Enhancements to constrained novelty search two-population novelty search for generating game content

Antonios Liapis, Georgios N. Yannakakis, Julian Togelius

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

    Abstract

    Novelty search is a recent algorithm geared to explore search spaces without regard to objectives; minimal criteria novelty search is a variant of this algorithm for constrained search spaces. For large search spaces with multiple constraints, however, it is hard to find a set of feasible individuals that is both large and diverse. In this paper, we present two new methods of novelty search for constrained spaces, Feasible-Infeasible Novelty Search and Feasible-Infeasible Dual Novelty Search. Both algorithms keep separate populations of feasible and infeasible individuals, inspired by the FI-2pop genetic algorithm. These algorithms are applied to the problem of creating diverse and feasible game levels, representative of a large class of important problems in procedural content generation for games. Results show that the new algorithms under certain conditions can produce larger and more diverse sets of feasible strategy game maps than existing algorithms. However, the best algorithm is contingent on the particularities of the search space and the genetic operators used. It is also shown that the proposed enhancement of offspring boosting increases performance in all cases.

    Original languageEnglish (US)
    Title of host publicationGECCO 2013 - Proceedings of the 2013 Genetic and Evolutionary Computation Conference
    Pages343-350
    Number of pages8
    DOIs
    StatePublished - 2013
    Event2013 15th Genetic and Evolutionary Computation Conference, GECCO 2013 - Amsterdam, Netherlands
    Duration: Jul 6 2013Jul 10 2013

    Other

    Other2013 15th Genetic and Evolutionary Computation Conference, GECCO 2013
    CountryNetherlands
    CityAmsterdam
    Period7/6/137/10/13

    Fingerprint

    Enhancement
    Game
    Search Space
    Population
    Genetic Operators
    Boosting
    Genetic algorithms
    Genetic Algorithm

    Keywords

    • Constrained novelty search
    • Feasible-Infeasible two-population GA
    • Level design
    • Procedural content generation

    ASJC Scopus subject areas

    • Genetics
    • Computational Mathematics

    Cite this

    Liapis, A., Yannakakis, G. N., & Togelius, J. (2013). Enhancements to constrained novelty search two-population novelty search for generating game content. In GECCO 2013 - Proceedings of the 2013 Genetic and Evolutionary Computation Conference (pp. 343-350) https://doi.org/10.1145/2463372.2463416

    Enhancements to constrained novelty search two-population novelty search for generating game content. / Liapis, Antonios; Yannakakis, Georgios N.; Togelius, Julian.

    GECCO 2013 - Proceedings of the 2013 Genetic and Evolutionary Computation Conference. 2013. p. 343-350.

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

    Liapis, A, Yannakakis, GN & Togelius, J 2013, Enhancements to constrained novelty search two-population novelty search for generating game content. in GECCO 2013 - Proceedings of the 2013 Genetic and Evolutionary Computation Conference. pp. 343-350, 2013 15th Genetic and Evolutionary Computation Conference, GECCO 2013, Amsterdam, Netherlands, 7/6/13. https://doi.org/10.1145/2463372.2463416
    Liapis A, Yannakakis GN, Togelius J. Enhancements to constrained novelty search two-population novelty search for generating game content. In GECCO 2013 - Proceedings of the 2013 Genetic and Evolutionary Computation Conference. 2013. p. 343-350 https://doi.org/10.1145/2463372.2463416
    Liapis, Antonios ; Yannakakis, Georgios N. ; Togelius, Julian. / Enhancements to constrained novelty search two-population novelty search for generating game content. GECCO 2013 - Proceedings of the 2013 Genetic and Evolutionary Computation Conference. 2013. pp. 343-350
    @inproceedings{11b788ecf8d34e91a7a3b6bd1db51a68,
    title = "Enhancements to constrained novelty search two-population novelty search for generating game content",
    abstract = "Novelty search is a recent algorithm geared to explore search spaces without regard to objectives; minimal criteria novelty search is a variant of this algorithm for constrained search spaces. For large search spaces with multiple constraints, however, it is hard to find a set of feasible individuals that is both large and diverse. In this paper, we present two new methods of novelty search for constrained spaces, Feasible-Infeasible Novelty Search and Feasible-Infeasible Dual Novelty Search. Both algorithms keep separate populations of feasible and infeasible individuals, inspired by the FI-2pop genetic algorithm. These algorithms are applied to the problem of creating diverse and feasible game levels, representative of a large class of important problems in procedural content generation for games. Results show that the new algorithms under certain conditions can produce larger and more diverse sets of feasible strategy game maps than existing algorithms. However, the best algorithm is contingent on the particularities of the search space and the genetic operators used. It is also shown that the proposed enhancement of offspring boosting increases performance in all cases.",
    keywords = "Constrained novelty search, Feasible-Infeasible two-population GA, Level design, Procedural content generation",
    author = "Antonios Liapis and Yannakakis, {Georgios N.} and Julian Togelius",
    year = "2013",
    doi = "10.1145/2463372.2463416",
    language = "English (US)",
    isbn = "9781450319638",
    pages = "343--350",
    booktitle = "GECCO 2013 - Proceedings of the 2013 Genetic and Evolutionary Computation Conference",

    }

    TY - GEN

    T1 - Enhancements to constrained novelty search two-population novelty search for generating game content

    AU - Liapis, Antonios

    AU - Yannakakis, Georgios N.

    AU - Togelius, Julian

    PY - 2013

    Y1 - 2013

    N2 - Novelty search is a recent algorithm geared to explore search spaces without regard to objectives; minimal criteria novelty search is a variant of this algorithm for constrained search spaces. For large search spaces with multiple constraints, however, it is hard to find a set of feasible individuals that is both large and diverse. In this paper, we present two new methods of novelty search for constrained spaces, Feasible-Infeasible Novelty Search and Feasible-Infeasible Dual Novelty Search. Both algorithms keep separate populations of feasible and infeasible individuals, inspired by the FI-2pop genetic algorithm. These algorithms are applied to the problem of creating diverse and feasible game levels, representative of a large class of important problems in procedural content generation for games. Results show that the new algorithms under certain conditions can produce larger and more diverse sets of feasible strategy game maps than existing algorithms. However, the best algorithm is contingent on the particularities of the search space and the genetic operators used. It is also shown that the proposed enhancement of offspring boosting increases performance in all cases.

    AB - Novelty search is a recent algorithm geared to explore search spaces without regard to objectives; minimal criteria novelty search is a variant of this algorithm for constrained search spaces. For large search spaces with multiple constraints, however, it is hard to find a set of feasible individuals that is both large and diverse. In this paper, we present two new methods of novelty search for constrained spaces, Feasible-Infeasible Novelty Search and Feasible-Infeasible Dual Novelty Search. Both algorithms keep separate populations of feasible and infeasible individuals, inspired by the FI-2pop genetic algorithm. These algorithms are applied to the problem of creating diverse and feasible game levels, representative of a large class of important problems in procedural content generation for games. Results show that the new algorithms under certain conditions can produce larger and more diverse sets of feasible strategy game maps than existing algorithms. However, the best algorithm is contingent on the particularities of the search space and the genetic operators used. It is also shown that the proposed enhancement of offspring boosting increases performance in all cases.

    KW - Constrained novelty search

    KW - Feasible-Infeasible two-population GA

    KW - Level design

    KW - Procedural content generation

    UR - http://www.scopus.com/inward/record.url?scp=84883056025&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=84883056025&partnerID=8YFLogxK

    U2 - 10.1145/2463372.2463416

    DO - 10.1145/2463372.2463416

    M3 - Conference contribution

    AN - SCOPUS:84883056025

    SN - 9781450319638

    SP - 343

    EP - 350

    BT - GECCO 2013 - Proceedings of the 2013 Genetic and Evolutionary Computation Conference

    ER -