Abstract

Multi-echelon facility location models are commonly employed to design transportation systems. While they provide cost-efficient designs, they are prone to severe financial loss in the event of the disruption of any of its facilities. Additionally, the recent crisis in the world motivates OR practitioners to develop models that better integrate disruptive event in the design phase of a distribution network.

In this research, we propose a two-echelon capacitated facility location model under the risk of a targeted attack, which identifies the optimal location of intermediate facilities by minimizing the weighted sum of pre and post interdiction flow cost and the fixed cost of opening intermediate facilities. The developed model results in a tri-level Mixed Integer Programming (MIP) formulation, reformulated in a two-level MIP. Hence, we prescribe solution methods based on Bender Decomposition as well as two variants that enhance the speed performance of the algorithm.

The results reveal the importance of selecting backup facilities and highlight that premium paid to design a robust distribution network is negligible given the benefit of reducing the post-interdiction cost when a disruptive event occurs.