City Research Online

Abstract

We present a framework for building agents that learn using SMART, a system that combines stochastic model acquisition with reinforcement learning to enable an agent to model its environment through experience and subsequently form action selection policies using the acquired model. We extend an existing algorithm for automatic creation of stochastic strips operators [9] as a preliminary method of environment modelling. We then define the process of generation of future states using these operators and an initial state and finally show the process by which the agent can use the generated states to form a policy with a standard reinforcement learning algorithm. The potential of SMART is exemplified using the well-known predator prey scenario. Results of applying SMART to this environment and directions for future work are discussed.

Publication Type:	Article
Publisher Keywords:	Science & Technology, Technology, Computer Science, Artificial Intelligence, Computer Science
Subjects:	Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Departments:	School of Science & Technology > Computer Science

Export

Downloads