Robust Optimization for the Risk Mitigation of Hazardous Material Road Transportation System

Road transport is the primary way of transporting hazardous material (HAZMAT). The risk of HAZMAT transportation has great uncertainty due to the time-varying conditions of its transportation location and environment. If an accident occurs, it will bring severe casualties and social losses, which makes risk management challenging. For this problem, considering the nonlinearity and complexity between risk factors, the Causal Bayesian Network (CBN) is applied to describe this relationship. In this work, we consider the intervention impact within a limited budget on risk factors in the CBN. In addition, due to the deep risk uncertainty and the data scarcity, interval probability is introduced. Then a new robust optimization approach that combines the do-calculus on CBN and a mathematical programming model is developed to minimize the worst-case HAZMAT road transportation risk. Finally, numerical experiments are conducted, and some insights are presented.