Australian Computer Society Inc.
Faculty of Computing, Health and Science
School of Computer and Security Science / Artificial Intelligence and Optimisation Research Group
Despite games often being used as a testbed for new computational intelligence techniques, the majority of artificial intelligence in commercial games is scripted. This means that the computer agents are non-adaptive and often inherently exploitable because of it. In this paper, we describe a learning system designed for team strategy development in a real time multi-agent domain. We test our system in a prey and predators domain, evolving adaptive team strategies for the predators in real time against a single prey opponent. Our learning system works by continually training and updating the predator strategies, one at a time for a designated length of time while the game us being played. We test the performance of the system for real-time learning of strategies in the prey and predators domain against a hand-coded prey opponent. We show that the resulting real-time team strategies are able to capture hand-coded prey of varying degrees of difficulty without any prior learning. The system is highly adaptive to change, capable of handling many different situations, and quickly learning to function in situations that it has never seen before.