Digital Twin-Driven Coordination of Multi-Level Supply Chain Resilience: Rolling Optimization of Local Recovery Decisions and System Resilience

Local recovery decisions made by individual supply chain members may improve node-level performance. Such improvements do not necessarily enhance system-level resilience. This study examines how information generated by a digital supply chain twin can be translated into coordinated recovery decisions. The study assesses whether such coordination can mitigate the mismatch between local recovery and system-level resilience. A digital twin-driven multi-agent simulation–optimization framework is developed to integrate state synchronization, causal forecasting, cross-node coordination, and rolling-horizon feedback within a common physical execution environment. Seven recovery strategies are evaluated through a structured capability comparison, functional ablation, information-quality sensitivity analysis, network-structure robustness tests, and paired statistical inference. The results show that the transition from decentralized local decision-making to a system-level coordinated optimization architecture produces the largest resilience improvement among the architecture transitions examined. Local prediction alone provides only limited gains. Rolling-horizon prediction and feedback provide conditional incremental value, primarily through improved intertemporal cost control rather than uniform improvements across all resilience indicators. Information delays and prediction errors weaken coordination effectiveness, while greater effective utilization of the latest available operational information generally improves recovery outcomes. Network structure further shapes the value of coordination: sparse networks constrain its effectiveness through insufficient alternatives, whereas high redundancy reduces its marginal value, with moderately redundant networks providing the clearest scope for coordination gains. This study conceptualizes the digital supply chain twin as an information-to-coordination mechanism. The mechanism links local recovery decisions with system-level resilience. The study clarifies the mechanisms and boundary conditions under which digital twin-driven coordination contributes to supply chain recovery.