• Search Research Projects
  • Search Researchers
  • How to Use
  1. Back to previous page

Segmentation of Time and Space in a Fully Online Reinforcement Learning System

Research Project

Project/Area Number 18K11473
Research Category

Grant-in-Aid for Scientific Research (C)

Allocation TypeMulti-year Fund
Section一般
Review Section Basic Section 61040:Soft computing-related
Research InstitutionOsaka Metropolitan University (2022)
Osaka Prefecture University (2018-2021)

Principal Investigator

Notsu Akira  大阪公立大学, 大学院現代システム科学研究科, 教授 (40405345)

Co-Investigator(Kenkyū-buntansha) 生方 誠希  大阪公立大学, 大学院情報学研究科, 准教授 (10755698)
本多 克宏  大阪公立大学, 大学院情報学研究科, 教授 (80332964)
Project Period (FY) 2018-04-01 – 2023-03-31
Project Status Completed (Fiscal Year 2022)
Budget Amount *help
¥4,290,000 (Direct Cost: ¥3,300,000、Indirect Cost: ¥990,000)
Fiscal Year 2020: ¥1,560,000 (Direct Cost: ¥1,200,000、Indirect Cost: ¥360,000)
Fiscal Year 2019: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000)
Fiscal Year 2018: ¥1,560,000 (Direct Cost: ¥1,200,000、Indirect Cost: ¥360,000)
Keywords強化学習 / クラスタリング / 最適化アルゴリズム / 転移学習 / 学習と進化 / 進化計算 / ニューラルネットワーク / バンディットアルゴリズム / 自己組織化マップ / オンライン学習 / 事前学習 / オンライン型 / 認知モデル
Outline of Final Research Achievements

We modified the growing self-organizing map for reinforcement learning and devised a method for unsupervised learning of state space and state transitions while maintaining learning efficiency, and demonstrated the usefulness of this method. We also showed that the method can adapt to the environment by adaptively changing the hyperparameter settings significantly. Furthermore, we proposed a method for switching methods while estimating several local environments for differential evolution, which is one of the best optimization algorithm methods, and were able to improve the performance. In addition, we were able to apply our findings to deep reinforcement learning, which had not been considered much at first, and propose a completely new deep reinforcement learning system.

Academic Significance and Societal Importance of the Research Achievements

本研究は強化学習が必要とする空間を統計学的に大量のデータを用いて獲得するのでは無く,幾何学的なミクロな観点から獲得したという意味で学術的な意義があると考えている.また,機械学習にとってハイパーパラメータの設定は大きな問題であるが,その適応的変化や並列学習で対応できることを明らかにしたことは,学術的にも産業応用を考えた上でも意義がある.さらに,ブラックボックス最適化アルゴリズムを発展させることは複雑化する社会問題など,ありとあらゆる最適化に貢献できることを意味しているので,社会的にも大きな意義がある.

Report

(6 results)
  • 2022 Annual Research Report   Final Research Report ( PDF )
  • 2021 Research-status Report
  • 2020 Research-status Report
  • 2019 Research-status Report
  • 2018 Research-status Report
  • Research Products

    (29 results)

All 2023 2022 2021 2020 2019 2018 Other

All Journal Article (13 results) (of which Peer Reviewed: 13 results,  Open Access: 2 results) Presentation (13 results) (of which Invited: 1 results) Remarks (3 results)

  • [Journal Article] Addition of Out-of-Population Search in JADE2023

    • Author(s)
      MIYAHIRA Yuichi、IGUCHI Makishi、NOTSU Akira、HONDA Katsuhiro
    • Journal Title

      Journal of Japan Society for Fuzzy Theory and Intelligent Informatics

      Volume: 35 Issue: 1 Pages: 532-537

    • DOI

      10.3156/jsoft.35.1_532

    • ISSN
      1347-7986, 1881-7203
    • Year and Date
      2023-02-15
    • Related Report
      2022 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Deep Reinforcement Learning Combined with Approximation of Number of State Experiences2022

    • Author(s)
      M. Iguchi, A. Notsu, K. Yasunaga, S. Ubukata, K. Honda
    • Journal Title

      Proc. of 2022 International Conference on Fuzzy Theory and Its Applications

      Volume: 1

    • NAID

      130008143592

    • Related Report
      2022 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Addition of Out-of-population Search Based on the Rate of Solution Updates in JADE2022

    • Author(s)
      Y. Miyahira, A. Notsu, K. Honda
    • Journal Title

      Proc. of 2022 International Conference on Fuzzy Theory and Its Applications

      Volume: 1

    • Related Report
      2022 Annual Research Report
    • Peer Reviewed
  • [Journal Article] Additional Out-group Search for JADE2022

    • Author(s)
      Y. Miyahira, A. Notsu
    • Journal Title

      Integrated Uncertainty in Knowledge Modelling and Decision Making

      Volume: 9 Pages: 105-116

    • Related Report
      2021 Research-status Report
    • Peer Reviewed
  • [Journal Article] A Study on Pre-Learning of State Similarity for Deep Reinforcement Learning2021

    • Author(s)
      K. Yasunaga, A. Notsu, S. Ubukata, K. Honda
    • Journal Title

      Proc. of 22nd International Symposium on Advanced Intelligent Systems

      Volume: G01-2 Pages: 7-16

    • NAID

      40022262669

    • Related Report
      2021 Research-status Report
    • Peer Reviewed
  • [Journal Article] Online state space generation by a growing self-organizing map and differential learning for reinforcement learning2020

    • Author(s)
      A. Notsu, K. Yasuda, S. Ubukata, K. Honda
    • Journal Title

      Applied Soft Computing

      Volume: 97 Pages: 1-9

    • DOI

      10.1016/j.asoc.2020.106723

    • Related Report
      2020 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Proposal of Adaptive Randomness in Differential Evolution2020

    • Author(s)
      J. Tsubamoto, A. Notsu, S. Ubukata, K. Honda
    • Journal Title

      Proc. of 2020 IEEE Congress on Evolutionary Computation

      Volume: 1 Pages: 1-8

    • NAID

      130007957968

    • Related Report
      2020 Research-status Report
    • Peer Reviewed
  • [Journal Article] Randomness Selection in Differential Evolution Using Thompson Sampling2020

    • Author(s)
      A. Notsu, J. Tsubamoto, Y. Miyahira, S. Ubukata, K. Honda
    • Journal Title

      Proc. of Joint 11th International Conference on Soft Computing and Intelligent Systems and 21st International Symposium on Advanced Intelligent Systems

      Volume: 1 Pages: 351-355

    • Related Report
      2020 Research-status Report
    • Peer Reviewed
  • [Journal Article] Designation of Candidate Solutions in Differential Evolution Based on Bandit Algorithm and its Evaluation2019

    • Author(s)
      M. Sakakibara, A. Notsu, S. Ubukata, K. Honda
    • Journal Title

      Journal of Advanced Computational Intelligence and Intelligent Informatics

      Volume: 23 Issue: 4 Pages: 758-766

    • DOI

      10.20965/jaciii.2019.p0758

    • NAID

      130007681855

    • ISSN
      1343-0130, 1883-8014
    • Year and Date
      2019-07-20
    • Related Report
      2019 Research-status Report
    • Peer Reviewed / Open Access
  • [Journal Article] Simple Pre-learning of State Similarity for Deep Reinforcement Learning2019

    • Author(s)
      A. Notsu, K. Yasuda, K. Yasunaga, S. Ubukata, K. Honda
    • Journal Title

      Proc. of the 20th International Symposium on Advanced Intelligent Systems and 2019 International Conference on Biometrics and Kansei Engineering

      Volume: 1 Pages: 63-68

    • Related Report
      2019 Research-status Report
    • Peer Reviewed
  • [Journal Article] Changes during the Search from Differential Evolution with Intervals to Other Methods2019

    • Author(s)
      J. Tsubamoto, A. Notsu, S. Ubukata, K. Honda
    • Journal Title

      Proc. of the 20th International Symposium on Advanced Intelligent Systems and 2019 International Conference on Biometrics and Kansei Engineering

      Volume: 1 Pages: 75-80

    • Related Report
      2019 Research-status Report
    • Peer Reviewed
  • [Journal Article] Optimization of Learning Cycles in Online Reinforcement Learning Systems2018

    • Author(s)
      A. Notsu, K. Yasuda, S. Ubukata, K. Honda
    • Journal Title

      Proc. of 2018 IEEE International Conference on Systems, Man, and Cybernetics

      Volume: #12428 Pages: 3520-3524

    • Related Report
      2018 Research-status Report
    • Peer Reviewed
  • [Journal Article] Setting of Candidate Solutions Considering Confidence Intervals in Differential Evolution2018

    • Author(s)
      A. Notsu, M. Sakakibara, S. Ubukata, K. Honda
    • Journal Title

      Proc. of 2018 International Conference on Fuzzy Theory and Its Applications

      Volume: #T1a-2 Pages: 7-11

    • NAID

      130007554574

    • Related Report
      2018 Research-status Report
    • Peer Reviewed
  • [Presentation] JADEにおける解の更新割合に基づいた集団外探索の追加2022

    • Author(s)
      宮平 裕一, 野津 亮, 本多 克宏
    • Organizer
      日本知能情報ファジィ学会
    • Related Report
      2022 Annual Research Report
  • [Presentation] 差分進化におけるパラメータのバンディットアルゴリズムによる適応的選択2021

    • Author(s)
      宮平 裕一,野津 亮,本多 克宏,生方 誠希
    • Organizer
      第65回システム制御情報学会研究発表講演会
    • Related Report
      2021 Research-status Report
  • [Presentation] 状態経験数の近似を併用した深層強化学習2021

    • Author(s)
      安永 恭平, 野津 亮, 生方 誠希, 本多 克宏
    • Organizer
      第37回ファジィシステムシンポジウム
    • Related Report
      2021 Research-status Report
  • [Presentation] JADEに対する集団外探索の追加2021

    • Author(s)
      宮平 裕一, 野津 亮, 生方 誠希, 本多 克宏
    • Organizer
      インテリジェント・システム・シンポジウム2021
    • Related Report
      2021 Research-status Report
  • [Presentation] 低コストな進化計算や強化学習のアルゴリズムの提案に向けて2021

    • Author(s)
      野津 亮
    • Organizer
      インテリジェント・システム・シンポジウム2021
    • Related Report
      2021 Research-status Report
    • Invited
  • [Presentation] 深層強化学習のための状態類似度の事前学習についての一考察2020

    • Author(s)
      安永 恭平,野津 亮,生方 誠希,本多 克宏
    • Organizer
      第64回システム制御情報学会研究発表講演会
    • Related Report
      2020 Research-status Report
  • [Presentation] ランダムネス適応型差分進化の提案2020

    • Author(s)
      鍔本 純也,野津 亮,生方 誠希,本多 克宏
    • Organizer
      第36回ファジィシステムシンポジウム
    • Related Report
      2020 Research-status Report
  • [Presentation] 区間を考慮した差分進化から他手法への探索途中での変更2019

    • Author(s)
      鍔本 純也,野津 亮,生方 誠希,本多 克宏
    • Organizer
      第63回システム制御情報学会研究発表講演会
    • Related Report
      2019 Research-status Report
  • [Presentation] 差分進化における探索点群の広がりとアルゴリズムの切り替え2019

    • Author(s)
      野津 亮,鍔本 純也,生方 誠希,本多 克宏
    • Organizer
      第35回ファジィシステムシンポジウム
    • Related Report
      2019 Research-status Report
  • [Presentation] 深層強化学習のための状態類似性のシンプルな事前学習2019

    • Author(s)
      野津 亮,安田 功嗣,安永 恭平
    • Organizer
      第29回インテリジェント・システム・シンポジウム
    • Related Report
      2019 Research-status Report
  • [Presentation] 成長型自己組織化マップによる強化学習システムについての考察2018

    • Author(s)
      安田 功嗣,野津 亮,生方 誠希,本多 克宏
    • Organizer
      第62回システム制御情報学会研究発表講演会
    • Related Report
      2018 Research-status Report
  • [Presentation] 差分進化における信頼区間を考慮した解候補の設定2018

    • Author(s)
      野津 亮,榊原 雅也,生方 誠希,本多 克宏
    • Organizer
      第34回ファジィシステムシンポジウム
    • Related Report
      2018 Research-status Report
  • [Presentation] 強化学習システムにおける学習周期の無作為抽出による適応2018

    • Author(s)
      野津 亮,安田 功嗣,生方 誠希,本多 克宏
    • Organizer
      第28回インテリジェント・システム・シンポジウム
    • Related Report
      2018 Research-status Report
  • [Remarks] 人間情報システム研究グループ

    • URL

      http://www.cs.osakafu-u.ac.jp/hi/index.html

    • Related Report
      2020 Research-status Report
  • [Remarks] 人間情報システム研究グループホームページ

    • URL

      http://www.cs.osakafu-u.ac.jp/hi/publications.html

    • Related Report
      2019 Research-status Report
  • [Remarks] http://www.cs.osakafu-u.ac.jp/hi/publications.html

    • Related Report
      2018 Research-status Report

URL: 

Published: 2018-04-23   Modified: 2024-01-30  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi