Mead and Geophysicist Bradford at Boise State University on the development of effective and practical approaches for solving general nonlinear inverse problems in which data coupling across different models is required. Her efforts are associated mainly with the theoretical background on the computational algorithms, rather than the solution of the motivating and test hydrogeophysics problem. At ASU she will be advising one doctoral student each year and up to 2 undergraduate students each year. The project involves a coordinated REU every summer, work with consultants over the summer and travel to Boise State. This project falls under the scope of Renauts faculty appointment during the academic year. She will receive limited salary support each summer for her research efforts. J-
StatusFinished
Effective start/end date7/1/149/30/18

Funding

  • National Science Foundation (NSF): $90,764.00

Fingerprint

inverse problem
summer
student
physical property
inversion
project