Optimal planning of safety improvements on road sites belonging to different categories within large networks: An integrated multi-layer framework

Planning road safety interventions on large road networks implies several layers of complexity in the decisionmaking process. In fact, the following simultaneous problems should be addressed: estimating safety performances on the different road elements of the network, identifying sites showing high potential for improvement with respect to reference values, defining the possible types of safety measures to be implemented and their anticipated effect on traffic safety, limiting the number of interventions given fixed budget constraints. This study proposes an integrated multi-layer framework which takes into account the above-defined problems into a single optimization procedure which provides the number and type of safety interventions to be implemented over a wide road network composed of different categories of road elements. The proposed framework is based on the following peculiar aspects: the potential for safety improvement is quantitatively assessed based on the estimation of safety performances for each road category, a bi-level thresholding process integrated in the optimization process is used to highlight sites for interventions, the anticipated outcome of safety measures is quantitatively assessed as well through available crash reduction factors. The proposed methodology is applied to a case study which analyzes a sample of real roads belonging to a province-wide road network composed of various road elements (i.e., different categories of segments and intersections), under different budget constraints. Results demonstrate the applicability and flexibility of the proposed approach, which could be used for planning purposes, independently of the particular geographic location. Clearly, the approach is valid at the planning stage, given that several details of the different layers of analysis are necessarily simplified, while they should be studied in detail at the single intervention project stage.