Terror Queues

The same techniques that help minimize the time customers have to wait on hold to speak to a customer service representative can help predict terror plots, finds new research by Edward H. Kaplan, the William N. and Marie A. Beach Professor of Management Sciences, Professor of Public Health, and Professor of Engineering.

Since the September 11 terrorist attacks, Kaplan has been applying the tools of operations research to address problems in counterterrorism and homeland security. In his study Terror Queues (pdf), published in the July/August 2010 issue of Operations Research, Kaplan borrows ideas from queuing theory to create a novel method to estimate the number of ongoing yet undetected terror plots and how many of those threats can be detected and stopped by undercover intelligence agents.

Queuing theory is most often used to improve service for customers waiting in checkout lines, call center phone queues, and hospital waiting rooms. Queuing models provide information about customers and servers in a system — for example, the number of customers receiving and waiting for service, the amount of time customers have to wait for service, and the number of servers who are either busy or available to provide service. Kaplan’s terror queue models treat terror plots as customers, undercover intelligence agents as servers, and detection of terror plots as the service they provide. Successful terror attacks are customers who drop out of the system before they receive service.

"This is a service system where, from the government point of view, it is really important to provide good service," says Kaplan. "On the other hand, from the customer point of view, lousy, and ideally, no service, is preferred."

The terror queue models are based on the relationship between the number of terror plots and the activity of undercover intelligence agents who spend their time either detecting and infiltrating new terror plots (in which case they are "available") or stopping previously detected terror plots (in which case they are "busy").

Kaplan explains that the insight for the model was prompted in part by a report released in May 2007 by MI5, Britain’s security service, which noted that the agency was monitoring approximately 2,000 Al Qaeda supporters in the United Kingdom. "Some people thought, 'Wow, that’s scary.' Other people thought, 'Only 2,000? I would have thought more.' But my reaction was: How many Al Qaeda sympathizers would MI5 think there are if they doubled the number of MI5 agents? Or halved them for that matter? It hit me that the estimated size of a terror threat depends not only on how big the threat really is, but also on the resources, like undercover agents and informants, that are invested in detecting terror plots."

The models also take into account an important reality in counterterror operations: the false detection of fake terror plots. Agents can make mistakes and spend time investigating suspects who are not involved in terrorism, and terrorists deliberately broadcast false information intended to divert agents.

Since information on counterterror operations is classified, Kaplan illustrates the models using a hypothetical example based on suicide bombing data from Israel. Intelligence agencies can estimate the terror threat level through information from the activity logs of undercover agents and statistics from past cases, such as the number of agents in the field and the average time to it takes agents to detect a new terror plot.

In addition to estimating the number of current terror threats in an area, Kaplan explains that the terror queue models will also benefit governments by helping to allocate the right amount of resources to counterterror efforts. "If the estimated number of ongoing but undetected plots is high, it suggests that additional undercover resources are needed to increase the detection rate," says Kaplan. "On the other hand, if you really don’t estimate that there are many more plots beyond those that you know about, then maybe the current effort is sufficient."

The research was conducted under the auspices of the Technion-Yale Initiative in Homeland Security and Counterterror Operations Research, a three-year partnership supported by a gift from Daniel Rose, chairman of Rose Associates, Inc., and a 1951 graduate of Yale College.