No abstract

No abstract

No abstract

ph: (865) 576-8401 fax: (865) 576-5728 email: reports@adonis.osti.gov Available to the public from the National Technical Information Service, U.S. Department of Commerce, 5285 Port Royal Rd., Springfield, VA 22161 ph: (800) 553-6847 fax: (703) 605-6900 email: orders@ntis.fedworld.gov online ordering: http://www.ntis.gov/ordering.htm This document was printed on recycled paper.A wide variety of security problems hinge on the detection of threats and discrimination of threats from innocuous objects. The theory that frames these problems is common among medical diagnostics, radar and sonar imaging, and detection of radiological, chemical, and biological agents. In many of these problems, the nature of the threat is subject to control by a malicious adversary, and the choice of a reference (or "design basis") threat is a very difficult, and often intractable, aspect of the problem. It is this class of problems that this report considers.This report formulates a threat detection problem from a decision theory (i.e. game theoretic) perspective and calculates the optimal strategies for both players. For this problem, containers pass a checkpoint which is monitored by a set of detectors. The adversary desires to introduce a container carrying a threat into the stream of "clean" containers and get it through the checkpoint without detection. The objective of the detector operator is to accomplish the opposite, to find and detain the threat containers. The specific "threat" detection problem we are most interested in evaluating is that of nuclear explosives. However, the framework developed also applies to other threat detection problems, so we present the problem in a more general form.The decision theoretic formulation most clearly describes the inter-relationships between three components of the problem; (1) the detector capabilities, (2) the player strategies, and (2) the player payoffs. Decision theory provides the best description of detector capability when the checkpoint must account for an intelligent adversary.

No abstract