Evolutionary Level Repair

2025-06-24

AI Summary

This work addresses the level repair problem: transforming procedurally generated levels—consistent in style but functionally flawed (e.g., unreachable objects, broken logic)—into playable, structurally sound versions under minimal modification constraints. We propose a novel hybrid framework integrating generative AI with search-based optimization: initial levels are produced via PCGML, then refined using evolutionary algorithms guided by Quality-Diversity (QD) search to satisfy functional constraints such as reachability and structural completeness. Our key contribution is the first application of QD to level repair, enabling robust multi-objective optimization while preserving solution diversity. Experiments across diverse failure modes demonstrate >92% repair success rate and an average edit ratio below 4.3%, significantly outperforming baseline methods. The approach is both practically effective and scalable to varied game domains and constraint configurations.

Technology Category

• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Humans and AI: Game Design — Procedural Content Generation & Storytelling
• Search and Optimization: Learning to Search

Application Category

• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Search and Retrieval-Augmented AI: Agentic search
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

We address the problem of game level repair, which consists of taking a designed but non-functional game level and making it functional. This might consist of ensuring the completeness of the level, reachability of objects, or other performance characteristics. The repair problem may also be constrained in that it can only make a small number of changes to the level. We investigate search-based solutions to the level repair problem, particularly using evolutionary and quality-diversity algorithms, with good results. This level repair method is applied to levels generated using a machine learning-based procedural content generation (PCGML) method that generates stylistically appropriate but frequently broken levels. This combination of PCGML for generation and search-based methods for repair shows great promise as a hybrid procedural content generation (PCG) method.

Problem

• Repair non-functional game levels for functionality
• Ensure level completeness and object reachability
• Use evolutionary algorithms for constrained level repair

Innovation

• Evolutionary algorithms for game level repair
• Quality-diversity algorithms for performance enhancement
• Hybrid PCGML and search-based repair method