Date of Award

5-2021

Document Type

Thesis open access

Department

Computer Science

First Advisor

Mark Lewis

Second Advisor

Matthew Hibbs

Abstract

Wave Function Collapse (WFC) is an image-based Procedural Content Generation (PCG) algorithm that uses constraints extracted from an input image to generate a similar, yet novel output. The goal of this thesis is to modify WFC with extra constraints and parameters that would allow a game designer to have more control over the algorithm to produce more varied and specific results. This thesis introduces the field of Procedural Content Generation (of which Wave Function Collapse is a part) and details the WFC algorithm. It then examines past work done in these two topics, along with developments in the field of content generation for the game Super Mario Bros. and developments of heuristics used to analyze PCG content. The thesis next explains the specifics of WFC’s core algorithm, introduces the new modifications made to it, and details heuristics used to analyze its output. Finally, experiments are run using these modifications to generate content using levels from the original Super Mario Bros. as input, and previously introduced heuristics are used to assess the results.

Download

Share

COinS