Modified shifted variable metric algorithms for solving unconstrained minimization problems
Author(s) -
Abbas Al-Bayati,
Basim A. Hassan
Publication year - 2010
Publication title -
mağallaẗ al-tarbiyaẗ wa-al-ʻilm
Language(s) - English
Resource type - Journals
eISSN - 2664-2530
pISSN - 1812-125X
DOI - 10.33899/edusj.2010.58398
Subject(s) - broyden–fletcher–goldfarb–shanno algorithm , algorithm , metric (unit) , positive definite matrix , minification , variable (mathematics) , quasi newton method , mathematics , computer science , mathematical optimization , newton's method , nonlinear system , mathematical analysis , computer network , operations management , eigenvalues and eigenvectors , physics , asynchronous communication , quantum mechanics , economics
In this paper, we propose a modified version of the shifted VM algorithms where the matrices k H have the form , k k k A I H + = ξ 1 ≥ k where 0 f k ξ and k A are symmetric positive semi definite matrices, usually 0 1 = A , 1 + k A is obtained from k A to satisfy the shifted modified quasi-Newton condition ,we consider our new QN-condition the form , ~ * * 1 k k k k v y A ρ = + * ~ k k k k y v v ξ − = , k k k k v m y y + = * where / k T k k k T k k y v y A y = ρ , and we derive of the modified algorithms. Experimental results indicate that the new proposed algorithms were efficient than the both standard BFGS & the shifted BFGSalgorithms. Modified shifted variable metric algorithms for solving unconstrained ... ٧١
Accelerating Research
Robert Robinson Avenue,
Oxford Science Park, Oxford
OX4 4GP, United Kingdom
Address
John Eccles HouseRobert Robinson Avenue,
Oxford Science Park, Oxford
OX4 4GP, United Kingdom