An Instance-Based State Representation for Network Repair

Michael L. Littman, Nishkam Ravi, Eitan Fenson, Richard E. Howard

Abstract: We describe a formal framework for diagnosis and repair problems that shares elements of the well known partially observable MOP and cost-sensitive classification models. Our cost-sensitive fault remediation model is amenable to implementation as a reinforcement-learning system, and we describe an instance-based state representation that is compatible with learning and planning in this framework. We demonstrate a system that uses these ideas to learn to efficiently restore network connectivity after a failure.