Answer Set Programming for PCG: the Good, the Bad, and the Ugly

Ian Horswill*

*Corresponding author for this work

Research output: Contribution to journalConference articlepeer-review

1 Scopus citations

Abstract

Declarative languages allow designers to build procedural content generation systems without having to design and debug specialized generation algorithms. Instead, the designer describes the desired properties of the objects to be generated, and a general-purpose constraint-solver constructs the desired artifact. Answer-Set Prolog (Gebser et al., 2012; Lifschitz, 2008b) is a popular family of languages and solvers used in procedural content generation research. Answer set programming is very powerful, with mature implementations and a significant user base outside the PCG community. However, ASP uses stable-model semantics (Gelfond & Lifschitz, 1992), which is subtle and difficult. In this paper, I will present some of the history and motivation underlying stable model semantics in as non-technical manner as I can manage, and discuss its advantages and disadvantages. I will argue that while it is appropriate for some very difficult PCG tasks, the simpler semantics of classical monotonic logic may be preferable for tasks not requiring ASP’s non-monotonicity.

Original languageEnglish (US)
JournalCEUR Workshop Proceedings
Volume3217
StatePublished - 2021
EventJoint of the Artificial Intelligence and Interactive Digital Entertainment 2021 Workshops, AIIDE-WS-2021 - Virtual, Lexington, United States
Duration: Oct 11 2021Oct 12 2021

ASJC Scopus subject areas

  • General Computer Science

Fingerprint

Dive into the research topics of 'Answer Set Programming for PCG: the Good, the Bad, and the Ugly'. Together they form a unique fingerprint.

Cite this