A hybrid search agent in pommerman

Hongwei Zhou, Yichen Gong, Luvneesh Mugrai, Ahmed Khalifa, Andy Nealen, Julian Togelius

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Abstract

    In this paper, we explore the possibility of search-based agents in games with resource-intensive forward models. We implemented a player agent in the Pommerman framework and put it against the baseline agent to measure its performance. We implemented a heuristic agent and improved it by enabling depth-limited tree search in specific gameplay moments. We also compared different node selection methods during depth-limited tree search. Our result shows that depth-limited tree search is still viable when presented with inefficient forward models and exploitation-driven selection method is the most efficient in this specific domain.

    Original languageEnglish (US)
    Title of host publicationProceedings of the 13th International Conference on the Foundations of Digital Games, FDG 2018
    EditorsSebastian Deterding, Mitu Khandaker, Sebastian Risi, Jose Font, Steve Dahlskog, Christoph Salge, Carl Magnus Olsson
    PublisherAssociation for Computing Machinery
    ISBN (Electronic)9781450365710
    DOIs
    StatePublished - Aug 7 2018
    Event13th International Conference on the Foundations of Digital Games, FDG 2018 - Malmo, Sweden
    Duration: Aug 7 2018Aug 10 2018

    Other

    Other13th International Conference on the Foundations of Digital Games, FDG 2018
    CountrySweden
    CityMalmo
    Period8/7/188/10/18

    Keywords

    • Monte carlo methods
    • Pommerman
    • State machines
    • Tree search

    ASJC Scopus subject areas

    • Human-Computer Interaction
    • Computer Networks and Communications
    • Computer Vision and Pattern Recognition
    • Software

    Cite this

    Zhou, H., Gong, Y., Mugrai, L., Khalifa, A., Nealen, A., & Togelius, J. (2018). A hybrid search agent in pommerman. In S. Deterding, M. Khandaker, S. Risi, J. Font, S. Dahlskog, C. Salge, & C. M. Olsson (Eds.), Proceedings of the 13th International Conference on the Foundations of Digital Games, FDG 2018 [46] Association for Computing Machinery. https://doi.org/10.1145/3235765.3235812

    A hybrid search agent in pommerman. / Zhou, Hongwei; Gong, Yichen; Mugrai, Luvneesh; Khalifa, Ahmed; Nealen, Andy; Togelius, Julian.

    Proceedings of the 13th International Conference on the Foundations of Digital Games, FDG 2018. ed. / Sebastian Deterding; Mitu Khandaker; Sebastian Risi; Jose Font; Steve Dahlskog; Christoph Salge; Carl Magnus Olsson. Association for Computing Machinery, 2018. 46.

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Zhou, H, Gong, Y, Mugrai, L, Khalifa, A, Nealen, A & Togelius, J 2018, A hybrid search agent in pommerman. in S Deterding, M Khandaker, S Risi, J Font, S Dahlskog, C Salge & CM Olsson (eds), Proceedings of the 13th International Conference on the Foundations of Digital Games, FDG 2018., 46, Association for Computing Machinery, 13th International Conference on the Foundations of Digital Games, FDG 2018, Malmo, Sweden, 8/7/18. https://doi.org/10.1145/3235765.3235812
    Zhou H, Gong Y, Mugrai L, Khalifa A, Nealen A, Togelius J. A hybrid search agent in pommerman. In Deterding S, Khandaker M, Risi S, Font J, Dahlskog S, Salge C, Olsson CM, editors, Proceedings of the 13th International Conference on the Foundations of Digital Games, FDG 2018. Association for Computing Machinery. 2018. 46 https://doi.org/10.1145/3235765.3235812
    Zhou, Hongwei ; Gong, Yichen ; Mugrai, Luvneesh ; Khalifa, Ahmed ; Nealen, Andy ; Togelius, Julian. / A hybrid search agent in pommerman. Proceedings of the 13th International Conference on the Foundations of Digital Games, FDG 2018. editor / Sebastian Deterding ; Mitu Khandaker ; Sebastian Risi ; Jose Font ; Steve Dahlskog ; Christoph Salge ; Carl Magnus Olsson. Association for Computing Machinery, 2018.
    @inproceedings{db0cf84413d2414295d17e097e8e115a,
    title = "A hybrid search agent in pommerman",
    abstract = "In this paper, we explore the possibility of search-based agents in games with resource-intensive forward models. We implemented a player agent in the Pommerman framework and put it against the baseline agent to measure its performance. We implemented a heuristic agent and improved it by enabling depth-limited tree search in specific gameplay moments. We also compared different node selection methods during depth-limited tree search. Our result shows that depth-limited tree search is still viable when presented with inefficient forward models and exploitation-driven selection method is the most efficient in this specific domain.",
    keywords = "Monte carlo methods, Pommerman, State machines, Tree search",
    author = "Hongwei Zhou and Yichen Gong and Luvneesh Mugrai and Ahmed Khalifa and Andy Nealen and Julian Togelius",
    year = "2018",
    month = "8",
    day = "7",
    doi = "10.1145/3235765.3235812",
    language = "English (US)",
    editor = "Sebastian Deterding and Mitu Khandaker and Sebastian Risi and Jose Font and Steve Dahlskog and Christoph Salge and Olsson, {Carl Magnus}",
    booktitle = "Proceedings of the 13th International Conference on the Foundations of Digital Games, FDG 2018",
    publisher = "Association for Computing Machinery",

    }

    TY - GEN

    T1 - A hybrid search agent in pommerman

    AU - Zhou, Hongwei

    AU - Gong, Yichen

    AU - Mugrai, Luvneesh

    AU - Khalifa, Ahmed

    AU - Nealen, Andy

    AU - Togelius, Julian

    PY - 2018/8/7

    Y1 - 2018/8/7

    N2 - In this paper, we explore the possibility of search-based agents in games with resource-intensive forward models. We implemented a player agent in the Pommerman framework and put it against the baseline agent to measure its performance. We implemented a heuristic agent and improved it by enabling depth-limited tree search in specific gameplay moments. We also compared different node selection methods during depth-limited tree search. Our result shows that depth-limited tree search is still viable when presented with inefficient forward models and exploitation-driven selection method is the most efficient in this specific domain.

    AB - In this paper, we explore the possibility of search-based agents in games with resource-intensive forward models. We implemented a player agent in the Pommerman framework and put it against the baseline agent to measure its performance. We implemented a heuristic agent and improved it by enabling depth-limited tree search in specific gameplay moments. We also compared different node selection methods during depth-limited tree search. Our result shows that depth-limited tree search is still viable when presented with inefficient forward models and exploitation-driven selection method is the most efficient in this specific domain.

    KW - Monte carlo methods

    KW - Pommerman

    KW - State machines

    KW - Tree search

    UR - http://www.scopus.com/inward/record.url?scp=85055591663&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=85055591663&partnerID=8YFLogxK

    U2 - 10.1145/3235765.3235812

    DO - 10.1145/3235765.3235812

    M3 - Conference contribution

    BT - Proceedings of the 13th International Conference on the Foundations of Digital Games, FDG 2018

    A2 - Deterding, Sebastian

    A2 - Khandaker, Mitu

    A2 - Risi, Sebastian

    A2 - Font, Jose

    A2 - Dahlskog, Steve

    A2 - Salge, Christoph

    A2 - Olsson, Carl Magnus

    PB - Association for Computing Machinery

    ER -