Comparative study of several classical and modified conjugate gradient using inexact line search / Uzir Ibrahim

• Main Author: Ibrahim, Uzir
• Format: Thesis
• Language: English
• Published: 2018
• Subjects: Equations; Mathematical statistics. Probabilities; Analysis; Analytical methods used in the solution of physical problems; Algorithms
• Online Access: http://ir.uitm.edu.my/id/eprint/41318/1/41318.pdf; http://ir.uitm.edu.my/id/eprint/41318/
• Institution: Universiti Teknologi Mara

Conjugate gradient (CG) method is one of the method in finding minimization or maximization of objective function to solve unconstrained optimization problems. Many studies have been conducted recently to improve this method but some of the modified CG seem to be more difficult and complicated than the older ones previously suggested. Therefore, in this paper, some simple classical and modified CG method will be compared based on number of iterations and CPU time to find the best solution point. The global convergence result is solved using strong Wolfe-Powell (SWP) inexact line search. Four difference initial point and five test functions is being used to test the efficiency of the CG coefficient.