Analyzing the Resilience of Complex Supply Network Topologies Against Random and Targeted Disruptions

In this paper, we study the resilience of supply networks against disruptions and provide insights to supply chain managers on how to construct a resilient supply network from the perspective of complex network topologies. Our goal is to study how different network topologies, which are created from different growth models, affect the network's resilience against both random and targeted disruptions. Of particular interest are situations where the type of disruption is unknown. Using a military logistic network as a case study, we propose new network resilience metrics that reflect the heterogeneous roles (e.g., supply, relay, and demand) of nodes in supply networks. We also present a hybrid and tunable network growth model called Degree and Locality-based Attachment (DLA), in which new nodes make connections based on both degree and locality. Using computer simulations, we compare the resilience of several supply network topologies that are generated with different growth models. The results show that the new resilience metrics can capture important resilience requirements for supply networks very well. We also found that the supply network topology generated by the DLA model provides balanced resilience against both random and targeted disruptions.