Automated Machine Learning Solutions To Gaming Applications
Hickam, Fred Quentin · 2017
This thesis project details the creation of a simple platform game and an artificially intelligent agent capable of learning to play the game. An overview of relevant concepts, including genetic algorithms, artificial neural networks, and neuroevolution is presented. The algorithm Neuroevolution of Augmenting Topologies, utilized by the agent, is discussed in detail. Gameplay mechanics and game implementation details are presented. The agent's behavior and its development over time are explained. *Includes CD.