Abstract:
We propose and study the Levenberg–Marquardt method globalized by means of linesearch
for unconstrained optimization problems with possibly nonisolated solutions.
It is well-recognized that this method is an efficient tool for solving systems
of nonlinear equations, especially in the presence of singular and even nonisolated
solutions. Customary globalization strategies for the Levenberg–Marquardt method
rely on linesearch for the squared Euclidean residual of the equation being solved.
In case of unconstrained optimization problem, this equation is formed by putting
the gradient of the objective function equal to zero, according to the Fermat principle.
However, these globalization strategies are not very adequate in the context of
optimization problems, as the corresponding algorithms do not have “preferences” for
convergence to minimizers, maximizers, or any other stationary points. To that end,
in this work we considers a different technique for globalizing convergence of
the Levenberg–Marquardt method, employing linesearch for the objective function of the
original problem. We demonstrate that the proposed algorithm possesses reasonable
global convergence properties, and preserves high convergence rate of the
Levenberg–Marquardt method under weak assumptions.
The work is partially supported by the Russian Foundation for Basic Research (projects no. 17-01-00125_a and 19-51-12003 ННИО_а) and by the Volkswagen Foundation (grant 90306).
Received: 10.01.2019
Bibliographic databases:
Document Type:
Article
UDC:
519
Language: Russian
Citation:
A. F. Izmailov, A. S. Kurennoy, P. I. Stetsyuk, “Levenberg–Marquardt method for unconstrained optimization”, Russian Universities Reports. Mathematics, 24:125 (2019), 60–74
This publication is cited in the following 5 articles:
D. N. Maykov, S. S. Makarov, “Numerical Investigation of Optimization Algorithms for Adapting the Hydrodynamic Model Based on the Results of Well Tests”, Math Models Comput Simul, 15:2 (2023), 289
G.A. Trutnev, A.V. Shchenyatskiy, “Computational Component in a Hemispherical Resonator Gyroscope with its Measuring Component Utilising Alternating Voltage”, HoBMSTU.SIE, 2022, no. 3 (140), 78
Monika Trojanová, Tomáš Čakurda, Alexander Hošovský, Tibor Krenický, “Estimation of Grey-Box Dynamic Model of 2-DOF Pneumatic Actuator Robotic Arm Using Gravity Tests”, Applied Sciences, 11:10 (2021), 4490
Vladimir Viktorovich Pekunov, “Modification of the Marquardt method for training a neural network predictor in eddy viscosity models”, Kibernetika i programmirovanie, 2021, no. 1, 27
P. Stetsyuk, V. Stovba, A. Suprun, “B-FORM OF THE DAVIDON–FLETCHER–POWELL METHOD”, JNAM, 2021, no. 2 (136), 93