{"created":"2021-10-15T04:07:02.006336+00:00","id":2000029,"links":{},"metadata":{"_buckets":{"deposit":"26d04fdd-ec58-47cf-bb2a-7285ec4fd66d"},"_deposit":{"id":"2000029","owners":[1],"pid":{"revision_id":0,"type":"depid","value":"2000029"},"status":"published"},"_oai":{"id":"oai:niigata-u.repo.nii.ac.jp:02000029","sets":["453:456","485:872:1583:1634266335548"]},"author_link":[],"item_1627361970403":{"attribute_name":"出版タイプ","attribute_value_mlt":[{"subitem_version_resource":"http://purl.org/coar/version/c_970fb48d4fbd8a85","subitem_version_type":"VoR"}]},"item_7_biblio_info_6":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicIssueDates":{"bibliographicIssueDate":"2021-02","bibliographicIssueDateType":"Issued"},"bibliographicPageEnd":"19","bibliographicPageStart":"13","bibliographicVolumeNumber":"73","bibliographic_titles":[{"bibliographic_title":"新潟大学農学部研究報告","bibliographic_titleLang":"ja"},{"bibliographic_title":"Bulletin of the Faculty of Agriculture","bibliographic_titleLang":"en"}]}]},"item_7_description_4":{"attribute_name":"抄録","attribute_value_mlt":[{"subitem_description":"自然科学で対象とされるシステムの現象は、微分方程式(微分モデル)で記述されることが多い。しかしながら、微分モデルの解である方程式は得られないことが多く、現象データにモデルを当てはめることを困難にしている。本研究では、進化戦略のアルゴリズムを適用して微分モデルの数値積分を繰り返すことにより、微分モデルを最適化する手法を提示した。工程として、微分モデルのパラメータを正規分布上で変異させて数値積分し、現象データにより適応したパラメータを選択する。この変異と選択の1回の処理を1世代とし、任意の世代数で処理を繰り返すことにより適切なパラメータを探索する。本手法を評価するため、牛の成長と泌乳期乳量のデータにそれぞれGompertzモデルとWoodモデルを当てはめた。両モデルを微分した方程式に本推定法で当てはめた結果と、両モデルに直接Gauss–Newton法で当てはめた結果を比較した。結果として、本推定法を3万世代実施することにより、両モデルともに、Gauss–Newton法で当てはめたものと同程度の精度でほぼ同じパラメータ値を示した。これらのことは、本推定法が微分モデルのパラメータ推定に有効であることを示唆している。ただし、本推定法の精度は世代数とパラメータ変異に利用する正規分布の変動幅に影響を受けることに留意する。","subitem_description_language":"ja","subitem_description_type":"Abstract"},{"subitem_description":"The phenomena of systems targeted in natural science are often described by differential equations (differential models). However, it is often impossible to obtain equations that are solutions of the differential models, which makes it difficult to fit the models to the phenomena data. We proposed a method for optimising differential models by applying an evolution strategy algorithm and repeating the numerical integration of the differential model. In this process, the parameter set for the differential model is mutated using a normal distribution and numerically integrated, and a parameter set adapted to the phenomena data is selected. One step of mutation and selection is regarded as one generation, and an appropriate parameter is sought by repeating the process for an arbitrary number of generations. To evaluate this method, we applied the Gompertz model to cow growth data and the Wood model to lactation milk yield data. We compared the results applying the Gauss–Newton method to both models directly with those obtained with our method applied to the differential equations of both models. As a result, by implementing this method for 30,000 generations, both models had similar parameter values with the same accuracy as those fitted by the Gauss–Newton method. This suggests that our method is effective for parameter estimation in differential models. Note, however, that the parameter estimation accuracy in this method is affected by the number of generations and the fluctuation range of the normal distribution used for parameter mutation.","subitem_description_language":"en","subitem_description_type":"Abstract"}]},"item_7_publisher_7":{"attribute_name":"出版者","attribute_value_mlt":[{"subitem_publisher":"新潟大学農学部","subitem_publisher_language":"ja"}]},"item_7_source_id_11":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN00183393","subitem_source_identifier_type":"NCID"}]},"item_7_source_id_9":{"attribute_name":"ISSN","attribute_value_mlt":[{"subitem_source_identifier":"0385-8634","subitem_source_identifier_type":"PISSN"}]},"item_access_right":{"attribute_name":"アクセス権","attribute_value_mlt":[{"subitem_access_right":"open access","subitem_access_right_uri":"http://purl.org/coar/access_right/c_abf2"}]},"item_creator":{"attribute_name":"著者","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"板野, 志郎","creatorNameLang":"ja"},{"creatorName":"Itano, Shiro","creatorNameLang":"en"}]},{"creatorNames":[{"creatorName":"坂上, 清一","creatorNameLang":"ja"},{"creatorName":"Sakanoue, Seiichi","creatorNameLang":"en"}]},{"creatorNames":[{"creatorName":"小野, ほのか","creatorNameLang":"ja"},{"creatorName":"Ono, Honoka","creatorNameLang":"en"}]},{"creatorNames":[{"creatorName":"亀山, 亜美","creatorNameLang":"ja"},{"creatorName":"Kameyama, Ami","creatorNameLang":"en"}]},{"creatorNames":[{"creatorName":"大和田, 章生","creatorNameLang":"ja"},{"creatorName":"Owada, Akio","creatorNameLang":"en"}]},{"creatorNames":[{"creatorName":"長谷川, 賢治","creatorNameLang":"ja"},{"creatorName":"Hasegawa, Kenji","creatorNameLang":"en"}]},{"creatorNames":[{"creatorName":"田中, 繁史","creatorNameLang":"ja"},{"creatorName":"Tanaka, Shigefumi","creatorNameLang":"en"}]},{"creatorNames":[{"creatorName":"山城, 秀昭","creatorNameLang":"ja"},{"creatorName":"Yamashiro, Hideaki","creatorNameLang":"en"}]},{"creatorNames":[{"creatorName":"吉田, 智佳子","creatorNameLang":"ja"},{"creatorName":"Yoshida, Chikako","creatorNameLang":"en"}]}]},"item_files":{"attribute_name":"ファイル情報","attribute_type":"file","attribute_value_mlt":[{"accessrole":"open_access","date":[{"dateType":"Available","dateValue":"2021-10-15"}],"displaytype":"detail","filename":"73_13-19.pdf","filesize":[{"value":"638KB"}],"format":"application/pdf","licensetype":"license_note","mimetype":"application/pdf","url":{"objectType":"fulltext","url":"https://niigata-u.repo.nii.ac.jp/record/2000029/files/73_13-19.pdf"},"version_id":"93739c14-4e8f-4d50-b29c-c1550e5617f6"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"最適化","subitem_subject_language":"ja","subitem_subject_scheme":"Other"},{"subitem_subject":"進化戦略","subitem_subject_language":"ja","subitem_subject_scheme":"Other"},{"subitem_subject":"数値積分","subitem_subject_language":"ja","subitem_subject_scheme":"Other"},{"subitem_subject":"微分モデル","subitem_subject_language":"ja","subitem_subject_scheme":"Other"},{"subitem_subject":"differential model","subitem_subject_language":"en","subitem_subject_scheme":"Other"},{"subitem_subject":"evolutionary strategy","subitem_subject_language":"en","subitem_subject_scheme":"Other"},{"subitem_subject":"numerical integration","subitem_subject_language":"en","subitem_subject_scheme":"Other"},{"subitem_subject":"optimization","subitem_subject_language":"en","subitem_subject_scheme":"Other"}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"jpn"}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourcetype":"departmental bulletin paper","resourceuri":"http://purl.org/coar/resource_type/c_6501"}]},"item_title":"進化戦略と数値積分を利用した微分モデルパラメータの最適化","item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"進化戦略と数値積分を利用した微分モデルパラメータの最適化","subitem_title_language":"ja"},{"subitem_title":"Optimization of differential model parameters using evolutionary strategy and numerical integration","subitem_title_language":"en"}]},"item_type_id":"7","owner":"1","path":["456","1634266335548"],"pubdate":{"attribute_name":"PubDate","attribute_value":"2021-10-15"},"publish_date":"2021-10-15","publish_status":"0","recid":"2000029","relation_version_is_last":true,"title":["進化戦略と数値積分を利用した微分モデルパラメータの最適化"],"weko_creator_id":"1","weko_shared_id":-1},"updated":"2022-12-15T04:33:06.742957+00:00"}