Prioritising highway safety improvement projects: a robust data envelopment analysis approach

Abstract

This paper presents and evaluates an application of a robust data envelopment analysis (RDEA) multi-criteria model for prioritising safety improvement projects with limited budget and data uncertainty. RDEA can be seen as an extension of classic data envelopment analysis (CDEA) that supports a more flexible and robust project selection by enabling decision-makers to adjust the level of conservation (or robustness) of decision-making against data uncertainty. Data from an existing case study were used to evaluate the performance of the model. The results indicated that the proposed methodology provides a useful tool for adjusting the level of robustness and that the efficiency of the candidate project decreases with the decision-maker's conservativeness and their ranking does not change as the decision-maker becomes more risk-averse. As a comparative study, the proposed approach was compared with incremental benefit–cost analysis and CDEA methods. The results indicated some changes in the list of selected projects considering the uncertainty impacts of data observed according to allocated budgets.