A method to compute the transition function of a piecewise deterministic Markov process with application to reliability (Q945770)

The main contribution of the paper is to provide a closed form solution for the transition function of a picewise deterministic Markov process defined through the first order differential equation \[ \dot{Z}_t=C(Z_t,X_t), \quad Z_0=z\; , \] where \(X_t\) is an irreducible Markov process with a countable state space. It is shown that the transition function of \((Z_t,X_t)\) is governed by a Markov renewal equation. As an application, a system is considered where a failure occurs when \(Z_t\) reaches a critical level. For this problem, a numerical solution of the Markov renewal equation is developed. The numerical solution is compared with the Monte Carlo estimator of the first hitting time.