Facility Protection Optimization Under Uncertainty
Hester, Patrick Thomas
The goal of this study is to develop a decision-making methodology to optimally allocate resources in designing a facility protection system, composed of detection, delay, and response elements, in order to balance cost and performance. The developed methodology extends current network interdiction techniques by including concepts such as timely detection, guard response time, and conflict between the facility response and adversary forces. Both single and multiple adversary teams are considered. The methodology decouples critical path selection and safeguards optimization, thus increasing computational efficiency. Further computational savings are gained by utilizing reliability-based design optimization with a first-order reliability method approach to analyzing uncertainty. These computational improvements allow for the facility protection methodology to be extended to large networks and to incorporate multiple adversary team threats. A real world example is utilized to demonstrate the applicability of the methodology to problems of large scale complexity.