Uncertainty modeling of hurricane-based disruptions to interdependent economic and infrastructure systems

Abstract

Extreme risks associated with natural and man-made disasters involve disruptions to the production of goods or provision of services in interdependent systems. The reduced supply of goods and services will degrade “as-planned” production and create ripple effects. Hence, maintaining above-minimum levels of inventory is a resilience strategy that could effectively reduce the onset of disruption. This research integrates the uncertainty in inventory levels to assess the economic impacts of moderate and extreme disastrous events on interdependent systems. The unique contribution of this research is the formulation of a stochastic inventory-based risk assessment model using a multi-objective optimization framework for minimizing (1) extreme economic losses and (2) sector inoperability. Empirical distributions are derived from inventory-to-sales ratio (ISR) of the manufacturing and trade sectors from the Bureau of Economic Analysis database. Simulations of inventory enable the initialization of inoperability functions of a dynamic inoperability input–output model (DIIM). The stochastic inventory-based DIIM-computed values of economic loss and inoperability are simultaneously minimized to identify inventory-enhancement opportunities for critically disrupted systems. A lean production case for a moderate-intensity hurricane in Virginia reveals an overestimation of regional economic loss relative to the expected inventory levels from the ISR data. The conditional expected value of regional economic loss for an extreme event is found to be $12 M higher than a moderate-intensity case. Identification of resilience-enhancement opportunities using the proposed multi-objective optimization framework could reduce expected economic loss by $24 M for an extreme-event and $17 M for a moderate-intensity case.