A STOCHASTIC APPROACH TO SAFETY MANAGEMENT OF ROADWAY SEGMENT

The targets of the thesis are road safety analysis based on crash event and road features for the benefit cost analysis where a treatment has to be applied. There is a growing attention to road safety in Europe. New regulations applied on TERN network push Agencies to introduce new methodological approaches to Road Safety, monitoring the treatment and controlling the level of safety on the managed road network. The regression to the mean effect as well as the functional relationship between crashes and exposure factors don t allow a reliable estimation of the effect of a treatment. Have a high reliability in the estimation of the effects of a countermeasures should be one of the most important issue in a Benefit-Cost analysis. The traditional techniques don t take into account the possible site to site variability of a treatment applying a deterministic approach to the problem. The present research work seek to address this issue introducing in the Benefit-Cost analysis a variability accounted by the cross site variance of the investigated treatment. The comparison between the stochastic and the determinist approach shown a general overestimation of the benefit of the latter. To do this two different methodologies were introduced either for existing infrastructure or for new infrastructure. For the latter a further variability for the predicted crashes was introduced in the computation of the possible benefit and loss. Furthermore a Montecarlo simulation was performed to consider the combination of two different Crash Modification Factor in the benefit-cost analysis.