QNQL processes -- \((H,Q)\)-processes and their applications (Q1367233)

QNQL is an abbreviation for `quantity and quality'. The authors define the \((H,Q)\)-processes, a class of such QNQL processes. Let \((\Omega,{\mathcal F},P)\) be a complete probability space, \((E,{\mathcal E})\) a Polish space, and \(X=(x(t,\omega), 0\leq t<\tau(\omega),\omega\in\Omega)\) a right-continuous with left limits \(E\)-valued stochastic process on \((\Omega,{\mathcal L},P)\). Let \(h,q: [0,\infty)\times E\times{\mathcal E}\to [0,1]\) be measurable functions in the first two variables and quasi-distributions on \(E\) in the third variable. The process \(X\) is called an \((H, Q)\)-process if there exists an increasing sequence of Markov times \((\tau_n)_{n\geq 0}\) with \(\tau_0\equiv 0\) and \(\lim_{n\to\infty} \tau_n(\omega)= \tau(\omega)\) \(P\)-a.s. in \(\Omega\) such that \[ \begin{multlined} E(x(\tau_n+ t)\in A,\;\tau_{n+1}- \tau_n> t\mid x(\tau_j),\;\tau_j,\;0\leq j\leq n)=\\ E(x(\tau_n+ t)\in A,\;\tau_{n+1}- \tau_n>t\mid x(\tau_n))= h(t,x(\tau_n),A)\end{multlined} \] and \[ \begin{multlined} E(x(\tau_{n+1})\in A,\;\tau_{n+1}- \tau_n\leq t\mid x(\tau_j),\;\tau_j,\;0\leq j\leq n)=\\ E(x(\tau_{n+1})\in A,\;\tau_{n+1}- \tau_n\leq t\mid x(\tau_n))= q(t,x(\tau_n), A)\end{multlined} \] for all \(n\geq 0\), \(t\geq 0\), and \(A\in{\mathcal E}\). Examples of \((H,Q)\)-processes are minimal \(Q\)-processes, Doob processes, semi-Markov processes, and some processes occurring in reservoir models and queueing theory. The forward and backward equations for \((H,Q)\)-processes are considered. No proofs are given. Let us also note that the authors' English is quite peculiar.