An efficient single-parameter scaling memoryless BFGS algorithm for solving large scale unconstrained optimization problems

Release time:2020-05-18

Impact Factor:3.78

DOI number:10.1109/ACCESS.2020.2992340

Affiliation of Author(s):数学与统计学院

Teaching and Research Group:信息与计算科学

Journal:IEEE-access

Place of Publication:USA

Key Words:Computational efficiency, Convergence of numerical methods, Optimization methods, Algorithm design and analysis

Abstract:In this paper, a new spectral scaling memoryless BFGS algorithm is developed for solving large scale unconstrained optimization problems, where the scaling parameter is chosen so as to minimize all the eigenvalues of search direction matrices. The search directions in this algorithm are also proved to satisfy the Dai-Liao conjugate condition. With this advantage of such search directions, a memoryless BFGS update formula is constructed and an algorithm is developed based on acceleration scheme and restart criterion. Under mild assumptions, global convergence of the algorithm is proved. Numerical experiments demonstrate that the developed algorithm is more robust and efficient than the existing similar ones.

Co-author:Songhai Deng

First Author:Jing, Lv

Indexed by:Journal paper

Document Code:10.1109/ACCESS.2020.2992340

Document Type:J

Volume:8

Page Number:85664-85674

Translation or Not:no

Date of Publication:2020-04-30

Included Journals:SCI