Resilience of weighted interdependent supply chain networks considering hybrid underload-overload cascading failures

This study investigates the resilience of the weighted interdependent supply chain network (WISCN), a novel model that can more accurately describe real-world supply chains. Resilience against disruptions is typically challenged by both structural (caused by dependency loss) and functional (triggered by overload or underload) cascading failures. Although the consideration of both overload and underload cascading failures is crucial, this integrated perspective has largely been lacking for WISCNs. Accordingly, we first develop a novel model for generating WISCNs with tunable characteristics, such as edge weights and node-selection strengths (with linear and exponential scaling). We then propose a hybrid underload-overload cascading failure model to capture the complex propagation of failures within the WISCNs. Using both model-generated and real-world WISCNs, we systematically investigate the impact of these network characteristics on WISCN resilience. The results yield the following key findings: (1) Edge weight exponent (γ) quantifies inter-node connection strength sensitivity to node degree; a larger γ strengthens core nodes disproportionately, their failure causing severe neighbor load shocks and rapid resilience decline. (2) Linear and exponential selection strengths indicate new nodes’ preference for high-degree nodes; higher values concentrate links on few hubs, reducing resilience. (3) Edge count (m) is the number of links added when a new node joins; resilience improves with m until m = 2, then drops sharply when m > 2 due to excessive density. Finally, from these findings we derive practical management insights for designing resilient WISCNs and outline promising avenues for future research.