Comparative Results from Applying a Multilayered Network Framework to Engineer Nuclear Security Systems

Building on the strong legacy of current state-of-the-art approaches for nuclear security—including the Design Evaluation Process Outline (DEPO)—current research at Sandia National Laboratories is exploring the effectiveness of describing nuclear security as a multidomain system visualized as multiple, interacting layers. This research effort is focusing on developing next-generation security engineering concepts to address challenges stemming from complex risk environments, innovative adversaries, and disruptive technologies. From an analytical perspective, this research re-examines core analytical assumptions to more formally incorporate dynamic interdependencies and nth-order effects observed—and anticipated—in nuclear security operations. More specifically, this research leverages key insights from systems, network, complexity and resilience theories to develop multilayer-network models (MLN) for security. MLN models conceptualize security performance in terms of emergent behaviors from multidomain interactions better capable of addressing new—and evolving—sources of complexity. The ability of MLNs to better capture, model and simulate non-linear and dynamic behaviors in nuclear security systems provides both quantitative and qualitative benefits over more highly linear traditional security models. After summarizing early lessons learned on how this multilayered network-based approach captures technical-personnel-digital-organizational interactions in nuclear security systems, this paper will introduce a set of principles and related metrics to better describe these interactions. Next, this paper will explain the experimental design used to compare the MLN security model against a DEPO-based path analytic model. This paper will then summarize the analytic results from each model and discuss a range of outcomes stemming the results comparison. Lastly, this paper will review conclusions and insights for the efficacy of MLN models, as well as implications for future R&D efforts to enhance nuclear security. SAND2022-1221A.