Tuesday, September 07, 2010

Spread Through a Network

As a sort of quick follow-up to yesterday's post. I've re-edited the graph from last time to demonstrate how a 'virus' might spread through that network.

For this, we assume a 0.5 contact rate - that is, a person in contact with an infective has a 50:50 chance of becoming infected in the next 'turn'. An Infective will recover after one turn.

In the diagrams, then, red circle means infected; orange circle means exposed and able to catch the infection; blue circle means recovered and not able to catch the infection again. White circles are susceptible to infection, but not exposed to it.

So in this case, that all important central connector gets infected (eventually) and as a result, 73% of the total network gets infected.

If it hadn't, only 40% of the network would've been infected - 53% if all the nodes on that side were infected.

Of course, this is only one possible course for the infection. And if the contact rate were lower, it probably wouldn't have spread through as much of the network.

Anyway. It's just a nice way of showing how modelling epidemics on an asymmetrical network works.

Good.


Oatzy.

No comments: