A handful of stats students at the University of Ottawa decided to create a working mathematical model for zombie outbreaks, and possible ways of thwarting the attack. What did they learn? We're screwed. Kind of.
First, some background. They based their methodology around three groups: zombies, those susceptible to zombie attacks, and those who are unaffected (dead zombies). They based their model around zombies who infect humans with saliva via bites, and walk in slow, irregular strides. They also allowed a 24 hour incubation period from the moment of infection to complete zombification.
What did they learn? Well if left unaddressed, a zombie attack on a sizable city would wipe out the population in a matter of 4-8 hours. If you tried to quarantine the zombies, it would essentially have no effect on the outcome because the zombies would inevitably escape, or infect the humans attempting to quarantine zombies. And if you tried to generate a zombie antidote, you'd still lose a lot of people in the process of creating the antidote, and it wouldn't revert the zombies back to a dead state, which means they could possibly infect people in other areas.
The best solution? The only hope of wiping out a possible zombie invasion is to attack the undead in focused, strategic attacks that progressively increase in intensity. This will help address the growing number of undead in the process. But even then it would prove difficult to emerge victorious, as it would take 10 days worth of heavy fighting to quell the outbreak.
But luckily, you don't have to worry about any of this because some sap mathematically proved it would be impossible for zombies to exist (along with vampires). Something about how they would feast themselves into oblivion. [University of Ottawa (PDF) via io9]