Phenotype building blocks and geometric crossover in structural optimisation

Alireza Maheri, T. Macquart, D. Safari, M. R. Maheri

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Macro or phenotype building blocks (PBBs) contain information at phenotype level. PBBs are either the components of a multi-component system, or different parts of a continuous system with different design qualities and/or evaluation measures. Using PBBs can lead to enhancement of the search efficiency by utilising problem specific search operators and heuristics applied on the PBBs. In order to preserve, propagate and recombine good building blocks efficiently, building blocks should have a low probability of being disrupted by crossover. Therefore, the crossover operation should be designed at phenotype level. Geometric crossover (GCO) is applied on the phenotype rather than the genotype. Identifying PBBs and using GCO, partial fitness can be defined and employed to improve the performance of the search algorithm. Another advantage of using GCO is as a result of its ease of application to nonfixed- length and variable-length chromosomes. Variable-length chromosomes are a common feature of topology optimisation problems when genetic algorithms are employed. GCO can be easily applied to the structural optimisation problems including, topology optimisation without predefining a grid. These two advantages have been demonstrated by implementing GCO in genetic algorithms employed for optimisation of plane trusses supporting distributed loads, two-dimensional steel frames and wind turbine blades.

Original languageEnglish
Title of host publicationProceedings of the 8th International Conference on Engineering Computational Technology, ECT 2012
PublisherCIVIL COMP PRESS
Pages728-741
Number of pages14
Volume100
ISBN (Print)9781905088553
Publication statusPublished - 2012
Event8th International Conference on Engineering Computational Technology, ECT 2012 - Dubrovnik, Croatia
Duration: 4 Sep 20127 Sep 2012

Conference

Conference8th International Conference on Engineering Computational Technology, ECT 2012
CountryCroatia
CityDubrovnik
Period4/09/127/09/12

Fingerprint

Structural optimization
Shape optimization
Chromosomes
Genetic algorithms
Trusses
Wind turbines
Turbomachine blades
Macros
Steel

Keywords

  • Genetic algorithm
  • Geometric crossover
  • Partial fitness
  • Phenotype building block
  • Structural optimisation
  • Topology optimisation

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Civil and Structural Engineering
  • Artificial Intelligence
  • Environmental Engineering

Cite this

Maheri, A., Macquart, T., Safari, D., & Maheri, M. R. (2012). Phenotype building blocks and geometric crossover in structural optimisation. In Proceedings of the 8th International Conference on Engineering Computational Technology, ECT 2012 (Vol. 100, pp. 728-741). CIVIL COMP PRESS.

Phenotype building blocks and geometric crossover in structural optimisation. / Maheri, Alireza; Macquart, T.; Safari, D.; Maheri, M. R.

Proceedings of the 8th International Conference on Engineering Computational Technology, ECT 2012. Vol. 100 CIVIL COMP PRESS, 2012. p. 728-741.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Maheri, A, Macquart, T, Safari, D & Maheri, MR 2012, Phenotype building blocks and geometric crossover in structural optimisation. in Proceedings of the 8th International Conference on Engineering Computational Technology, ECT 2012. vol. 100, CIVIL COMP PRESS, pp. 728-741, 8th International Conference on Engineering Computational Technology, ECT 2012, Dubrovnik, Croatia, 4/09/12.
Maheri A, Macquart T, Safari D, Maheri MR. Phenotype building blocks and geometric crossover in structural optimisation. In Proceedings of the 8th International Conference on Engineering Computational Technology, ECT 2012. Vol. 100. CIVIL COMP PRESS. 2012. p. 728-741
Maheri, Alireza ; Macquart, T. ; Safari, D. ; Maheri, M. R. / Phenotype building blocks and geometric crossover in structural optimisation. Proceedings of the 8th International Conference on Engineering Computational Technology, ECT 2012. Vol. 100 CIVIL COMP PRESS, 2012. pp. 728-741
@inproceedings{6d119600d9ff4310aacd4587f1f753e9,
title = "Phenotype building blocks and geometric crossover in structural optimisation",
abstract = "Macro or phenotype building blocks (PBBs) contain information at phenotype level. PBBs are either the components of a multi-component system, or different parts of a continuous system with different design qualities and/or evaluation measures. Using PBBs can lead to enhancement of the search efficiency by utilising problem specific search operators and heuristics applied on the PBBs. In order to preserve, propagate and recombine good building blocks efficiently, building blocks should have a low probability of being disrupted by crossover. Therefore, the crossover operation should be designed at phenotype level. Geometric crossover (GCO) is applied on the phenotype rather than the genotype. Identifying PBBs and using GCO, partial fitness can be defined and employed to improve the performance of the search algorithm. Another advantage of using GCO is as a result of its ease of application to nonfixed- length and variable-length chromosomes. Variable-length chromosomes are a common feature of topology optimisation problems when genetic algorithms are employed. GCO can be easily applied to the structural optimisation problems including, topology optimisation without predefining a grid. These two advantages have been demonstrated by implementing GCO in genetic algorithms employed for optimisation of plane trusses supporting distributed loads, two-dimensional steel frames and wind turbine blades.",
keywords = "Genetic algorithm, Geometric crossover, Partial fitness, Phenotype building block, Structural optimisation, Topology optimisation",
author = "Alireza Maheri and T. Macquart and D. Safari and Maheri, {M. R.}",
year = "2012",
language = "English",
isbn = "9781905088553",
volume = "100",
pages = "728--741",
booktitle = "Proceedings of the 8th International Conference on Engineering Computational Technology, ECT 2012",
publisher = "CIVIL COMP PRESS",

}

TY - GEN

T1 - Phenotype building blocks and geometric crossover in structural optimisation

AU - Maheri, Alireza

AU - Macquart, T.

AU - Safari, D.

AU - Maheri, M. R.

PY - 2012

Y1 - 2012

N2 - Macro or phenotype building blocks (PBBs) contain information at phenotype level. PBBs are either the components of a multi-component system, or different parts of a continuous system with different design qualities and/or evaluation measures. Using PBBs can lead to enhancement of the search efficiency by utilising problem specific search operators and heuristics applied on the PBBs. In order to preserve, propagate and recombine good building blocks efficiently, building blocks should have a low probability of being disrupted by crossover. Therefore, the crossover operation should be designed at phenotype level. Geometric crossover (GCO) is applied on the phenotype rather than the genotype. Identifying PBBs and using GCO, partial fitness can be defined and employed to improve the performance of the search algorithm. Another advantage of using GCO is as a result of its ease of application to nonfixed- length and variable-length chromosomes. Variable-length chromosomes are a common feature of topology optimisation problems when genetic algorithms are employed. GCO can be easily applied to the structural optimisation problems including, topology optimisation without predefining a grid. These two advantages have been demonstrated by implementing GCO in genetic algorithms employed for optimisation of plane trusses supporting distributed loads, two-dimensional steel frames and wind turbine blades.

AB - Macro or phenotype building blocks (PBBs) contain information at phenotype level. PBBs are either the components of a multi-component system, or different parts of a continuous system with different design qualities and/or evaluation measures. Using PBBs can lead to enhancement of the search efficiency by utilising problem specific search operators and heuristics applied on the PBBs. In order to preserve, propagate and recombine good building blocks efficiently, building blocks should have a low probability of being disrupted by crossover. Therefore, the crossover operation should be designed at phenotype level. Geometric crossover (GCO) is applied on the phenotype rather than the genotype. Identifying PBBs and using GCO, partial fitness can be defined and employed to improve the performance of the search algorithm. Another advantage of using GCO is as a result of its ease of application to nonfixed- length and variable-length chromosomes. Variable-length chromosomes are a common feature of topology optimisation problems when genetic algorithms are employed. GCO can be easily applied to the structural optimisation problems including, topology optimisation without predefining a grid. These two advantages have been demonstrated by implementing GCO in genetic algorithms employed for optimisation of plane trusses supporting distributed loads, two-dimensional steel frames and wind turbine blades.

KW - Genetic algorithm

KW - Geometric crossover

KW - Partial fitness

KW - Phenotype building block

KW - Structural optimisation

KW - Topology optimisation

UR - http://www.scopus.com/inward/record.url?scp=84893842258&partnerID=8YFLogxK

M3 - Conference contribution

SN - 9781905088553

VL - 100

SP - 728

EP - 741

BT - Proceedings of the 8th International Conference on Engineering Computational Technology, ECT 2012

PB - CIVIL COMP PRESS

ER -