Preconditioning a non-linear problem and its application to bidirectional deconvolution
نویسندگان
چکیده
Non-linear optimization problems suffer from local minima. When we use gradient based iterative solvers on these problems, we often find the final solution highly dependent on the initial guess. Here we introduce preconditioning and show how it helps resolve in our current problem—bidirectional deconvolution. The results in three data examples show that preconditioning helps us get a more spiky result when compared with the results without preconditioning. Additionally field data results with preconditioning have fewer precursors, cleaner salt body, more symmetric wavelet, and faster convergence rate than those without preconditioning. In addition to the field data, we theoretically illustrate and practically apply two ways of preconditioning: Prediction-error filter (PEF) preconditioning and Gapped anti-causal leaky integration followed by PEF (GALIP) preconditioning. Unlike PEF preconditioning, GALIP helps produce the result in the central wavelet or other position of the wavelet if we change the length of gap.
منابع مشابه
Logarithm Bidirectional Deconvolution Method With Regularization
In this paper, we theoretically discuss about regularization on logarithm bidirectional deconvolution proposed by Claerbout et al. (2011). We hope model fitting helps reduce the unwelcome strong precursors and noise in the bidirectional deconvolution results. INTRODUCTION Bidirectional deconvolution is an ill-posed and highly non-linear problem with many unexpected traps-local minima. When we t...
متن کاملSome properties of band matrix and its application to the numerical solution one-dimensional Bratu's problem
A Class of new methods based on a septic non-polynomial spline function for the numerical solution one-dimensional Bratu's problem are presented. The local truncation errors and the methods of order 2th, 4th, 6th, 8th, 10th, and 12th, are obtained. The inverse of some band matrixes are obtained which are required in proving the convergence analysis of the presented method. Associated boundary f...
متن کاملPresentation and Solving Non-Linear Quad-Level Programming Problem Utilizing a Heuristic Approach Based on Taylor Theorem
The multi-level programming problems are attractive for many researchers because of their application in several areas such as economic, traffic, finance, management, transportation, information technology, engineering and so on. It has been proven that even the general bi-level programming problem is an NP-hard problem, so the multi-level problems are practical and complicated problems therefo...
متن کاملDiscrete-Time Blind Deconvolution for Distributed Parameter Systems with Dirichlet Boundary Input and Unbounded Output with Application to a Transdermal Alcohol Biosensor
A scheme for the blind deconvolution of blood or breath alcohol concentration from biosensor measured transdermal alcohol concentration based on a parabolic PDE with Dirichlet boundary input and point-wise boundary output is developed. The estimation of the convolution filter corresponding to a particular patient and device is formulated as a nonlinear least squares fit to data. The deconvoluti...
متن کاملPreconditioning non-monotone gradient methods for retrieval of seismic reflection signals
The core problem in seismic exploration is to invert the subsurface reflectivity from the surface recorded seismic data. However, most of the seismic inverse problems are ill-posed by nature. To overcome the ill-posedness, different regularized least squares methods are introduced in the literature. In this paper, we developed a preconditioning non-monotone gradient method, proved it converges ...
متن کامل