Damping–undamping strategies for the Levenberg–Marquardt nonlinear least-squares method
نویسنده
چکیده
The speed of the Levenberg–Marquardt ~LM! nonlinear iterative least-squares method depends upon the choice of damping strategy when the fitted parameters are highly correlated. Additive damping with small damping increments and large damping decrements permits LM to efficiently solve difficult problems, including those that otherwise cause stagnation. © 1997 American Institute of Physics. @S0894-1866~97!01801-4#
منابع مشابه
A Levenberg–Marquardt method for estimating polygonal regions
The problem is considered of the estimation of a polygonal region in two dimensions from data approximately marking the outline of the region. A solution is sought by formulating and solving a nonlinear least squares problem. A Levenberg–Marquardt method is developed for this problem, with an implementation which exploits the special structure so that the Levenberg–Marquardt step can be compute...
متن کاملHybrid Levenberg–Marquardt and weak-constraint ensemble Kalman smoother method
The ensemble Kalman smoother (EnKS) is used as a linear least-squares solver in the Gauss–Newton method for the large nonlinear least-squares system in incremental 4DVAR. The ensemble approach is naturally parallel over the ensemble members and no tangent or adjoint operators are needed. Furthermore, adding a regularization term results in replacing the Gauss–Newton method, which may diverge, b...
متن کاملGlobal and Local Convergence of a Levenberg-Marquadt Algorithm for Inverse Problems
The Levenberg-Marquardt algorithm is one of the most popular algorithms for the solution of nonlinear least squares problems. In this paper, we propose and analyze the global and local convergence results of a novel Levenberg-Marquadt method for solving general nonlinear least squares problems. The proposed algorithm enjoys strong convergence properties (global convergence as well as quadratic ...
متن کاملNonlinear least squares and Sobolev gradients
Least squares methods are effective for solving systems of partial differential equations. In the case of nonlinear systems the equations are usually linearized by a Newton iteration or successive substitution method, and then treated as a linear least squares problem. We show that it is often advantageous to form a sum of squared residuals first, and then compute a zero of the gradient with a ...
متن کاملBundle Adjustment with and without Damping
The least squares adjustment (LSA) method is studied as an optimisation problem and shown to be equivalent to the undamped Gauss-Newton (GN) optimisation method. Three problem-independent damping modifications of the GN method are presented: the line-search method of Armijo (GNA); the LevenbergMarquardt algorithm (LM); and Levenberg-Marquardt with Powell dogleg (LMP). Furthermore, an additional...
متن کامل