A New Type of Step Sizes for Unconstrained Optimization

Basim A. Hassan; Maha S. Younis; Mohammed W. Taha; Abdulkarim Hassan Ibrahim

doi:10.1088/1742-6596/1999/1/012099

A New Type of Step Sizes for Unconstrained Optimization

Basim A. Hassan^*, Maha S. Younis, Mohammed W. Taha, Abdulkarim Hassan Ibrahim

^*Corresponding author for this work

Research output: Contribution to journal › Conference article › peer-review

7 Scopus citations

Abstract

Step sizes is very important for a global convergence gradient method for solving the problems of unconstrained optimization. The new step sizes formulas techniques proposed, the key idea used in the construction of the algorithm is to approximate Hessian by a suitable diagonal matrix, which has been found to be the most efficient in this paper. Under weaker conditions, we define the convergences of the proposed methods. In addition, we will show that performance of proposed algorithm is better than of the gradient descent (GD) method.

Original language	English
Article number	012099
Journal	Journal of Physics: Conference Series
Volume	1999
Issue number	1
DOIs	https://doi.org/10.1088/1742-6596/1999/1/012099
State	Published - 21 Sep 2021
Externally published	Yes
Event	2nd International Virtual Conference on Pure Science, 2IVCPS 2021 - Diwaniyah, Iraq Duration: 21 Apr 2021 → 22 Apr 2021

Bibliographical note

Publisher Copyright:
© Published under licence by IOP Publishing Ltd.

ASJC Scopus subject areas

General Physics and Astronomy

Access to Document

10.1088/1742-6596/1999/1/012099

Cite this

@article{03d6bfb4274241dea7798ebbd79edbc0,

title = "A New Type of Step Sizes for Unconstrained Optimization",

abstract = "Step sizes is very important for a global convergence gradient method for solving the problems of unconstrained optimization. The new step sizes formulas techniques proposed, the key idea used in the construction of the algorithm is to approximate Hessian by a suitable diagonal matrix, which has been found to be the most efficient in this paper. Under weaker conditions, we define the convergences of the proposed methods. In addition, we will show that performance of proposed algorithm is better than of the gradient descent (GD) method.",

author = "Hassan, \{Basim A.\} and Younis, \{Maha S.\} and Taha, \{Mohammed W.\} and Ibrahim, \{Abdulkarim Hassan\}",

note = "Publisher Copyright: {\textcopyright} Published under licence by IOP Publishing Ltd.; 2nd International Virtual Conference on Pure Science, 2IVCPS 2021 ; Conference date: 21-04-2021 Through 22-04-2021",

year = "2021",

month = sep,

day = "21",

doi = "10.1088/1742-6596/1999/1/012099",

language = "English",

volume = "1999",

journal = "Journal of Physics: Conference Series",

issn = "1742-6588",

publisher = "IOP Publishing Ltd.",

number = "1",

}

TY - JOUR

T1 - A New Type of Step Sizes for Unconstrained Optimization

AU - Hassan, Basim A.

AU - Younis, Maha S.

AU - Taha, Mohammed W.

AU - Ibrahim, Abdulkarim Hassan

PY - 2021/9/21

Y1 - 2021/9/21

N2 - Step sizes is very important for a global convergence gradient method for solving the problems of unconstrained optimization. The new step sizes formulas techniques proposed, the key idea used in the construction of the algorithm is to approximate Hessian by a suitable diagonal matrix, which has been found to be the most efficient in this paper. Under weaker conditions, we define the convergences of the proposed methods. In addition, we will show that performance of proposed algorithm is better than of the gradient descent (GD) method.

AB - Step sizes is very important for a global convergence gradient method for solving the problems of unconstrained optimization. The new step sizes formulas techniques proposed, the key idea used in the construction of the algorithm is to approximate Hessian by a suitable diagonal matrix, which has been found to be the most efficient in this paper. Under weaker conditions, we define the convergences of the proposed methods. In addition, we will show that performance of proposed algorithm is better than of the gradient descent (GD) method.

UR - https://www.scopus.com/pages/publications/85116854143

U2 - 10.1088/1742-6596/1999/1/012099

DO - 10.1088/1742-6596/1999/1/012099

M3 - Conference article

AN - SCOPUS:85116854143

SN - 1742-6588

VL - 1999

JO - Journal of Physics: Conference Series

JF - Journal of Physics: Conference Series

IS - 1

M1 - 012099

T2 - 2nd International Virtual Conference on Pure Science, 2IVCPS 2021

Y2 - 21 April 2021 through 22 April 2021

ER -

A New Type of Step Sizes for Unconstrained Optimization

Abstract

Bibliographical note

ASJC Scopus subject areas

Access to Document

Fingerprint

Cite this