Development of extensible AI systems in a video game

Abstract

In recent years, the video game market has experienced a significant growth in titles that has led to its saturation. The Metroidvania landscape is no different, which has resulted in video games with repetitive mechanics and designs, offering similar experiences with little innovation and limited chance of succeeding. This bachelor’s thesis is part of a collaborative project that presents the creation and design of a Metroidvania video game, that aims to bring uniqueness into such a crowded market. The project is undertaken by five students, each assuming a different role in the game’s development, including programming, game design, and art direction. The scope of this thesis is the development of the AI systems, including combat system, pathfinding and decision-making. The primary goal is to develop a prototype that works as a vertical slice, showing the core features of the full game and providing a representation of its mechanics and aesthetics. The specific goal in the development of the AI systems is that they are extensible so that new features can be added without hindering the development of the video game. The thesis covers the full development process, from reviewing the different techniques used to create complex behaviours and selecting which one is best suited for the project, to coding the AI systems and implementing the different types of enemies. It also outlines the collaborative development framework, with a deep study of the Metroidvania subgenre, the application of agile methodology, and the consideration of potential project costs. The results of this development can be seen in the demo release, which showcases multiple enemy types, a functioning A* pathfinding system and a modular state machine for decision-making that allows the creation of diverse of behaviours