A 2D Defense-in-Depth Waypoint Graph Framework for Evaluation of Physical Protection Systems at Research Reactors

Research reactors represent potential targets for sabotage scenarios with severe radiological consequences, requiring rigorous evaluation of the effectiveness of physical protection systems (PPSs). Traditional evaluation tools based on the Estimate of Adversary Sequence Interruption model and its extensions employ one-dimensional adversary sequence diagrams (ASDs) that abstract facility geometry into sequential protection layers, limiting the spatial analysis of alternative adversary routes. This study presents a two-dimensional waypoint graph framework for evaluating the effectiveness of PPSs. Facility layouts are represented as sparse waypoint graphs encoding spatial coordinates, probability of detection PD, and barrier delay times td, with adversary movement constrained by the progression of the defense-in-depth layer. A backward Dijkstra-based pathfinding formulation with exhaustive critical detection point enumeration is used to identify the most vulnerable path under a conservative adversary strategy, without relying on heuristic assumptions. The global pathfinding solution was independently verified through full enumeration of feasible adversary paths. A Monte Carlo simulation was integrated to quantify the influence of the detection probability variability on the probability of interruption distribution. A hypothetical research reactor case study demonstrates that the proposed framework identifies more vulnerable adversary paths than the ASD-based approach under equivalent spatial traversal conditions, resulting in probability of interruption PI values up to 4.25% lower under varying response force timing scenarios. The complete evaluation demonstrated computational tractability suitable for iterative PPS evaluation. Analysis of modified PPS configurations showed where the greatest vulnerability lies within the facility.