Over the past several years, airport and transportation security has become a major issue in the US and around the world.  The problem is that there are so many different scenarios for what could go wrong, why and how, that it’s unrealistic for security officials to reason about all the variables at play and to make appropriate policy from this. This is where Monte Carlo methods come in. As the name suggests, Monte Carlo methods rely on random samplings of scenarios in conjunction with some more structured algorithms to make logical inferences on the various likelihoods for certain scenarios to play out.

In June 2007, the Transportation Security Administration (TSA) announced that they would be working with Boeing in the implementation of Monte Carlo methods to assess vulnerabilities in the commercial aviation system.  The reason why the Monte Carlo method is so useful in this field is that there are simply too many variables to factor in when it comes to air travel. Beyond the often talked about factors such as screening for racial backgrounds and passengers buying one way tickets, there are an incredible number of other variables that may actually prove to be more likely to cause a catastrophic incident. Despite the wide range of factors to possibly lead to incident, this method also works well to enhance the effectiveness of existing policies on screening and other areas such as airline employee security clearance.

To get a basic idea of how the Monte Carlo method works see the example below:

(source: Wikipedia)

Here we a simple game of battleship, but with Monte Carlo methods being used.  Initially a player would fire randomly at an opponent to create the necessary random sampling.  Following this, certain restraints and conditions would be applied using an algorithm. In this case the boat would be constrained to taking up four continuous spaces.  Finally, from this a shot would be fired in a spot that fits with the random data as well as the constraints presented by the algorithm.

So while the Monte Carlo method is not deterministic, it’s not too hard imagine that for using it in a field such as airline security, that its results would be far more useful than considering significantly fewer scenarios in a more deterministic manner.

Sources:

Wikipedia  - http://en.wikipedia.org/wiki/Monte_Carlo_method#Application_areas

GNC – Government Computer News - http://www.gcn.com/print/26_13/44398-1.html