Learning from experience using a decision-theoretic intelligent agent in multi-agent systems

Abstract

This paper proposes a decision-theoretic intelligent agent model to solve a herding problem and studies the learning from experience capabilities of the agent model. The proposed intelligent agent model is designed by combining Bayesian networks (BN) and influence diagrams (ID). The online Bayesian network learning method is proposed to accomplish the learning from experience. Intelligent agent software, IntelliAgent, is written to realize the proposed intelligent agent model and to simulate the agents in a problem domain. The same software is then used to simulate the herding problem with one sheep and one dog. Simulation results show that the proposed intelligent agent is successful in establishing a goal (herding) and learning other agents' behaviors.