Globalized robust bilevel optimization model for hazmat transport network design considering reliability

Hazardous materials (hazmat) have great risks in transportation because of their inflammable, explosive and corrosive properties. Once an accident occurs, it will cause irreversible and huge impact. Notably, the design of hazmat transport network is a complex transport network, usually faces two challenges: special hierarchical relationship and only partial distribution information about accident probability. Motivated by these two challenges, we develop a bilevel optimization method for the particular hierarchical relationship. Moreover, for the ambiguous accident probability, we construct two uncertainty sets containing the true probability distribution and a “soft constraint†with phi-divergence distance to ensure the connectivity reliability of network arcs. In order to improve the safety of transportation, a globalized robust bilevel hazmat transport network design model is proposed. In addition, our new robust model can be transformed into a computationally tractable mixed integer linear programming model by cone duality theory, which has certain practicability. Finally, we apply the proposed model to a real transport case about a hazmat transport company in the Beijing–Tianjin–Hebei region of China to show its superiority, and compare it with the traditional optimization method. The important parameters in the model, safety parameters and global sensitivity parameters, are also analyzed in detail.