A NEW WEIGHTED CONJUGATE GRADIENT METHOD
Abstract. A new weighted multi-update conjugate gradient method is proposed for unconstrained optimization which is considered as memoryless variable metric methods. It is derived for inexact line searches and evaluated numerically against Shanno's two memoryless quasi-Newton methods. The numerical results indicate that, in general, the new method is numerically superior to Shanno's methods.
AMS Subject classification: 65K10


Download full article from here (pdf format).

DOI: 10.12732/ijam.v26i4.1

Volume: 26
Issue: 4
Year: 2013