Passenger grouping under constant threat probability in an airport security system

In a typical security system at a US airport, there are a series of inspections done on passengers as well as on baggage to check whether any item of threat is entering into the system. Considering the large number of passengers and items using our airports, one hundred percent check on all the items is not practical. This paper investigates the benefit of classifying passengers into different groups, with the idea that the number of checks and the degree of inspection may vary for different groups. The threat probability is assumed to be known and identical for all passengers. We develop a model to determine the number of groups, the fractions of passengers and the assignment of check stations for each group. The constraint is that the false clear probability is within Federal Aviation Administration (FAA) specifications, and the objective is that the number of false alarms (a surrogate measure of passenger inconvenience) in the system is minimized. The model studies the effect of the system parameters on the number of items checked at various check stations and on the false alarm rate. The major conclusion of this paper is that passenger grouping is beneficial even when the threat probability is assumed constant across all passengers. A further conclusion is that the optimal grouping and check station assignments can depend on the overall threat probability.