The special structure of a class of Markovian decision problems is exploited to simplify the determination of optimum policies. For certain pairs consisting of state [i] and decision [k], the immediate expected cost separates into one component that...
