Numerical experiments on size optimization of truss structures by use of Genetic Algorithm (an approach for optimization using fully stressed design and Genetic Algorithm)

M. Asayama, Hiroshi Hasegawa, K. Kawamo

研究成果: Article

2 引用 (Scopus)

抄録

Genetic Algorithm (GA) is known to be excellent in searching globally optimal solutions, though it requires a large population size of the chromosome and a large number of generation. Some studies are carried out in order to reduce the population size and the number of generations to obtain the solutions. For the purpose, the Authors try to discuss size optimization of truss structures whose members are discretized in the cross-sectional areas by using GA. The structure is coded by a binary string consisting of substrings, which represent the cross-sectional areas of the members and whose numbers are equal to the numbers of the members. Two types of GA models are adopted. In one type the operations are applied to all the substrings simultaneously during a generation. On the other hand, in the other the operations are applied to each substring, one at time. Various kinds of GA strategies, such as simple GA, elite strategy, hybrid strategy combined GA with fully stressed design, etc, are tested in numerical experiments. From their results the authors confirm that the hybrid strategy are most efficient and stable for searching globally optimal solutions.

元の言語English
ページ(範囲)1234-1241
ページ数8
ジャーナルNippon Kikai Gakkai Ronbunshu, A Hen/Transactions of the Japan Society of Mechanical Engineers, Part A
62
発行部数597
出版物ステータスPublished - 1996 5
外部発表Yes

Fingerprint

Genetic algorithms
Experiments
Chromosomes

ASJC Scopus subject areas

  • Mechanical Engineering

これを引用

@article{150e5c67c5cc46d1bc48bc6880f70b8f,
title = "Numerical experiments on size optimization of truss structures by use of Genetic Algorithm (an approach for optimization using fully stressed design and Genetic Algorithm)",
abstract = "Genetic Algorithm (GA) is known to be excellent in searching globally optimal solutions, though it requires a large population size of the chromosome and a large number of generation. Some studies are carried out in order to reduce the population size and the number of generations to obtain the solutions. For the purpose, the Authors try to discuss size optimization of truss structures whose members are discretized in the cross-sectional areas by using GA. The structure is coded by a binary string consisting of substrings, which represent the cross-sectional areas of the members and whose numbers are equal to the numbers of the members. Two types of GA models are adopted. In one type the operations are applied to all the substrings simultaneously during a generation. On the other hand, in the other the operations are applied to each substring, one at time. Various kinds of GA strategies, such as simple GA, elite strategy, hybrid strategy combined GA with fully stressed design, etc, are tested in numerical experiments. From their results the authors confirm that the hybrid strategy are most efficient and stable for searching globally optimal solutions.",
author = "M. Asayama and Hiroshi Hasegawa and K. Kawamo",
year = "1996",
month = "5",
language = "English",
volume = "62",
pages = "1234--1241",
journal = "Nihon Kikai Gakkai Ronbunshu, A Hen/Transactions of the Japan Society of Mechanical Engineers, Part A",
issn = "0387-5008",
publisher = "Japan Society of Mechanical Engineers",
number = "597",

}

TY - JOUR

T1 - Numerical experiments on size optimization of truss structures by use of Genetic Algorithm (an approach for optimization using fully stressed design and Genetic Algorithm)

AU - Asayama, M.

AU - Hasegawa, Hiroshi

AU - Kawamo, K.

PY - 1996/5

Y1 - 1996/5

N2 - Genetic Algorithm (GA) is known to be excellent in searching globally optimal solutions, though it requires a large population size of the chromosome and a large number of generation. Some studies are carried out in order to reduce the population size and the number of generations to obtain the solutions. For the purpose, the Authors try to discuss size optimization of truss structures whose members are discretized in the cross-sectional areas by using GA. The structure is coded by a binary string consisting of substrings, which represent the cross-sectional areas of the members and whose numbers are equal to the numbers of the members. Two types of GA models are adopted. In one type the operations are applied to all the substrings simultaneously during a generation. On the other hand, in the other the operations are applied to each substring, one at time. Various kinds of GA strategies, such as simple GA, elite strategy, hybrid strategy combined GA with fully stressed design, etc, are tested in numerical experiments. From their results the authors confirm that the hybrid strategy are most efficient and stable for searching globally optimal solutions.

AB - Genetic Algorithm (GA) is known to be excellent in searching globally optimal solutions, though it requires a large population size of the chromosome and a large number of generation. Some studies are carried out in order to reduce the population size and the number of generations to obtain the solutions. For the purpose, the Authors try to discuss size optimization of truss structures whose members are discretized in the cross-sectional areas by using GA. The structure is coded by a binary string consisting of substrings, which represent the cross-sectional areas of the members and whose numbers are equal to the numbers of the members. Two types of GA models are adopted. In one type the operations are applied to all the substrings simultaneously during a generation. On the other hand, in the other the operations are applied to each substring, one at time. Various kinds of GA strategies, such as simple GA, elite strategy, hybrid strategy combined GA with fully stressed design, etc, are tested in numerical experiments. From their results the authors confirm that the hybrid strategy are most efficient and stable for searching globally optimal solutions.

UR - http://www.scopus.com/inward/record.url?scp=0030143698&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0030143698&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:0030143698

VL - 62

SP - 1234

EP - 1241

JO - Nihon Kikai Gakkai Ronbunshu, A Hen/Transactions of the Japan Society of Mechanical Engineers, Part A

JF - Nihon Kikai Gakkai Ronbunshu, A Hen/Transactions of the Japan Society of Mechanical Engineers, Part A

SN - 0387-5008

IS - 597

ER -