1: \begin{definition}[Continuous time Markov chain]	 \label{D:contMC}

2: Let $\mathcal{I}=\R_{\ge0}$. Further, let $\mat{G}\in\R^{n\times n}$ be such that

3: \begin{itemize}

4: 	\item[(i)] $\mat{G}_{ij}\ge 0$ for $i\neq j$, and

5: 	\item[(ii)] $\sum_{i=1}^n \mat{G}_{ij}\le 0$.

6: \end{itemize}

7: Define $\mat{P}^t = e^{t\mat{G}} = \sum_{k=0}^{\infty}\tfrac{t^k\mat{G}^k}{k!}$. Then $\mat{P}^t$ is a sub-stochastic matrix for every $t\ge 0$.\footnote{For a proof, see Theorem~2.1.2 in~\cite{Nor97}.} A process $\{Z^t\}_{t\in\mathcal{I}}$ with

8: \[

9: \mat{P}^t_{ji} = \prob(Z^{t+s}=j\,\vert\, Z^s=i)\quad\text{for all }t,s\ge 0,\text{ and }i,j\in \set{Y}

10: \]

11: is called a \textit{continuous time Markov chain}. It is also often called a Markov jump process (MJP). The matrix $\mat{G}$ is called the (infinitesimal) \textit{generator} of the process.

12: \end{definition}

13: