Building sustainable hazardous products supply chain against ambiguous risk with accelerated Benders decomposition algorithm

Abstract

The detrimental consequences of accidents in the supply chain pose a major challenge to the management of transportation risks in the hazardous products supply chain. The development of a sustainable hazardous products risk management method is an important research problem. For this purpose, this paper investigates the hazardous products supply chain design problem with the deployment of an emergency response team (ERT) before and after supply accidents, and proposes a new hazardous products supply chain risk avoidance approach. Our approach considers the government as the upper decision-maker and the company as the lower decision-maker, and adopts a bilevel optimization framework to characterize the hierarchical relationship in our problem. To model the uncertainty of arcs risks and availability of partial distribution information, this paper constructs an ambiguous joint chance constraint based on the Wasserstein ambiguity set. To improve the reliability of the network arc connection and the sustainability of supply, a multi-objective bilevel distributionally robust (MBDR) model is developed. Moreover, we reformulate the proposed model as a computationally tractable mixed-integer linear programming (MILP) model. To further improve the solving efficiency, we design an accelerated Benders decomposition (BD) algorithm by incorporating two sets of valid inequalities. Finally, a practical case in Guangdong Province is presented to illustrate the superiority of our proposed method and algorithm. The computational results show that our MBDR model and BD algorithm perform better in terms of out-of-sample performance and exhibit good solving efficiency.