1: \begin{definition}

2: The SDE~(\ref{eq:SDE}) is called an \emph{affine SDE} with state

3: space $\mathcal{X}\subset\R^p$ if

4: \begin{enumerate}

5: \item it has a unique strong solution;

6: \item $\mathcal{X}$ is stochastically invariant;

7: %i.e.\ for all $x_0\in\mathcal{X}$ there exists a weak solution $X$ to

8: %(\ref{eq:SDE}) with initial condition $x_0$ such that

9: %$X_t\in\mathcal{X}$ a.s.;

10: \item the drift $\mu$ and diffusion matrix $\theta=\sigma\sigma^\top$ are affine in $x$ on $\mathcal{X}$, i.e.\

11: \begin{align}\label{eq:affinemusigma}

12: \mu(x)=a x +b,\quad\theta(x)= A^0+\sum_{i=1}^p A^i x_i,\mbox{ for

13: all $x\in\mathcal{X}$},

14: \end{align}

15: for some $a\in\R^{p\times p}$, $b\in\R^p$, symmetric

16: $A^i\in\R^{p\times p}$.

17: \end{enumerate}

18: %We call a stochastic process $X$ an \emph{affine diffusion} with

19: %state space $\mathcal{X}$ if $X_0\in\mathcal{X}$ and $X$ solves an

20: %affine SDE, which implies that $\mu(X_t)$ and

21: %$\sigma(X_t)\sigma(X_t)^\top$ are affine in $X_t$ for all

22: %$t\geq0$.

23: The unique strong solution to an affine SDE is called an affine

24: diffusion, which corresponds to the definition in

25: \cite{fm09,part1} in view of \cite[Theorem~2.5]{part1}.

26: \end{definition}