Offline planning with hierarchical task networks in video games

John Paul Kelly, Adi Botea, Sven Koenig

    Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

    29 Citations (Scopus)

    Abstract

    Artificial intelligence (AI) technology can have a dramatic impact on the quality of video games. AI planning techniques are useful in a wide range of game components, including modules that control the behavior of fully autonomous units. However, planning is computationally expensive, and the CPU and memory resources available to game AI modules at runtime are scarce. Offline planning can be a good strategy to avoid runtime performance bottlenecks. In this work, we apply hierarchical task network (HTN) planning to video games. We describe a system that computes plans offline and then represents them as game scripts. This can be seen as a form of generating game scripts automatically, replacing the traditional approach of composing them by hand. We apply our ideas to the commercial game THE ELDER SCROLLS IV: OBLIVION, with encouraging results. Our system generates scripts automatically at a level of complexity that would require a great human effort to achieve.

    Original languageEnglish
    Title of host publicationProceedings of the 4th Artificial Intelligence and Interactive Digital Entertainment Conference, AIIDE 2008
    Pages60-65
    Number of pages6
    Publication statusPublished - 2008
    Event4th Artificial Intelligence and Interactive Digital Entertainment Conference,AIIDE 2008 - Stanford, CA, United States
    Duration: 22 Oct 200824 Oct 2008

    Publication series

    NameProceedings of the 4th Artificial Intelligence and Interactive Digital Entertainment Conference, AIIDE 2008

    Conference

    Conference4th Artificial Intelligence and Interactive Digital Entertainment Conference,AIIDE 2008
    Country/TerritoryUnited States
    CityStanford, CA
    Period22/10/0824/10/08

    Fingerprint

    Dive into the research topics of 'Offline planning with hierarchical task networks in video games'. Together they form a unique fingerprint.

    Cite this