Continuous-time, constant causative Markov chains (Q1098175)

The paper extends the definition of constant causative Markov chain in discrete time to the case of continuous time. Recall that a discrete-time constant causative Markov chain is a nonhomogeneous Markov chain with one-step transition matrices satisfying \(P(n,n+1)=P(n-1,n)C\) for some matrix C (called causative). Generalizing this definition to continuous- time processes, the author finds that a constant causative Markov chain in continuous time must have intensity matrices of the form \(Q(t)=tC+Q\), where C and Q are intensity matrices. Next, he considers long-run properties and derives a sufficient condition for ergodicity, which extends earlier results for discrete time. Finally, he shows that in the case when Q and C commute, then the irreducibility and ergodicity of the chain can be directly related to that of two corresponding discrete time, homogeneous chains.