A robust optimization approach to risk-averse routing of marine crude oil tankers

Marine accidents involving crude oil tankers are rare though such incidents have led to catastrophic socioeco-nomic and environmental consequences. This low probability - high consequence nature of marine oil spills renders the existing expected consequence technique ineffective and calls for the development of a risk-averse approach. We propose a conditional value-at-risk (CVaR) based methodology for selecting the appropriate size of crude oil tankers and then routing them over the intercontinental network such that the weighted sum of transport cost and CVaR based transport risk is minimized. In addition, we also develop a robust formulation of the mixed-integer routing problem, which not only facilitates tiding over imprecision in the risk data but also prepares risk-averse shipment plan. The proposed methodology was used to analyze the realistic marine network of a major oil supplier and to demonstrate its benefits, and to underscore the value of incorporating robustness.