{"metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00209858","sets":["1164:5305:10533:10534"]},"path":["10534"],"owner":"44499","recid":"209858","title":["強化学習による連鎖型落ち物パズルゲームの研究"],"pubdate":{"attribute_name":"公開日","attribute_value":"2021-02-26"},"_buckets":{"deposit":"d3486995-2569-4808-aa7f-6e1196a97492"},"_deposit":{"id":"209858","pid":{"type":"depid","value":"209858","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"強化学習による連鎖型落ち物パズルゲームの研究","author_link":["530169","530166","530168","530167"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"強化学習による連鎖型落ち物パズルゲームの研究"},{"subitem_title":"A Study on Falling Block Puzzle Game by Reinforcement Learning","subitem_title_language":"en"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"強いAIプレイヤ1","subitem_subject_scheme":"Other"}]},"item_type_id":"4","publish_date":"2021-02-26","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"松江工業高等専門学校"},{"subitem_text_value":"松江工業高等専門学校"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"National Institute of Technology, Matsue College","subitem_text_language":"en"},{"subitem_text_value":"National Institute of Technology, Matsue College","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/209858/files/IPSJ-GI21045009.pdf","label":"IPSJ-GI21045009.pdf"},"date":[{"dateType":"Available","dateValue":"2023-02-26"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-GI21045009.pdf","filesize":[{"value":"1.1 MB"}],"mimetype":"application/pdf","priceinfo":[{"tax":["include_tax"],"price":"660","billingrole":"5"},{"tax":["include_tax"],"price":"330","billingrole":"6"},{"tax":["include_tax"],"price":"0","billingrole":"18"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"8ae93b3f-c5e0-422a-8bd4-87312f829f02","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2021 by the Information Processing Society of Japan"}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"杉江, 矢"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"橋本, 剛"}],"nameIdentifiers":[{}]}]},"item_4_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Nao, Sugie","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Tsuyoshi, Hashimoto","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AA11362144","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":"2188-8736","subitem_source_identifier_type":"ISSN"}]},"item_4_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"近年，ゲームAIの汎用機械学習の研究が注目されているが，成功しているゲームはブロック崩しなど単純なものがほとんどである．テトリスは人間にとっては単純に感じるが，汎用機械学習の題材としては難しいことが報告されている．DQNのような強化学習がうまくいく条件として，ランダムな操作である程度の報酬が得られる必要がある．そして，連鎖型落ち物パズルゲームと呼ばれるジャンルはこの条件を満たすと考えた．本稿では，DQNを用いて連鎖型落ち物パズルゲームであるパネルでポン・ぷよぷよ・コラムスにおいて大連鎖・高得点を獲得するAIを作成することを目的とする．結果として，ニューロエボリューションと比較して多くの連鎖を達成し高い汎用性があることがわかった．","subitem_description_type":"Other"}]},"item_4_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"In recent years, research on general-purpose machine learning of game AI has attracted attention, but successful games are as simple as Breakout. Although Tetris feels simple to humans, it is reported that it is difficult as a subject of general-purpose machine learning. A condition for reinforcement learning such as DQN to work is that some reward must be obtained at random operations. With this background, the game genre called chained falling block puzzle game met this requirement. The aim of this paper is to create an AI that performs many chains by machine learning using DQN, applying Tetris Attack and Puyopuyo and Columns as a chained falling block puzzle game. As a result, this method achieved the larger number of chains and more versatility than neuro-evolution.","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"6","bibliographic_titles":[{"bibliographic_title":"研究報告ゲーム情報学（GI）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2021-02-26","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"9","bibliographicVolumeNumber":"2021-GI-45"}]},"relation_version_is_last":true,"weko_creator_id":"44499"},"id":209858,"updated":"2025-01-19T18:22:02.418613+00:00","links":{},"created":"2025-01-19T01:11:08.463026+00:00"}