RISK MANAGEMENT IN INTERNATIONAL SUPPLY CHAINS: SCENARIO PLANNING AND ROBUST OPTIMIZATION

Abstract

Global supply chains face heightened exposure to compound disruptions arising from geopolitical shocks, seasonal peaks, labor actions, and force majeure events. These drivers increase lead-time variability, jeopardize on-time-in-full performance, and raise cost-to-serve; emergency rerouting and mode shifts also add Scope-3 carbon externalities. This study develops and evaluates an integrated framework that combines scenario planning with robust optimization to support cross-border network, inventory, and transport decisions under distributional ambiguity. We generate a portfolio of plausible, internally consistent scenarios that cover seasonal demand surges, capacity shortfalls from strikes, corridor closures or delays linked to geopolitical risk, and weather-driven outages, then apply scenario reduction to obtain tractable, representative sets. The optimization layer implements robust and distributionally robust models with service-level and carbon (Scope-3) constraints, benchmarked against deterministic and two-stage stochastic baselines. Using industrially realistic instances, we assess performance in terms of expected cost, tail risk (measured by Conditional Value at Risk, CVaR), on-time-in-full, lead-time variance, and CO₂e. Results indicate that robust designs improve worst-case service and lower tail risk with modest cost premia. Multimodal flexibility, dual sourcing, and targeted buffers are especially effective when shocks are correlated across lanes and periods. For managers, the framework provides a repeatable cadence for scenario reviews and contracting choices; for policymakers, it underscores the importance of corridor governance, customs cooperation, and labor-mediation mechanisms. Limitations include a static planning horizon, proxy-based geopolitical indicators, and a single-industry testbed. Future research should examine adaptive robust control, learning-augmented forecasting, and higher-fidelity carbon accounting.