Neighbor Integration Centrality: Prioritizing Critical Intersections in Urban Road Infrastructure under Targeted Disruptions

Urban road networks are critical infrastructure systems whose local disruptions can degrade mobility, emergency response, and access to essential services. For city-scale protection planning, a practical first step is to rapidly identify structurally vulnerable intersections. This paper proposes Directed Weighted Neighbor Integration Centrality (DW-NIC), a lightweight structural pre-screening metric for directed and weighted urban street graphs. DW-NIC ranks intersections by combining road-class-weighted directed through-movement load with the lack of bounded-hop local bypass capacity, thereby targeting junctions whose removal may degrade directed accessibility, local detour availability, and cost-sensitive network efficiency. The proposed metric has a worst-case time complexity of O(mΔ²), where m is the number of directed edges and Δ is the maximum directed neighborhood size. We evaluate DW-NIC on twelve real-world OpenStreetMap-derived urban road networks and compare it with eleven baselines. We conduct targeted node-removal experiments under 5% and 10% removal budgets, complemented by statistical testing, runtime analysis, and ranking-overlap analysis, with residual damage quantified using nine indicators. Across the full city-indicator-budget evaluation space, DW-NIC achieves the best aggregate mean rank, remains in the top three in 211 of 216 cases (97.7%), and is particularly effective on weighted efficiency loss, directed reachability loss, largest strongly connected component reduction, and weak connectivity loss. Runtime and overlap analyses further show that DW-NIC provides a favorable trade-off between scalability, stability, and distinct critical-node selection. These results support DW-NIC as an interpretable structural first-filter metric for prioritizing candidate intersections in critical-infrastructure protection workflows.