Constrained Videogame Content Generation with Answer Set Programming.
Adam M. Smith · Tiny Trans. Comput. Sci. · 2012
Automatically generating content for videogames (often called procedural content generation or PCG) requires developing systems that can synthesize artifacts that are both deeply relevant to a designer’s vision and appropriate for a player’s experience in a game so far. PCG work has historically proceeded by developing algorithms for each combination of content type and design requirements. As ill-defined requirements on generated content are refined, the design of generative procedures increasingly requires expertise in combinatorial search that exceeds any reasonable bound on a game developer’s skillset. To address this problem, recent work [3] shows how to declaratively model videogame content design spaces (specifying what to generate instead of how to generate it) as very concise answer set programs (ASP is a modern development from the logic programming community [1]). Powered by domain-independent algorithms capable of efficiently solving NP-hard problems, ASP offers a way to rapidly produce systems that are able to much more efficiently and expressively generate artifacts from tightly-constrained design spaces than a fair comparison with bespoke algorithm design by competent developers who are not experts in combinatorial search [2]. BODY Answer set programming allows the rapid development of tightly-constrained videogame content generators without inventing new algorithms.