Troubleshooting Under Uncertainty

We develop a series of approximations for decision-theoretic troubleshooting under uncertainty. Our approach generates troubleshooting plans in the face of uncertainty in the relationships among components and device status, observations, as well as the affect of actions on device status. Included in our approach is a Bayesian-network representation of these relationships. We have applied our technique successfully to troubleshooting problems with printing, photocopier feeders, automobiles, and gas turbines. We report empirical findings demonstrating the high quality of plans produced by our approach. 1 Introduction The purpose of troubleshooting is to generate a low cost plan for the repair of a device, where a plan consists of observations and component-repair actions. 1 In this paper, we develop a series of approximations for decision-theoretic troubleshooting under uncertainty. Our approach generates troubleshooting plans in the face of uncertainty in (1) the relationships among...