Modeling the impact of unmet demand in supply chain resiliency planning

This paper seeks to optimize the pre-event mitigation and post-event restoration decisions for disruptive events impacting supply chain systems, while specifically accounting for how unmet demand impacts the customers within the supply chain. The pre-event mitigation strategies will prepare the supply chain for various potential disruptions and then the post-event restoration decisions are the response of it to a specific disruption. In the context of the restoration, we consider customer behaviors which require fulfillment of demand at a certain time after the disruption and those which require that a cumulative percentage of their demand must be met on-time after the disruption or, otherwise, the customer will leave the system. In addition, we then require that the system backorders unmet demand in order to maintain the customer. We build two-stage stochastic programming models to increase the resiliency of single-product, single-echelon supply chain systems that include the cost of lost customers. We further provide theoretical insights into the impact of the two distinct customer types on the restoration planning, including conditions under which the restoration planning is independent of the customers’ behaviors. Our insights from our computational analysis include that a ‘chain reaction’ occurs in most scenarios, which means the supply chain system tends to lose the same subset of customers even when they are not served by the facilities impacted by disruption. Furthermore, different sets of post-event customer behaviors tend to favor the same pre-event mitigation strategies.