A memoryless symmetric rank-one method with sufficient descent property for unconstrained optimization

Shummin Nakayama, Yasushi Narushima, Hiroshi Yabe

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Quasi-Newton methods are widely used for solving unconstrained optimization problems. However, it is difficult to apply quasi-Newton methods directly to large-scale unconstrained optimization problems, because they need the storage of memories for matrices. In order to overcome this difficulty, memoryless quasi-Newton methods were proposed. Shanno (1978) derived the memoryless BFGS method. Recently, several researchers studied the memoryless quasi-Newton method based on the symmetric rank-one formula. However existing memoryless symmetric rank-one methods do not necessarily satisfy the sufficient descent condition. In this paper, we focus on the symmetric rank-one formula based on the spectral scaling secant condition and derive a memoryless quasi-Newton method based on the formula. Moreover we show that the method always satisfies the sufficient descent condition and converges globally for general objective functions. Finally, preliminary numerical results are shown.

Original languageEnglish
Pages (from-to)53-70
Number of pages18
JournalJournal of the Operations Research Society of Japan
Volume61
Issue number1
DOIs
Publication statusPublished - 2018 Jan 1
Externally publishedYes

Fingerprint

Optimization problem
Scaling
Objective function
Large-scale optimization

Keywords

  • Memoryless quasi-Newton method
  • Nonlinear programming
  • Sufficient descent condition
  • Symmetric rank-one formula
  • Unconstrained optimization

ASJC Scopus subject areas

  • Decision Sciences(all)
  • Management Science and Operations Research

Cite this

A memoryless symmetric rank-one method with sufficient descent property for unconstrained optimization. / Nakayama, Shummin; Narushima, Yasushi; Yabe, Hiroshi.

In: Journal of the Operations Research Society of Japan, Vol. 61, No. 1, 01.01.2018, p. 53-70.

Research output: Contribution to journalArticle

@article{4dc282a989bb4f19b741a04d8afb5cf3,
title = "A memoryless symmetric rank-one method with sufficient descent property for unconstrained optimization",
abstract = "Quasi-Newton methods are widely used for solving unconstrained optimization problems. However, it is difficult to apply quasi-Newton methods directly to large-scale unconstrained optimization problems, because they need the storage of memories for matrices. In order to overcome this difficulty, memoryless quasi-Newton methods were proposed. Shanno (1978) derived the memoryless BFGS method. Recently, several researchers studied the memoryless quasi-Newton method based on the symmetric rank-one formula. However existing memoryless symmetric rank-one methods do not necessarily satisfy the sufficient descent condition. In this paper, we focus on the symmetric rank-one formula based on the spectral scaling secant condition and derive a memoryless quasi-Newton method based on the formula. Moreover we show that the method always satisfies the sufficient descent condition and converges globally for general objective functions. Finally, preliminary numerical results are shown.",
keywords = "Memoryless quasi-Newton method, Nonlinear programming, Sufficient descent condition, Symmetric rank-one formula, Unconstrained optimization",
author = "Shummin Nakayama and Yasushi Narushima and Hiroshi Yabe",
year = "2018",
month = "1",
day = "1",
doi = "10.15807/jorsj.61.53",
language = "English",
volume = "61",
pages = "53--70",
journal = "Journal of the Operations Research Society of Japan",
issn = "0453-4514",
publisher = "Operations Research Society of Japan",
number = "1",

}

TY - JOUR

T1 - A memoryless symmetric rank-one method with sufficient descent property for unconstrained optimization

AU - Nakayama, Shummin

AU - Narushima, Yasushi

AU - Yabe, Hiroshi

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Quasi-Newton methods are widely used for solving unconstrained optimization problems. However, it is difficult to apply quasi-Newton methods directly to large-scale unconstrained optimization problems, because they need the storage of memories for matrices. In order to overcome this difficulty, memoryless quasi-Newton methods were proposed. Shanno (1978) derived the memoryless BFGS method. Recently, several researchers studied the memoryless quasi-Newton method based on the symmetric rank-one formula. However existing memoryless symmetric rank-one methods do not necessarily satisfy the sufficient descent condition. In this paper, we focus on the symmetric rank-one formula based on the spectral scaling secant condition and derive a memoryless quasi-Newton method based on the formula. Moreover we show that the method always satisfies the sufficient descent condition and converges globally for general objective functions. Finally, preliminary numerical results are shown.

AB - Quasi-Newton methods are widely used for solving unconstrained optimization problems. However, it is difficult to apply quasi-Newton methods directly to large-scale unconstrained optimization problems, because they need the storage of memories for matrices. In order to overcome this difficulty, memoryless quasi-Newton methods were proposed. Shanno (1978) derived the memoryless BFGS method. Recently, several researchers studied the memoryless quasi-Newton method based on the symmetric rank-one formula. However existing memoryless symmetric rank-one methods do not necessarily satisfy the sufficient descent condition. In this paper, we focus on the symmetric rank-one formula based on the spectral scaling secant condition and derive a memoryless quasi-Newton method based on the formula. Moreover we show that the method always satisfies the sufficient descent condition and converges globally for general objective functions. Finally, preliminary numerical results are shown.

KW - Memoryless quasi-Newton method

KW - Nonlinear programming

KW - Sufficient descent condition

KW - Symmetric rank-one formula

KW - Unconstrained optimization

UR - http://www.scopus.com/inward/record.url?scp=85040363909&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85040363909&partnerID=8YFLogxK

U2 - 10.15807/jorsj.61.53

DO - 10.15807/jorsj.61.53

M3 - Article

VL - 61

SP - 53

EP - 70

JO - Journal of the Operations Research Society of Japan

JF - Journal of the Operations Research Society of Japan

SN - 0453-4514

IS - 1

ER -