Musharraf Ahmad Khan

Hazardous material transportation problems have widely been studied in the past especially in the context of routing, scheduling, and network design problems. Yet, the combined hazardous material facility location-routing problem has not been studied adequately. We emphasize that locating a hazardous material facility is a rich process, and a good site can mitigate the potential transportation risk beforehand. A methodological framework is proposed which allows evaluation and ranking of potential sites based on hierarchical relationship utilities. The proposed method attempts to improve the risk functions and applies a stochastic analysis to measure the risk, which relaxes some assumptions in deterministic analysis, and is more realistic while avoiding overestimation of the risk. The study covers multi-objective optimization considering the decision-makers’ preferences on network segments and risk to the population and water bodies. Potential hazardous material facility sites’ rank is determined by the probability of optimality and one-to-one relationship utilities with the points of interests. Results show that the proposed stochastic analysis offers more flexibility to select and rank the potential sites.

Resilience is a desired attribute for critical networks such as a transportation network which can be vulnerable to several internal and external hazards. Network resilience provides a broader-level network evaluation regarding its ability to absorb the hazards. However, this broader level evaluation seems abstract for a path-specific evaluation. The shipments often follow designated paths, especially in transporting hazardous materials (HAZMAT). We suggest that path-level resilience provides a direct and effective way to evaluate and mitigate transportation risks. We specify a ‘path resilience’ model considering multi-criteria settings, and propose a novel stochastic model to optimize and rank hazardous materials facility siting based on the path resilience concept. In the proposed model, designated paths are stochastically disturbed to generate disruption scenarios and foster alternate paths. Subsequently, Cluster-based Range-specific Maximum Likelihood Sampling is applied to sample disruption scenarios. Next, the Trade-off concept is applied to model the path resilience as Discrete Markov Chain (DMC) to capture the resilient behaviour of designated paths in response to random disruption scenarios. DMC enables ranking the potential sites for HAZMAT facilities while effectively handling the big data generated during the modelling process. Finally, site rankings are integrated by applying Markov Ranking Method and Cross-Entropy Monte Carlo Method on multiple points of interest considering multi-criteria settings. Evidently, the routing-location problems (LRPs) in HAZMAT transportation deal with locating sites only. By introducing ranking explicitly, the proposed model presents a new variant of LRP as the Ranked Location-Routing Problem or rLRP, in which locations are ranked as part of the solution. The significance of ranked locations is empowering, especially in a multi-criteria decision support system.