Hybrid Methods of Two Conjugate Gradient Coefficients with Modified Condition for Solving Unconstrained Optimization

Authors

Yasir Salih, Faculty of Education, Department of Mathematics, Red Sea University, Port-Sudan, Sudan (Postdoctoral program at Universitas Padjadjaran Indonesia)Follow
Sudradjat, Faculty of Mathematics and Natural Sciences, Department of Mathematics, Universitas Padjadjaran, Bandung, IndonesiaFollow

Abstract

The conjugate gradient (CG) technique is an efficient method for solving nonlinear optimization problems, especially for large scale problems. This is due to simplicity, low memory requirement, and computational cost, which satisfy the sufficient descent condition. This paper presents a new modification of the PRP method which can be considered as a convex combination of PRP and DPRP methods. The performance of the new method is analyzed to ensure the descent property under exact line search. Numerical experiments indicate that this method has good convergence performance and is promising.

Recommended Citation

Salih, Yasir and Sudradjat (2025) "Hybrid Methods of Two Conjugate Gradient Coefficients with Modified Condition for Solving Unconstrained Optimization," Iraqi Journal for Computer Science and Mathematics: Vol. 6: Iss. 2, Article 23.
DOI: https://doi.org/10.52866/2788-7421.1264
Available at: https://ijcsm.researchcommons.org/ijcsm/vol6/iss2/23