Some environments or tasks may require multiple agents to work together to survive or thrive. Turn-based combat, found in many popular role-playing games, is one such environment. This paper examines a method to expose agents controlled by neural networks to both the solo and group cases in this domain, as well as difficulties presented by the transition from solo to group combat. The results suggest that, dependent on the diversity of agents' skills, agents may choose heterogeneous behaviors regardless of reward scheme.

PDF

Technical Report HR-10-09, Department of Computer Science, The University of Texas at Austin, 2010.

@techreport{johnston:ugthesis10, title={An Analysis of Distributed Decision Making Methodologies in Role Playing Video Games}, author={Matthew Johnston}, number={HR-10-09}, institution={Department of Computer Science, The University of Texas at Austin}, type={Undergraduate Honors Thesis}, url="http://nn.cs.utexas.edu/?johnston:ugthesis10", year={2010} }