Scheduling policies for an antiterrorist surveillance system

Abstract

"A surveillance system is designated to detect terrorists in a crowded area to prevent a potential attack. Such a system usually does not have the capacity to screen all the people in the area. If the sojourn time distribution of a terrorist is different from that of the other people in the crowd, then it is possible to increase the probability of detecting a terrorist in time by using a decision rule to choose whom to inspect next. We use a queueing model with impatient customers to describe the interaction between the surveillance system and the surveyed people. We identify a few cases when a simple service rule--such as the first-come-first-serve-rule--is optimal. In general, we develop a heuristic policy that is particularly effective in an area with heavy traffic, such as an airport check-in lobby."-- p. v