Decision-Theoretic: troubleshooting : Real-World Applications of Bayesian Networks

We have described a decision-theoretic approach for generating troubleshooting plans under uncertainty that interleaves both observations and repair actions. Our approach is based on a set of approximations to an exact method for a simple special case. Despite our approximations, we have seen that our planner produces expected troubleshooting costs that are close to optimal and significantly lower than those of simple planners