A Hybrid Simulation Modelling Framework for Stress Testing Rare Earth Element-Centric Semiconductor Supply Chains

Semiconductor supply chains (SCs) depend on geographically concentrated rare earth element inputs and face systemic risks from geopolitical disruption, policy intervention, and capacity constraints across multiple decision-making layers. Existing stress testing approaches often quantify fulfilment loss without attributing it to strategic capacity shortages, tactical sourcing-access constraints, or operational congestion, even though each requires a different intervention. This study develops a hybrid simulation modelling framework integrating System Dynamics (SD), Agent-Based Modelling (ABM), and Discrete Event Simulation (DES) as necessary analytical approaches per decision-making layer: SD captures long-term capacity stocks and policy feedback (macro-level); ABM captures heterogeneous monthly sourcing and inventory decisions (meso-level); and DES captures stochastic weekly fabrication and shipment execution (micro-level). Applied to India's cerium-centric semiconductor SC through 247 experiments, 12,350 simulation runs and an additional 1,400 runs of sensitivity analysis, the results show that upstream cerium capacity disruptions are largely buffered, whereas loss of qualified fabrication capacity causes severe fulfilment collapse. Stable DES cycle times further confirm that this failure arises from sourcing access rather than operational congestion. This cross-layer diagnostic signal is only detectable when all three modelling layers are present. Managerially, the hybrid simulation modelling framework reframes stress testing as a multi-level tool that identifies not only whether an SC fails, but also where failure emerges and which interventions are appropriate.