On the Global Convergence of A Parameter-Adjusting Levenberg-Marquardt Method

Liyan Qi,Xiantao Xiao,Liwei Zhang
DOI: https://doi.org/10.3934/naco.2015.5.25
2015-01-01
Numerical Algebra, Control and Optimization
Abstract:The Levenberg-Marquardt (LM) method is a classical but popular method for solving nonlinear equations. Based on the trust region technique, we propose a parameter-adjusting LM (PALM) method, in which the LM parameter mu(k) is self-adjusted at each iteration based on the ratio between actual reduction and predicted reduction. Under the level-bounded condition, we prove the global convergence of PALM. We also propose a modified parameter-adjusting LM (MPALM) method. Numerical results show that the two methods are very efficient.
What problem does this paper attempt to address?