A Sufficient Descent Property for a Different Parameter to Enhance Three-Term Method | ||
AL-Rafidain Journal of Computer Sciences and Mathematics | ||
Volume 17, Issue 1, June 2023, Pages 107-113 PDF (819.53 K) | ||
Document Type: Research Paper | ||
DOI: 10.33899/csmj.2023.179508 | ||
Authors | ||
Ghada Moayid Al-Naemi; Samaa Abdulkader Al-bakri* | ||
Department of Mathematics College of Computer Science and Mathematics, University of Mosul, Mosul, Iraq | ||
Abstract | ||
In this paper, we derive a new parameter µk-1 for the three-term CG (N3T) algorithm for solving unconstrained optimization problems. As demonstrated by its calculations and proof, the parameter µk-1 worth is determined by T , and the study mentions four different types of T . The search directions of this algorithm are always sufficiently descent when using strong Wolfe line search (SWC). Under reasonable assumptions, the proposed algorithm achieves global convergence. The numerical comparison demonstrates that our proposed method works well for solving unconstrained optimization problems. | ||
Keywords | ||
Unconstrained optimization; Three term; Conjugate gradient; strong Wolfe condition; sufficient descent method; global convergent | ||
Statistics Article View: 110 PDF Download: 127 |