Regularization of Large-scale Ill-conditioned Least Squares Problems Marielba Rojas L. Danny C. Sorensen October 18, 1996 Abstract Ill-conditioned problems arise in important areas like geophysics, medical imaging and signal processing. The fact that the ill--conditioning is an intrinsic feature of these problems makes it necessary to develop special numerical methods to treat them. Regularization methods belong to this class. The lack of robust regularization methods for large-scale ill-conditioned problems motivated this project. Our goal is to develop a regularization method for the least squares problem as a large-scale discrete ill-posed problem arising in seismic inversion. From crpc.tr@dawn.cs.rice.edu Mon Dec 9 13:21:33 1996 Received: from sparky.caam.rice.edu (sparky.caam.rice.edu [128.42.17.40]) by cs.rice.edu (8.7.1/8.7.1) with ESMTP id OAA15655 for ; Wed, 23 Oct 1996 14:08:58 -0500 (CDT) Received: (from mrojas@localhost) by sparky.caam.rice.edu (8.7.6/8.7.3) id OAA12610 for crpc.tr@cs.rice.edu; Wed, 23 Oct 1996 14:08:58 -0500 (CDT) Date: Wed, 23 Oct 1996 14:08:58 -0500 (CDT) From: Marielba Rojas Larrazabal Message-Id: <199610231908.OAA12610@sparky.caam.rice.edu> To: crpc.tr@cs.rice.edu Subject: abstract new tr -- corrected Regularization of Large-scale Ill-conditioned Least Squares Problems Marielba Rojas L. October 18, 1996 Abstract Ill-conditioned problems arise in important areas like geophysics, medical imaging and signal processing. The fact that the ill-conditioning is an intrinsic feature of these problems makes it necessary to develop special numerical methods to treat them. Regularization methods belong to this class. The lack of robust regularization methods for large-scale ill-conditioned problems motivated this project. Our goal is to develop a regularization method for the least squares problem as a large-scale discrete ill-posed problem arising in seismic inversion.