Controllable procedural map generation via multiobjective evolution

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

    Research output: Contribution to journalArticle

    Abstract

    This paper shows how multiobjective evolutionary algorithms can be used to procedurally generate complete and playable maps for real-time strategy (RTS) games. We devise heuristic objective functions that measure properties of maps that impact important aspects of gameplay experience. To show the generality of our approach, we design two different evolvable map representations, one for an imaginary generic strategy game based on heightmaps, and one for the classic RTS game StarCraft. The effect of combining tuples or triples of the objective functions are investigated in systematic experiments, in particular which of the objectives are partially conflicting. A selection of generated maps are visually evaluated by a population of skilled StarCraft players, confirming that most of our objectives correspond to perceived gameplay qualities. Our method could be used to completely automate in-game controlled map generation, enabling player-adaptive games, or as a design support tool for human designers.

    Original languageEnglish (US)
    Pages (from-to)245-277
    Number of pages33
    JournalGenetic Programming and Evolvable Machines
    Volume14
    Issue number2
    DOIs
    StatePublished - Jun 2013

    Fingerprint

    Game
    Objective function
    Real-time
    Multi-objective Evolutionary Algorithm
    Tool Support
    Evolutionary algorithms
    Heuristics
    Experiment
    Strategy
    Experiments
    Design

    Keywords

    • Evolutionary computation
    • Multiobjective optimisation
    • Procedural content generation
    • Real-time strategy games
    • RTS
    • StarCraft

    ASJC Scopus subject areas

    • Software
    • Hardware and Architecture
    • Computer Science Applications
    • Theoretical Computer Science

    Cite this

    Togelius, J., Preuss, M., Beume, N., Wessing, S., Hagelbäck, J., Yannakakis, G. N., & Grappiolo, C. (2013). Controllable procedural map generation via multiobjective evolution. Genetic Programming and Evolvable Machines, 14(2), 245-277. https://doi.org/10.1007/s10710-012-9174-5

    Controllable procedural map generation via multiobjective evolution. / Togelius, Julian; Preuss, Mike; Beume, Nicola; Wessing, Simon; Hagelbäck, Johan; Yannakakis, Georgios N.; Grappiolo, Corrado.

    In: Genetic Programming and Evolvable Machines, Vol. 14, No. 2, 06.2013, p. 245-277.

    Research output: Contribution to journalArticle

    Togelius, J, Preuss, M, Beume, N, Wessing, S, Hagelbäck, J, Yannakakis, GN & Grappiolo, C 2013, 'Controllable procedural map generation via multiobjective evolution', Genetic Programming and Evolvable Machines, vol. 14, no. 2, pp. 245-277. https://doi.org/10.1007/s10710-012-9174-5
    Togelius J, Preuss M, Beume N, Wessing S, Hagelbäck J, Yannakakis GN et al. Controllable procedural map generation via multiobjective evolution. Genetic Programming and Evolvable Machines. 2013 Jun;14(2):245-277. https://doi.org/10.1007/s10710-012-9174-5
    Togelius, Julian ; Preuss, Mike ; Beume, Nicola ; Wessing, Simon ; Hagelbäck, Johan ; Yannakakis, Georgios N. ; Grappiolo, Corrado. / Controllable procedural map generation via multiobjective evolution. In: Genetic Programming and Evolvable Machines. 2013 ; Vol. 14, No. 2. pp. 245-277.
    @article{16c283008af444b4b87d5b420bb6810b,
    title = "Controllable procedural map generation via multiobjective evolution",
    abstract = "This paper shows how multiobjective evolutionary algorithms can be used to procedurally generate complete and playable maps for real-time strategy (RTS) games. We devise heuristic objective functions that measure properties of maps that impact important aspects of gameplay experience. To show the generality of our approach, we design two different evolvable map representations, one for an imaginary generic strategy game based on heightmaps, and one for the classic RTS game StarCraft. The effect of combining tuples or triples of the objective functions are investigated in systematic experiments, in particular which of the objectives are partially conflicting. A selection of generated maps are visually evaluated by a population of skilled StarCraft players, confirming that most of our objectives correspond to perceived gameplay qualities. Our method could be used to completely automate in-game controlled map generation, enabling player-adaptive games, or as a design support tool for human designers.",
    keywords = "Evolutionary computation, Multiobjective optimisation, Procedural content generation, Real-time strategy games, RTS, StarCraft",
    author = "Julian Togelius and Mike Preuss and Nicola Beume and Simon Wessing and Johan Hagelb{\"a}ck and Yannakakis, {Georgios N.} and Corrado Grappiolo",
    year = "2013",
    month = "6",
    doi = "10.1007/s10710-012-9174-5",
    language = "English (US)",
    volume = "14",
    pages = "245--277",
    journal = "Genetic Programming and Evolvable Machines",
    issn = "1389-2576",
    publisher = "Springer New York",
    number = "2",

    }

    TY - JOUR

    T1 - Controllable procedural map generation via multiobjective evolution

    AU - Togelius, Julian

    AU - Preuss, Mike

    AU - Beume, Nicola

    AU - Wessing, Simon

    AU - Hagelbäck, Johan

    AU - Yannakakis, Georgios N.

    AU - Grappiolo, Corrado

    PY - 2013/6

    Y1 - 2013/6

    N2 - This paper shows how multiobjective evolutionary algorithms can be used to procedurally generate complete and playable maps for real-time strategy (RTS) games. We devise heuristic objective functions that measure properties of maps that impact important aspects of gameplay experience. To show the generality of our approach, we design two different evolvable map representations, one for an imaginary generic strategy game based on heightmaps, and one for the classic RTS game StarCraft. The effect of combining tuples or triples of the objective functions are investigated in systematic experiments, in particular which of the objectives are partially conflicting. A selection of generated maps are visually evaluated by a population of skilled StarCraft players, confirming that most of our objectives correspond to perceived gameplay qualities. Our method could be used to completely automate in-game controlled map generation, enabling player-adaptive games, or as a design support tool for human designers.

    AB - This paper shows how multiobjective evolutionary algorithms can be used to procedurally generate complete and playable maps for real-time strategy (RTS) games. We devise heuristic objective functions that measure properties of maps that impact important aspects of gameplay experience. To show the generality of our approach, we design two different evolvable map representations, one for an imaginary generic strategy game based on heightmaps, and one for the classic RTS game StarCraft. The effect of combining tuples or triples of the objective functions are investigated in systematic experiments, in particular which of the objectives are partially conflicting. A selection of generated maps are visually evaluated by a population of skilled StarCraft players, confirming that most of our objectives correspond to perceived gameplay qualities. Our method could be used to completely automate in-game controlled map generation, enabling player-adaptive games, or as a design support tool for human designers.

    KW - Evolutionary computation

    KW - Multiobjective optimisation

    KW - Procedural content generation

    KW - Real-time strategy games

    KW - RTS

    KW - StarCraft

    UR - http://www.scopus.com/inward/record.url?scp=84891902031&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=84891902031&partnerID=8YFLogxK

    U2 - 10.1007/s10710-012-9174-5

    DO - 10.1007/s10710-012-9174-5

    M3 - Article

    VL - 14

    SP - 245

    EP - 277

    JO - Genetic Programming and Evolvable Machines

    JF - Genetic Programming and Evolvable Machines

    SN - 1389-2576

    IS - 2

    ER -