Image fragmented learning for data-driven topology design

Yang, Yusibo; Yaji, Kentaro; Yamasaki, Shintaro; Fujita, Kikuo

doi:https://doi.org/10.1007/s00158-025-04054-3

Image fragmented learning for data-driven topology design

Yang, Yusibo; Yaji, Kentaro; Yamasaki, Shintaro et al.

Structural and Multidisciplinary Optimization, 2025, 68(8), 148

アクセス数:5件（2025-09-03 17:05 集計）

固定URL: https://hdl.handle.net/11094/102593

閲覧可能ファイル

ファイル	フォーマット	利用条件	サイズ	閲覧回数	利用開始日	説明	information
StructMutltidiscipOpt_68_8	pdf	なし	5.38 MB	6	2025.08.29

Creative Commons : 表示

論文情報

ファイル出力 EndNote Basic出力 Mendeley出力

タイトル	Image fragmented learning for data-driven topology design
著者	著者名 Yang, Yusibo
	NRID 1000090779373 1000090779373 ORCID 0000-0002-4309-2043 0000-0002-4309-2043 researchmap permalink yajiken0609 yajiken0609 Scopus Author ID 55787947500 55787947500 著者名 Yaji, Kentaro
	NRID 1000070581601 1000070581601 著者名 Yamasaki, Shintaro
	NRID 1000010228992 1000010228992 ORCID 0000-0002-2197-1012 0000-0002-2197-1012 researchmap permalink read0014111 read0014111 Scopus Author ID 7404057902 7404057902 著者名 Fujita, Kikuo
キーワード等	Data-driven topology design
	Image fragmented learning
	Multi-objective optimization
	Topology optimization
内容	This version of the article has been accepted for publication, after peer review (when applicable) and is subject to Springer Nature’s AM terms of use, but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: https://doi.org/10.1007/s00158-025-04054-3.
抄録	This paper proposes a data-driven topology design (DDTD) framework, incorporating imagefragmentedlearning that leverages the technique of dividing an image into smaller segments for learning each fragment. This framework is designed to tackle the challenges of high-dimensional, multi-objective optimization problems. Original DDTD methods leverage the sensitivity-free nature and high capacity of deep generative models to effectively address strongly nonlinear problems. However, their training effectiveness significantly diminishes as input size exceeds a certain threshold, which poses challenges in maintaining the high degrees of freedom crucial for accurately representing complex structures. To address this limitation, we split a trained conditional generative adversarial network into two interconnected modules: the first performs dimensionality reduction, compressing high-dimensional data into a lower-dimensional representation, which is then fed into a variational autoencoder (VAE) to generate new low-dimensional data. The second module reconstructs the generated low-dimensional data back into the high-dimensional design space. The effectiveness of the proposed approach is demonstrated through two case studies: the optimization of an L-bracket design problem and a turbulent heat transfer design problem, both involving design variables at a scale unattainable by the conventional VAE-based method.
出版者	Springer Science and Business Media Deutschland GmbH
収録物名	Structural and Multidisciplinary Optimization
巻(号)	68(8)
ページ	148
刊行年月	2025-08-12
言語	英語
ハンドルURL	https://hdl.handle.net/11094/102593
PISSN	1615147X
eISSN	16151488
関連情報 (isIdenticalTo)	DOI (出版社版) https://doi.org/10.1007/s00158-025-04054-3
アクセス権	open access
権利情報	This article is licensed under a Creative Commons Attribution 4.0 International License.
出版タイプ	Version of Record
カテゴリ	学術雑誌論文 Journal Article

資源タイプ	学術雑誌論文
ローカル資源タイプ	学術雑誌論文
DCMI資源タイプ	text
DCTERMS.bibliographicCitation	Structural and Multidisciplinary Optimization. 2025, 68(8), p. 148
DC.title	Image fragmented learning for data-driven topology design
DC.creator	Yang, Yusibo
	Yaji, Kentaro
	Yamasaki, Shintaro
	Fujita, Kikuo
DC.publisher	Springer Science and Business Media Deutschland GmbH
DC.language' scheme='DCTERMS.RFC1766	英語
DC.type' scheme='DCTERMS.DCMIType	text
DCTERMS.issued' scheme='DCTERMS.W3CDTF	2025-08-12
DC.identifier	https://doi.org/10.1007/s00158-025-04054-3
DC.subject	Data-driven topology design
	Image fragmented learning
	Multi-objective optimization
	Topology optimization
DC.description	This version of the article has been accepted for publication, after peer review (when applicable) and is subject to Springer Nature’s AM terms of use, but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: https://doi.org/10.1007/s00158-025-04054-3.
DCTERMS.abstract	This paper proposes a data-driven topology design (DDTD) framework, incorporating imagefragmentedlearning that leverages the technique of dividing an image into smaller segments for learning each fragment. This framework is designed to tackle the challenges of high-dimensional, multi-objective optimization problems. Original DDTD methods leverage the sensitivity-free nature and high capacity of deep generative models to effectively address strongly nonlinear problems. However, their training effectiveness significantly diminishes as input size exceeds a certain threshold, which poses challenges in maintaining the high degrees of freedom crucial for accurately representing complex structures. To address this limitation, we split a trained conditional generative adversarial network into two interconnected modules: the first performs dimensionality reduction, compressing high-dimensional data into a lower-dimensional representation, which is then fed into a variational autoencoder (VAE) to generate new low-dimensional data. The second module reconstructs the generated low-dimensional data back into the high-dimensional design space. The effectiveness of the proposed approach is demonstrated through two case studies: the optimization of an L-bracket design problem and a turbulent heat transfer design problem, both involving design variables at a scale unattainable by the conventional VAE-based method.
DC.rights	This article is licensed under a Creative Commons Attribution 4.0 International License.
DC.format	102593
citation_title	Image fragmented learning for data-driven topology design
citation_author	Yang, Yusibo
	Yaji, Kentaro
	Yamasaki, Shintaro
	Fujita, Kikuo
citation_publisher	Springer Science and Business Media Deutschland GmbH
citation_language	英語
citation_date	2025-08-12
citation_journal_title	Structural and Multidisciplinary Optimization
citation_volume	68
citation_issue	8
citation_firstpage	148
citation_doi	https://doi.org/10.1007/s00158-025-04054-3
citation_public_url	https://hdl.handle.net/11094/102593

Image fragmented learning for data-driven topology design

閲覧可能ファイル

論文情報

論文詳細を表示