Acquisition Path Analysis Quantified Shaping the Success of the IAEAs State-level Concept

The IAEAs Department of Safeguards has embarked on an evolutionary process to more fully develop and apply the state-level concept (SLC) for safeguards implementation. In an attempt to direct safeguards to areas of significant proliferation concerns within a state, this concept makes use of all safeguards-relevant information available in order to focus and prioritize its safeguards activities for a state. A key component is the development of a customized safeguards approach for an individual state. Developing these state-level approaches consists of analyzing acquisition paths, establishing and prioritizing technical objectives, and identifying applicable safeguards measures. The paper presents a methodology to accomplish this process based on a threestep- approach: network modeling, network analysis, and strategic assessment. The network modeling step assesses and models the states nuclear capabilities as well as other state-specific factors concerning relevant proliferation scenarios. The network analysis step gives a ranking of all plausible acquisition paths including a visualization of the paths. Finally, the strategic assessment step evaluates the states proliferation and compliance options as well as the IAEAs set of technical objectives and subsequent safeguards measures. In this paper, a hypothetical state model was developed in order to test the methodologys performance. Therefore, an Excel spreadsheet with all necessary state-level factors was created. Afterwards, a Python software module based on graph theoretical algorithms was applied to produce a comprehensive list of ranked acquisition paths including their visualization. The following step of the strategic evaluation is mainly based on the concept of the Nash equilibrium resulting in a stable combination of the states and the IAEAs strategies. This formal and automatic procedure offers the advantage of gaining results in a comprehensive and non-discriminative manner. Besides presenting and discussing the methodology in detail, results from an example case study will also show how this process could be carried out in practice. Furthermore, the problem of how to determine the model parameter detection probability will be discussed. The paper ends with conclusions and an outlook on planned future work in this area.