= Poisson moment closure for a birth-death process
{c}
Poisson <moment> closure approximates a nonnegative integer-valued population law at each time by a <Poisson distribution> with its current <mean> $m$. Its <factorial moment>[factorial moments] then become $\mathbb E(N)_r\approx m^r$; in particular $\mathbb EN=m$ and $\mathbb E[N(N-1)]\approx m^2$.
For example, consider birth rates $b_n=an$ and death rates $d_n=bn+cn(n-1)$, with $a,b\geq0$ and $c>0$. Assuming the relevant <moments> are finite, the exact <first-moment equation of a birth-death process>[first-moment equation] is
$$
m'=(a-b)m-c\,\mathbb E[N(N-1)].
$$
Poisson closure gives the <Riccati equation> $m'\approx(a-b)m-cm^2$. The exact identity $\mathbb E[N(N-1)]=m^2+\operatorname{Var}(N)-m$ shows what is discarded: the <variance> correction $-c(\operatorname{Var}(N)-m)$. Nonlinear rates do not generally preserve a <Poisson distribution>.
For the $k=1$ <Batch-birth linear-death process>, a Poisson initial law does remain Poisson. <independent> survival thins the initial <Poisson distribution>[Poisson population], and surviving immigrants form an <independent> <Poisson distribution>[Poisson population]; their sum is Poisson. That gives an exact linear example, rather than making the nonlinear closure exact.<Lectures on the Poisson Process by Günter Last and others>[Chapter 5: marking, thinning and superposition]{ref}
Back to article page