Developments in the direction of solving extremly large problems in Geophysics

TitleDevelopments in the direction of solving extremly large problems in {Geophysics}
Publication TypeConference
Year of Publication2017
AuthorsEmmanouil Daskalakis, Rachel Kuske, Felix J. Herrmann
Conference NameSEG Technical Program Expanded Abstracts
Keywordscycling, least-squares migration, linearized Bregman, SEG, weighted increment

Often in exploration Geophysics we are forced to work with extremely large problems. Acquisitions via dense grids of receivers translate into very large mathematical systems. Usually, depending on the acquisition, the size of the matrix of the system to be solved, can be measured in the millions. The proper way to address this problem is by subsampling. Even though subsampling can reduce the computational efforts required, it can not address stability problems caused by preconditioning and/or instrumental response errors. In this abstract, we introduce a modification of the linearized Bregman solver for these large problems that resolves stability issues.


(SEG, Houston)


Citation Keydaskalakis2017SEGdds