Assessing the stochastic variability of the Benefit-Cost ratio in roadway safety management

To deal with the uncertainty inherent in the decision making process, a reliability based assessment of benefits must be performed introducing a stochastic approach. In the paper the variability of the CMFs, the predicted number of crashes and the crash costs are taken into account in a reliability based BCA to address improvements and issues of an accurate probabilistic approach when compared to the deterministic results or other approximated procedures. A case study is presented comparing different safety countermeasures selected to reduce crash frequency and severity on sharp curves in motorways. These measures include retrofitting of old safety barriers, delineation systems and shoulder rumble strips. The methodology was applied using the Monte Carlo simulations to calculate the probability of failure of BCA statements. Results and comparisons with alternative approaches, like the one proposed in the HSM, are presented showing remarkable differences in the evaluation of outcomes which can be achieved.