Author(s)

A. Polimeni & A. Vitetta

Abstract

In this paper, a shortest paths algorithm to find the paths in a time-dependent network is implemented and tested both on test and real networks. In the application on the real network, the data (obtained with real-time observations) are integrated with forecast data (to define the costs on non-observed network elements) obtained from a dynamic traffic simulation. The real case application is performed in emergency conditions; the aim is to optimize emergency vehicle routes (defined as a path succession) to reduce intervention times. Keywords: time-dependent network, dynamic shortest path algorithm, route design, emergency conditions. 1 Introduction In this paper we report an experiment related to paths and routes design for emergency vehicles in a time-dependent network. In a previous work [1] a model for path and route design in time-dependent networks was discussed. For simplicity sake, we refer to [1] for the full model formulation. Travel time forecasts and best vehicle routing are particularly important in emergency conditions. In a road network in emergency conditions it is essential to provide system users’ information in real time. To achieve this objective, we must know the evolution of link cost functions. Starting from the observed data, it is possible to calibrate the link cost functions, simulating the system in a time slice [1, 2]. This approach allows us to forecast link costs as time varies and hence to design paths (at the limit, instant by instant). The time -dependent link cost [3, 4] is

Keywords

time-dependent network, dynamic shortest path algorithm, route design, emergency conditions