Yıl:2022   Cilt: 12   Sayı: 2   Alan: Matematik

  1. Anasayfa
  2. Makale Listesi
  3. ID: 33

Isam H. HALİ ORCID Icon,Khalil K. ABBO ORCID Icon,Hassan H. IBRAHİM ORCID Icon

A New Conjugate Gradient Algorithm for Unconstrained Optimization

Parallel gradient methods (CG) include a class of unrestricted optimization algorithms with low memory requirements and strong local and global convergence characteristics .The conjugate gradient methods began with the seminal paper of Hestense and Stiefel in 1952 .The algorithm is presented as an approach to solving similar and positive linear systems, but is an alternative to gausi justice that is perfectly suited to solving big problems. In 1964 Fletcher generalized it to solve the unconstrained optimization problems today, the results of unrestricted improvement are applied in different branches of science, as well as generally in practice. Recently different nonlinear conjugate gradient methods are developed. In this paper, we developed a new nonlinear conjugate gradient algorithm. Deriving this technique on the basis of the ratios property and the case of association, the descent property and global convergence with some mild assumption of the algorithm was proved, Numerical comparison of the algorithm with other related (CG) methods are given. The new algorithm is very effective for solving unrestricted optimization problems.

Anahtar Kelimeler: Conjugate gradient algorithm, unconstrained optimization


A New Conjugate Gradient Algorithm for Unconstrained Optimization

Parallel gradient methods (CG) include a class of unrestricted optimization algorithms with low memory requirements and strong local and global convergence characteristics .The conjugate gradient methods began with the seminal paper of Hestense and Stiefel in 1952 .The algorithm is presented as an approach to solving similar and positive linear systems, but is an alternative to gausi justice that is perfectly suited to solving big problems. In 1964 Fletcher generalized it to solve the unconstrained optimization problems today, the results of unrestricted improvement are applied in different branches of science, as well as generally in practice. Recently different nonlinear conjugate gradient methods are developed. In this paper, we developed a new nonlinear conjugate gradient algorithm. Deriving this technique on the basis of the ratios property and the case of association, the descent property and global convergence with some mild assumption of the algorithm was proved, Numerical comparison of the algorithm with other related (CG) methods are given. The new algorithm is very effective for solving unrestricted optimization problems.

Keywords: Conjugate gradient algorithm, unconstrained optimization

Sayfa Aralığı: 1-12


Atıf İçin

Hali, I. H., Abbo, K. K. & Ibrahim, H. H. (2022). A New Conjugate Gradient Algorithm for Unconstrained Optimization. Journal of Current Researches on Educational Studies, 12 (2), 1-12.


211