The modified quasi‐Newton methods for solving unconstrained optimization problems. (28th May 2018)

• Record Type: Journal Article
• Authors: Dehghani, R.; Hosseini, M.M.; Bidabadi, N.
• Is Part Of: International journal of numerical modelling. Volume 32:Number 1(2019)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2018-05-28
• Subjects: global convergence -- quasi‐Newton method -- quasi‐Newton equation -- unconstrained optimization; Electric networks -- Mathematical models -- Periodicals; Electronics -- Mathematical models -- Periodicals; 621.3011
• DOI: 10.1002/jnm.2459
• Languages: English
• ISSNs: 0894-3370

Abstract

Abstract: The usual quasi‐Newton method at each iteration utilize the gradients and ignores the available function value information. In this paper, we employ Taylor formula and introduce a new quasi‐Newton method, which uses both available gradient and function value information. This method approximates the Hessian matrix with excellent accuracy. The global convergence of these method associated to a general line search rule will be also shown. In addition, we will show that average performance of proposed algorithm is better than some proposed methods.