||Zhang, Nevin Lianwen
||Partially observable decision processes (POMDP) can be used as a model for planning in stochastic domains. This paper considers the problem of computing an optimal policy for a finite horizon POMDP. The task is difficult because the decision at any time point depends upon information from all previous time points. We propose to filter out inconsistencies and insignificant details in the collection of information being passed from one time point to the next. This reduces the number of possible information states and hence speeds up computation. A bound on the sacrifice of optimality due to information filtering is given, which provides a way for rading off between computational complexity and optimality.
Files in this item: