Stable fluid controller for slotted ALOHA (source code)

= Stable fluid controller for slotted ALOHA
{title2=$a=b=-A,\quad c=2A/(e-2),\quad A\geq1$}

Let $N_t$ be the backlog in <slotted ALOHA>, and let stations use attempt probability $1/S_t$ with estimate $S_t\geq1$. Update the estimate by $a$ after an idle slot, $b$ after a success and $c$ after a collision, clipping its value below at one. These feedback parameters therefore specify an actual controller, rather than only a drift formula.<Stochastic Networks by Frank Kelly and Elena Yudovina>[Section 5.2, equations 5.2–5.3, printed pp. 116–117]{ref}

In the interior <fluid model>, write $s,n>0$ and $\kappa=n/s$. The idle and success probabilities approach $e^{-\kappa}$ and $\kappa e^{-\kappa}$. For $a=b=-A$, $c=2A/(e-2)$ and $A\geq1$, their weighted mean gives
$$
\dot s=g(\kappa)=\frac{A}{e-2}[2-e(1+\kappa)e^{-\kappa}],\qquad
\dot n=\nu-\kappa e^{-\kappa}.
$$
Here $\nu$ is the mean number of arrivals per slot. Since $g(1)=0$ and $g'(\kappa)=Ae\kappa e^{-\kappa}/(e-2)>0$, the estimate falls below contention ratio one and rises above it.

Use the <ratio time change for a homogeneous fluid model>. Its ratio drift is $h=\nu-\kappa e^{-\kappa}-\kappa g$. For $\kappa\geq1$,
$$
h'=(\kappa-1)e^{-\kappa}-g-\kappa g'<0,
$$
because $g\geq0$ and $Ae\kappa^2/(e-2)>\kappa-1$. If $\nu<e^{-1}$, then $h(1)<0$; <continuity> supplies a trapping upper bound below one. There $g$ is bounded strictly below zero, giving <finite-time draining of a fluid model>. If $\nu>e^{-1}$, instead $\dot n\geq\nu-e^{-1}>0$. These are fluid conclusions; stochastic stability requires a separate argument.