Non-monotone Nonlinear Conjugate Gradient Method Based on the Trust Region Technique

SUN Qing-ying,XU Min-cai,LIU Li-min

DOI: https://doi.org/10.3969/j.issn.1005-3085.2011.05.013

2011-01-01

Abstract:The conjugate gradient method has played a special role for solving the large scale un-constrained optimization problems due to the simplicity and the low storage.In this paper,based on the trust region technique and the modified quasi-Newton equation,a new non-monotone nonlinear conjugate gradient method for the unconstrained optimization problem is presented by combining with Zhang’s non-monotone strategy.In each iteration,the new algorithm generates a suitable trust region radius automatically and obtain the next iterate by solving a simple subproblem.The trust region approach is used in the new algorithm to guarantee the global convergence.The new method has stable convergence and simple computation,so as to solve the ill-posed and large scale problems.The numerical results show that the new method is effective.