Thesiss

The classical optimal control problems for discrete-time, transient Markov processes are infinite horizon, undiscounted expected total cost or reward models. Some examples of these models are optimal stopping problems and stochastic shortest or longest path problems, which may h…