Procedural Content Generation using Neuroevolution and Novelty Search for Diverse Video Game Levels

TL;DR

This paper presents a novel evolutionary approach combining neuroevolution and novelty search to generate diverse, high-quality video game levels efficiently in real time, without training data or domain knowledge.

Contribution

It introduces a new method that significantly speeds up level generation and generalizes to different level sizes without retraining.

Findings

Achieves an order of magnitude faster generation than existing methods.

Produces levels with comparable quality and diversity metrics.

Successfully generalizes to arbitrary level sizes without retraining.

Abstract

Procedurally generated video game content has the potential to drastically reduce the content creation budget of game developers and large studios. However, adoption is hindered by limitations such as slow generation, as well as low quality and diversity of content. We introduce an evolutionary search-based approach for evolving level generators using novelty search to procedurally generate diverse levels in real time, without requiring training data or detailed domain-specific knowledge. We test our method on two domains, and our results show an order of magnitude speedup in generation time compared to existing methods while obtaining comparable metric scores. We further demonstrate the ability to generalise to arbitrary-sized levels without retraining.