= Diffusion approximation of a birth-death process
A <birth-death process> with unit upward and downward jumps and rates $b(x)$ and $d(x)$ has <infinitesimal generator (stochastic processes)>[infinitesimal generator]
$$
Lf(x)=b(x)[f(x+1)-f(x)]+d(x)[f(x-1)-f(x)].
$$
If $f$ varies slowly over one jump, <Taylor series>[Taylor expansion] through second order gives $Lf\simeq(b-d)f'+(b+d)f''/2$. Thus its diffusion approximation has <drift coefficient> $u=b-d$ and, in the convention
$$
\partial_tP=-\partial_x(uP)+\partial_x^2(DP),
$$
<diffusion coefficient> $D=(b+d)/2$. Equivalently it solves $dX_t=u(X_t)\,dt+\sqrt{b(X_t)+d(X_t)}\,dB_t$, where $B$ is <Brownian motion>. The sum measures the <variance> of the short-time increment: births and deaths both add noise even when their <mean> drifts cancel.
For a population density $x=n/N$ with jump size $1/N$ and birth/death rates $N\beta(x),N\delta(x)$, the drift is $\beta-\delta$ and the <diffusion coefficient> is $(\beta+\delta)/(2N)$. Fluctuations therefore shrink as $N^{-1/2}$. A useful approximation requires many reaction <events> while the rates change little; it can fail near an <absorbing state> or when only a few individuals remain. A continuous diffusion can even visit negative values unless its boundary treatment is specified. The approximation replaces the discrete jump model rather than asserting equality of their laws.<Stochastic Simulation of Chemical Kinetics by Daniel T. Gillespie>[Section 4, printed pp. 44–46, equations (18)–(20), leap conditions and system-size scaling]{ref}
Back to article page