Robust facility location and protection under facility disruptions with decision-dependent uncertainty

Planning to mitigate the impacts of disruptions on supply facilities in advance is an important decision when designing a supply chain network (SCN). This paper proposes a new two-stage robust optimization framework for enhancing the resilience of the supply chain network against supply facility disruptions with decision-dependent uncertainty (DDU). In this framework, facility location and protection decisions are here-and-now decisions, anticipating the worst realization of uncertainty regarding facility disruption, while allocation decisions are wait-and-see decisions. A two-stage robust optimization model with decision-dependent uncertainty (TRO-DDU) is developed to minimize the system's total cost in the worst-case scenario, including fixed opening costs, protection costs, allocation costs, and penalty costs. To solve the model exactly, a decomposition method based on the simultaneous column-and-constraint generation (C&CG) algorithm is presented. Experimental results show that the importance of supply facilities in the supply chain network varies depending on protection budgets and disruption risk levels. Furthermore, we propose optimal protection plans for managers based on the principles of maximizing marginal efficiency (ME) and average marginal efficiency (AME). Considering the improvement of ME and AME, additional protection budgets are ineffective if the corresponding improvement in efficiency is low. Overall, the results demonstrate how our model strategically mitigates facility disruption risks and enhances supply network resilience.