Monitoring plan preconditions can allow for re-planning when a precondition fails, generally far in advance of the point in the plan where the precondition is relevant. However, moni-toring is generally costly, and some precondi-tion failures have a very small impact on plan quality. We formulate a model for optimal precondition monitoring, using partially-observable Markov decisions processes, and describe meth-ods for solving this model effectively, though approximately. Specifically, we show that the single-precondition monitoring problem is generally tractable, and the multiple-precondition monitoring policies can be effectively approximated using single-precondition solutions.
展开▼