Risk-averse Hazmat Network Design Considering Endogenous Risk and Uncertainty

Abstract

We consider a hazmat network design problem where designer selects a feasible set of facility locations and flow assignments so that the total cost and transportation risk are minimized. While hazmat carriers choose preferred routes to transport, in particular, the route-choice is uncertain and depends on the available facilities and travel links. To improve service reliability under uncertainty, we incorporate risk-averse measures based on Conditional value-at-risk (CVaR). We model the problem as a mixed-integer trilinear optimization problem, then an equivalent linearization reformulation and a Benders decomposition algorithm with several acceleration strategies are proposed to solve this model. Numerical experiments demonstrate the effectiveness of proposed model and algorithm and give management insights.