A Note on Exact Solution of Partially Observed Markov Decision Processes

This note points out an error in the only published algorithm for exactsolution of partially observed Markov decision processes. The major insights and contributions of the original work remain intact, and the error can be remedied, albeit at increased computational expense.