Within the framework of the SICURO research project, the main objective of this paper is to define the procedures to be planned and activated in emergencies in order to allow the evacuation of weak users (disabled, old persons, etc.) from the area affected by a disaster and design the optimal path for emergency vehicles to reduce evacuation times. Specification, calibration and validation of a path choice generation model in order to simulate the behaviour of emergency vehicle drivers at an urban level during an evacuation is proposed. We specify the factors that affect path choice behaviour and the two main approaches: one to one and many to one. The first regards the minimization of generalized cost of a path that connects an origin to a destination; the second instead considers the connection of one origin to many destinations. We also report some experimental results, obtained in the context of the SICURO Project, by applying the proposed model to a real road transport network at urban scale during a simulation of an evacuation.

In this paper an advancement on the SICURO Research Project relating to the design of path choice models for emergency vehicles is proposed. In particular, with respect to the previous work in which the procedures to be planned and activated in emergency conditions were defined in order to allow the evacuation of weak users (the disabled, senior citizens, etc.) from the area affected by a disaster, some advanced instruments and scenarios for designing the optimal path for emergency vehicles to reduce evacuation times are introduced. The problem of an emergency vehicle that has to pick up some users at fixed points of the network and to take them to the refuge area is schematized (as in a previous work) with two different approaches: as a shortest-path problem (one to one) and as a vehicle routing problem (many to one). In the first exact approach the k shortest paths algorithm is applied to obtain the best k paths that satisfy specific choice criteria; in the second approach a metaheuristic procedure (genetic algorithm) which allows route optimization of a fleet of emergency vehicles is proposed. Some new experimental results obtained by applying the proposed advanced model to a real road transport network during a simulation of evacuation at urban scale in the context of the SICURO Project are also reported and compared with those obtained previously.

This paper deals with path design for emergency vehicles (ambulances) in emergency conditions (hurricanes, earthquakes, etc.) to rescue weak users (the elderly, disabled persons, etc.). The problem of an emergency vehicle that has to pick up weak users at fixed points of the network and to take them to the refuge area is schematized with two different approaches: one-to-one and many-to-one. The first regards the minimization of generalized cost of a path that connects an origin to a destination (as a shortest-path problem); the second considers the connection of one origin to many destinations (as a vehicle routing problem). Starting from a brief review of both these problems, the available tools for planning evacuation of weak users in emergency condition in the within of oneto-one and many-to-one are analysed. Finally, on the basis of a real simulation of evacuation, we provide experimental results obtained by testing the effectiveness of a commercial tool when little time is available for planning evacuation operations.

No abstract

In this paper what to models are considered in order to define optimal emergency vehicle distribution (in terms of vehicle numbers, weak users sequence to visit) and hence paths/routes with the objective to optimize the total time for safe users in evacuation condition. The design problem is tackled with a multilevel approach, which allows, by subsequent steps, to consider the network performances, the vehicle paths (One to One Problem) and the vehicle routes (Vehicle Routing Problem). In this paper the whole model is specified in terms of input and output variables and objective function.