Finding Needles in a Haystack

Detecting terrorist weapon development is a fundamental goal of the intelligence and law enforcement communities. Achieving this goal can be quite difficult as many of the actions taken by terrorists can be executed covertly or may be seemingly innocuous against a background of non-terrorist related activities. This dissertation presents a systematic resource allocation methodology to design strategies to detect terrorist weapon development. First, a framework to approach the problem of detection of terrorist weapon development is introduced. Then, weapon pathways are generated, which define the target set of potential evidence the intelligence and law enforcement communities could pursue to discover terrorist weapon development. Finally, Bayesian networks are used to create a logical structure for how potential observations would affect our belief a weapon is being developed. Information entropy measures how much uncertainty is present in a system and can be used to assess the relative information content of potential evidence in the Bayesian networks. Resource allocations can be guided by these information-theoretic measures. The dissertation then shows how these methods might be used to detect terrorist development of improvised explosive devices (IEDs) and radiological dispersal devices (RDDs). This method is an example of how expert judgments made prior to observations can guide collection and analytic resource allocations.