On artificial economy with applications to reinforcement learning

Abstract

This thesis discusses Artificial Economy with applications to Reinforcement Learning. It focuses first on Reinforcement Learning, comparing it to other types of learning such as Supervised Learning. It then covers a historical overview of the field and summarizes the current works. The work described in this thesis has a psychological approach as we focus on reward and punishment. We experiment Artificial Economy with a prototype which solves the Blocks World problem. We then show how the Artificial Economy approach can be integrated with other concepts of Reinforcement Learning techniques.