Multiobjective exploration of the StarCraft map space

Julian Togelius, Mike Preuss, Nicola Beume, Simon Wessing, Johan Hagelbäck, Georgios N. Yannakakis

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

    Abstract

    This paper presents a search-based method for generating maps for the popular real-time strategy (RTS) game StarCraft. We devise a representation of StarCraft maps suitable for evolutionary search, along with a set of fitness functions based on predicted entertainment value of those maps, as derived from theories of player experience. A multiobjective evolutionary algorithm is then used to evolve complete StarCraft maps based on the representation and selected fitness functions. The output of this algorithm is a Pareto front approximation visualizing the tradeoff between the several fitness functions used, and where each point on the front represents a viable map. We argue that this method is useful for both automatic and machine-assisted map generation, and in particular that the Pareto fronts are excellent design support tools for human map designers.

    Original languageEnglish (US)
    Title of host publicationProceedings of the 2010 IEEE Conference on Computational Intelligence and Games, CIG2010
    Pages265-272
    Number of pages8
    DOIs
    StatePublished - 2010
    Event2010 IEEE Conference on Computational Intelligence and Games, CIG2010 - Copenhagen, Denmark
    Duration: Aug 18 2010Aug 21 2010

    Other

    Other2010 IEEE Conference on Computational Intelligence and Games, CIG2010
    CountryDenmark
    CityCopenhagen
    Period8/18/108/21/10

    Fingerprint

    Evolutionary algorithms

    Keywords

    • evolutionary multiobjective optimization
    • procedural content generation
    • Real-time strategy games
    • RTS

    ASJC Scopus subject areas

    • Artificial Intelligence
    • Computer Graphics and Computer-Aided Design
    • Software

    Cite this

    Togelius, J., Preuss, M., Beume, N., Wessing, S., Hagelbäck, J., & Yannakakis, G. N. (2010). Multiobjective exploration of the StarCraft map space. In Proceedings of the 2010 IEEE Conference on Computational Intelligence and Games, CIG2010 (pp. 265-272). [5593346] https://doi.org/10.1109/ITW.2010.5593346

    Multiobjective exploration of the StarCraft map space. / Togelius, Julian; Preuss, Mike; Beume, Nicola; Wessing, Simon; Hagelbäck, Johan; Yannakakis, Georgios N.

    Proceedings of the 2010 IEEE Conference on Computational Intelligence and Games, CIG2010. 2010. p. 265-272 5593346.

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

    Togelius, J, Preuss, M, Beume, N, Wessing, S, Hagelbäck, J & Yannakakis, GN 2010, Multiobjective exploration of the StarCraft map space. in Proceedings of the 2010 IEEE Conference on Computational Intelligence and Games, CIG2010., 5593346, pp. 265-272, 2010 IEEE Conference on Computational Intelligence and Games, CIG2010, Copenhagen, Denmark, 8/18/10. https://doi.org/10.1109/ITW.2010.5593346
    Togelius J, Preuss M, Beume N, Wessing S, Hagelbäck J, Yannakakis GN. Multiobjective exploration of the StarCraft map space. In Proceedings of the 2010 IEEE Conference on Computational Intelligence and Games, CIG2010. 2010. p. 265-272. 5593346 https://doi.org/10.1109/ITW.2010.5593346
    Togelius, Julian ; Preuss, Mike ; Beume, Nicola ; Wessing, Simon ; Hagelbäck, Johan ; Yannakakis, Georgios N. / Multiobjective exploration of the StarCraft map space. Proceedings of the 2010 IEEE Conference on Computational Intelligence and Games, CIG2010. 2010. pp. 265-272
    @inproceedings{c076eefb472a43ccbc05599ae35aec66,
    title = "Multiobjective exploration of the StarCraft map space",
    abstract = "This paper presents a search-based method for generating maps for the popular real-time strategy (RTS) game StarCraft. We devise a representation of StarCraft maps suitable for evolutionary search, along with a set of fitness functions based on predicted entertainment value of those maps, as derived from theories of player experience. A multiobjective evolutionary algorithm is then used to evolve complete StarCraft maps based on the representation and selected fitness functions. The output of this algorithm is a Pareto front approximation visualizing the tradeoff between the several fitness functions used, and where each point on the front represents a viable map. We argue that this method is useful for both automatic and machine-assisted map generation, and in particular that the Pareto fronts are excellent design support tools for human map designers.",
    keywords = "evolutionary multiobjective optimization, procedural content generation, Real-time strategy games, RTS",
    author = "Julian Togelius and Mike Preuss and Nicola Beume and Simon Wessing and Johan Hagelb{\"a}ck and Yannakakis, {Georgios N.}",
    year = "2010",
    doi = "10.1109/ITW.2010.5593346",
    language = "English (US)",
    isbn = "9781424462971",
    pages = "265--272",
    booktitle = "Proceedings of the 2010 IEEE Conference on Computational Intelligence and Games, CIG2010",

    }

    TY - GEN

    T1 - Multiobjective exploration of the StarCraft map space

    AU - Togelius, Julian

    AU - Preuss, Mike

    AU - Beume, Nicola

    AU - Wessing, Simon

    AU - Hagelbäck, Johan

    AU - Yannakakis, Georgios N.

    PY - 2010

    Y1 - 2010

    N2 - This paper presents a search-based method for generating maps for the popular real-time strategy (RTS) game StarCraft. We devise a representation of StarCraft maps suitable for evolutionary search, along with a set of fitness functions based on predicted entertainment value of those maps, as derived from theories of player experience. A multiobjective evolutionary algorithm is then used to evolve complete StarCraft maps based on the representation and selected fitness functions. The output of this algorithm is a Pareto front approximation visualizing the tradeoff between the several fitness functions used, and where each point on the front represents a viable map. We argue that this method is useful for both automatic and machine-assisted map generation, and in particular that the Pareto fronts are excellent design support tools for human map designers.

    AB - This paper presents a search-based method for generating maps for the popular real-time strategy (RTS) game StarCraft. We devise a representation of StarCraft maps suitable for evolutionary search, along with a set of fitness functions based on predicted entertainment value of those maps, as derived from theories of player experience. A multiobjective evolutionary algorithm is then used to evolve complete StarCraft maps based on the representation and selected fitness functions. The output of this algorithm is a Pareto front approximation visualizing the tradeoff between the several fitness functions used, and where each point on the front represents a viable map. We argue that this method is useful for both automatic and machine-assisted map generation, and in particular that the Pareto fronts are excellent design support tools for human map designers.

    KW - evolutionary multiobjective optimization

    KW - procedural content generation

    KW - Real-time strategy games

    KW - RTS

    UR - http://www.scopus.com/inward/record.url?scp=79955815556&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=79955815556&partnerID=8YFLogxK

    U2 - 10.1109/ITW.2010.5593346

    DO - 10.1109/ITW.2010.5593346

    M3 - Conference contribution

    SN - 9781424462971

    SP - 265

    EP - 272

    BT - Proceedings of the 2010 IEEE Conference on Computational Intelligence and Games, CIG2010

    ER -