Modern production planning in Global Production Networks faces a multitude of uncertainties, ranging from short-term cost and demand fluctuations to political instabilities and large-scale disruptions. The varying types and time horizons of both uncertainties and planning parameters create the need for multi-scale scenarios that capture their different characteristics and developments across flexible time scales. This paper presents a novel scenario generation framework for deriving multi-scale scenarios along both type and time dimensions, serving as input to enhance the robustness and resilience of Global Production Networks. The scenario generation follows six consecutive steps, applying MonteCarlo-Simulation to defined Change Drivers, Receptor Key Figures, and their interdependencies. The generated scenarios are clustered into representative scenarios using the k-means algorithm. The framework is integrated into a Decision Support System featuring an optimization module for longterm planning of Global Production Networks. An application to an industrial use case shows that the resulting network configuration is more robust than the forecast-based configuration.
Martin et al. (Fri,) studied this question.