Integration of Machine Learning in the Tower Defense subgenre of video games

Abstract

his document presents the Final Degree Work report of Iván Gomis Moreno in Bache lor’s Degree in Video Game Design and Development. The main concept is making a tower attack video game using Unity, employing the various systems that this game engine provides for the creation of non-playable characters, specifically, ML Agent Libraries. The AI implementation will possess the capability to adapt and learn based on its current conditions and interactions with the player, enabling it to respond to specific events variably, thus achieving unpredictable behaviours. Furthermore, several behaviours, actions, and systems developed for the enemy’s creation can be repurposed across various video game genres.