Identifying geographically diverse routes for the transportation of hazardous materials

Often, the carrier/shipper of hazardous materials is interested in a collection of routes with approximately the same performance so that they can switch between different routes to avoid exposing the same population and potentially as a security measure. We develop a K shortest path algorithm for which the performance of each highway facility, with respect to each objective, can be stochastic and can vary over time. We also devise a mixed integer program to identify a subset of paths, which represents an acceptable trade-off between geographic diversity and performance. These models are then applied to a realistic case study.