{"updated":"2025-01-19T23:58:52.133926+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00192849","sets":["1164:2240:9411:9646"]},"path":["9646"],"owner":"44499","recid":"192849","title":["スケーラブルな並列探索による最適化問題の求解"],"pubdate":{"attribute_name":"公開日","attribute_value":"2018-12-10"},"_buckets":{"deposit":"063df0e4-2d64-47f1-b286-90992b34796a"},"_deposit":{"id":"192849","pid":{"type":"depid","value":"192849","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"スケーラブルな並列探索による最適化問題の求解","author_link":["451081","451084","451083","451080","451082","451085"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"スケーラブルな並列探索による最適化問題の求解"},{"subitem_title":"Solving Optimization Problems with Scalable Parallel Search","subitem_title_language":"en"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"最適化問題","subitem_subject_scheme":"Other"}]},"item_type_id":"4","publish_date":"2018-12-10","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"福井大学"},{"subitem_text_value":"福井大学"},{"subitem_text_value":"理化学研究所"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"University of Fukui","subitem_text_language":"en"},{"subitem_text_value":"University of Fukui","subitem_text_language":"en"},{"subitem_text_value":"RIKEN","subitem_text_language":"en"}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"jpn"}]},"item_publisher":{"attribute_name":"出版者","attribute_value_mlt":[{"subitem_publisher":"情報処理学会","subitem_publisher_language":"ja"}]},"publish_status":"0","weko_shared_id":-1,"item_file_price":{"attribute_name":"Billing file","attribute_type":"file","attribute_value_mlt":[{"url":{"url":"https://ipsj.ixsq.nii.ac.jp/record/192849/files/IPSJ-HPC18167024.pdf","label":"IPSJ-HPC18167024.pdf"},"date":[{"dateType":"Available","dateValue":"2020-12-10"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-HPC18167024.pdf","filesize":[{"value":"812.3 kB"}],"mimetype":"application/pdf","priceinfo":[{"tax":["include_tax"],"price":"660","billingrole":"5"},{"tax":["include_tax"],"price":"330","billingrole":"6"},{"tax":["include_tax"],"price":"0","billingrole":"14"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"3a045de1-6680-4a45-9a6b-31968d485afb","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2018 by the Information Processing Society of Japan"}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"泉, 翔太"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"石井, 大輔"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"美添, 一樹"}],"nameIdentifiers":[{}]}]},"item_4_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Shota, Izumi","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Daisuke, Ishii","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Kazuki, Yoshizoe","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN10463942","subitem_source_identifier_type":"NCID"}]},"item_4_textarea_12":{"attribute_name":"Notice","attribute_value_mlt":[{"subitem_textarea_value":"SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc."}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourceuri":"http://purl.org/coar/resource_type/c_18gh","resourcetype":"technical report"}]},"item_4_source_id_11":{"attribute_name":"ISSN","attribute_value_mlt":[{"subitem_source_identifier":"2188-8841","subitem_source_identifier_type":"ISSN"}]},"item_4_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"Branch-and-bound 等の探索に基づく最適化手法の並列化は HPC の古典的テーマである．Saraswat らが提案した大域的負荷分散 （GLB） ライブラリは，同型の並列ワーカーからなる機構を提供し，不均一なタスク （例 : 並列探索） を 16K コア程度までスケールさせることができる．しかし，GLB ライブラリを最適化問題へ適用すると，暫定解を効率よく並列ワーカー間で共有することが課題となる．そこで，本研究では GLB ライブラリを拡張し，暫定解の共有速度と通信コストのトレードオフを解決することを目指す．また，拡張 GLB ライブラリのベンチマークとして，木構造に基づく最適化問題を用いる．ワーカー間で暫定解を共有する機能として，通信ワーカーを動的に選択する方式と，ワーカー間の超立方体ネットワークを用いる方式を実装した．拡張 GLB ライブラリを用いてベンチマーク問題の並列ソルバーを実装し，複数インスタンスを求解する評価を実験したところ，784 コア使用時に 429-824 倍の速度向上 （並列化効率 0.55-1.05） を達成した．","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"6","bibliographic_titles":[{"bibliographic_title":"研究報告ハイパフォーマンスコンピューティング（HPC）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2018-12-10","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"24","bibliographicVolumeNumber":"2018-HPC-167"}]},"relation_version_is_last":true,"weko_creator_id":"44499"},"created":"2025-01-19T00:58:31.912478+00:00","id":192849,"links":{}}