Optimal strategies for a replacement model

We examine a replacement system with discrete-time Markovian deterioration and finite state space $\{0,\ldots,N\}$. State 0 stands for a new system, and the higher the state the worse the system; a system in state $N$ is considered to be in a {\it bad state}. We impose the condition that the fraction of replacements in state $N$ should not be larger than some fixed number. We prove that a generalized control limit policy maximizes the expected time between two successive replacements and we explain explicitly how to derive this optimal policy. Some numerical examples are given.