Title :
A Modified BFGS Trust Region Method
Author :
Lu, Yunlong ; Jiang, Xiaowei ; Yang, Yueting
Author_Institution :
Sch. of Math., Beihua Univ., Jilin, China
Abstract :
We propose a new trust region method that employs both the modified BFGS update and Amijio line search. The method exploits the information of function and gradient, and ensures the Hessian matrix of trust region subproblem positive-definite. At some assumptions, the global convergence and superlinear convergence property are proposed. Finally, numerical experiments show that the method is efficient.
Keywords :
Hessian matrices; convergence; gradient methods; optimisation; search problems; Armijo line search; Hessian matrix; global convergence; modified BFGS trust region method; superlinear convergence; trust region subproblem positive-definite; Convergence; Eigenvalues and eigenfunctions; Helium; Iterative algorithms; Iterative methods; Mathematics; Optimization methods; Armijio line search; global convergence; superlinear convergence; the modified BFGS update; trust region;
Conference_Titel :
Computational Science and Optimization (CSO), 2010 Third International Joint Conference on
Conference_Location :
Huangshan, Anhui
Print_ISBN :
978-1-4244-6812-6
Electronic_ISBN :
978-1-4244-6813-3
DOI :
10.1109/CSO.2010.106
