Date of Award:


Document Type:


Degree Name:

Master of Science (MS)


Computer Science

Committee Chair(s)

Mario Harper


Mario Harper


Dean Mathias


Steve Petruzza


This thesis introduces Acacia, a game engine with built-in artificial intelligence (AI) capabilities. Acacia allows game developers to effortlessly incorporate Reinforcement Learning (RL) algorithms into their creations. By tagging game elements to convey information about the game state or rewards, developers gain precise control over how RL algorithms interact with their games, mirroring real player behavior or providing full knowledge of the game world.

To showcase Acacia’s versatility, the thesis presents three games across different genres, each demonstrating the engine’s AI plugin. The goal is to establish Acacia as a preferred resource for creating 2D games with RL support without confining developers to specific machine-learning libraries, ensuring adaptability to current and emerging frameworks.

A key advantage of Acacia is its flexibility in deployment, addressing common challenges encountered in other simulators like Unity. Unlike Unity’s confined ML library, Acacia enables developers to deploy trained models beyond the simulation environment, even onto real-world objects such as robots.

Moreover, Acacia’s architecture facilitates efficient training by enabling multiple instances of the same virtual world and smart rendering optimization. This strategic approach accelerates training times, allowing for hundreds of hours of simulation within a fraction of the usual timeframe, significantly enhancing AI model development efficiency within the game engine.



Creative Commons License

Creative Commons Attribution 4.0 License
This work is licensed under a Creative Commons Attribution 4.0 License.