Resilience of supply-chain systems under perturbations: A network approach.

Supply-chain systems (SCSs) are an indispensable part of our daily infrastructures. Note that a small perturbation in a SCS can be amplified, eliciting cascading failures. It is of significant value to ensure a high resilience of SCSs. However, due to the complexity of SCSs, it is quite challenging to study their resilience under conditions of perturbations. In view of this, this paper presents a complex network perspective toward the resilience of SCSs. To achieve this goal, a complex SCS is modeled as a multilayer supply-chain network (SCN) with physical organizations being modeled as nodes and interactions among them as edges. A modeled SCN contains three types of nodes, i.e., suppliers, manufacturers, and retailers. An algorithm is proposed to construct a multilayer SCN. For each layer of a multilayer SCN, two kinds of networks, i.e., networks with Poisson degree distributions and networks with power-law degree distributions, are considered. For a given multilayer SCN, a ripple-effect network model is proposed to analyze its resilience under perturbations. Regarding the perturbations, two scenarios, i.e., random node failures and target node failures, are adopted in this research. In order to validate the effectiveness of the proposed network perspective, simulations on computer-generated SCNs are carried out. Interestingly, it is found that the resilience of SCNs under both random and target perturbations presents a discontinuous phase-change phenomenon, which indicates that SCNs are quite fragile under perturbations. It is further noticed that SCNs with power-law degree distributions are relatively more robust than SCNs with Poisson degree distributions. Although SCNs are found to be fragile, it has been discovered that denser interactions between different system organizations can enhance the network's resilience.