(Georgia Institute of Technology, 2021-06)
Kulkarni, Onkar; Cohen, Yaarit; Dahan, Mathieu; Montreuil, Benoit
Logistics networks frequently face disruptions inducing an increase in delivery costs and delays. This paper studies the design of resilient hyperconnected logistics hub networks for the Physical Internet, modeled as an integer programming problem. The objective is to open logistics hubs in order to connect each origin and destination using multiple minimum length edge-disjoint paths. To estimate the resilience of the designed networks, we propose graph theoretic measures involving (i) the maximum number of edge-disjoint paths connecting each origin and destination, and (ii) the number of short paths traversing each edge. We develop a case study to design a class of parcel delivery networks in China and evaluate the impact of various disruption scenarios on the resulting distance traveled by parcels. Our results show the relevance of the proposed resilience measures and the increased capability of the designed networks to sustain disruptions in comparison to traditional logistics networks.