A Benders Decomposition Method for Designing Reliable Supply Chain Networks Accounting for Multimitigation Strategies and Demand Losses

This paper investigates the design of reliable supply networks to make them resilient to unpredictable disruptions. We develop an optimization model that incorporates several features, including (1) partial failure of facilities (instead of complete shutdown) resulting in interrupted supply capacity, (2) the effect of disruption on customer demand, and (3) the possibility to use multistrategies to mitigate disruption. We formulate a mixed-integer linear programming model to determine the optimal location of facilities and assignment of customers to opened facilities. An accelerated Benders decomposition method with valid inequalities is proposed to solve the problem. We discuss the computational efficiency of this decomposition procedure using two case studies as well as randomized data. For medium- and large-sized instances, our approach can decrease computational times by as much as 60% on average. We analyze the effect of multimitigation policies on the optimal solution and the model performance. Compared with the existing single-mitigation strategy models, we find that our model reduces the need for redundancy by as much as 50% and improves the total cost by as much as 8% in our case studies.