MAS4Games: A Reinforced Learning-Based Multi-agent System to Improve Player Retention in Virtual Reality Video Games

Maury-Castañeda N.; Villarruel-Vasquez S.; Ugarte W. · Communications in Computer and Information Science (2023) · DOI: 10.1007/978-3-031-49368-3_7

Resumen

In this paper, we present a Q-learning-based multi-agent system designed for Dynamic Difficulty Adjustment (DDA) in a 3D fighting game. Our primary goal is to enhance the player’s gaming experience through dynamic game difficulty adjustments based on their performance. We leverage the Unity game development platform and the ML-Agents framework to implement the Q-learning algorithm, training intelligent agents to adapt the game’s difficulty. Our findings underscore the efficacy of Q-learning and multi-agent systems in improving DDA for video games. In the conclusion section, we discuss potential implications and future directions for our research.

Instituciones: Universidad Peruana de Ciencias Aplicadas