Reliability of transportation systems can be affected by events both exogenous to the system (i.e. earthquakes, floods, terrorism attack, etc.) and endogenous to the system (i.e. dangerous freight, etc.). Hazardous events can introduce disconnections in transportation systems whose effects can be relevant both in the short term (difficulties in evacuation procedures and aids to population) and in the long term (effects in local services and local economies). For these main reasons, in particular considering non urban situations, to guarantee an acceptable level of accessibility among urban centers, an analysis of network vulnerability is relevant. The interest of the scientific community on those topics concerning transportation network reliability has been increased in 90s, in particular after the Kobe earthquake of 1995. Several definitions and classifications have been reported in literature concerning reliability (Berdica, 2002); a first classification can be made between those methodologies considering only supply system e those ones taking also into account demand-supply interaction. Within the first case it is possible to distinguish two aspects: the first one concerns the connective reliability, and some analytical procedures are proposed (Bell and Iida, 1997, Iida, 1999) in order to individuate the reliability of connections; moreover (D'Este and Taylor, 2003) some indicators are proposed to investigate on the weakness points of a network. The second one concerns the infrastructural reliability where reliability of capacity is taken into account (Chen et al., 1999, Yang et al., 2000); it is defined as the probability that a network, whose conditions can be either normal or deteriorated (i.e. due to accidents), can satisfy a certain level of demand that is can be able to accept a certain amount of traffic volume. This probability (Chen et al., 1999) is defined equal to the probability that residual capacity of a network results, once given the loss of capacity due to the deteriorating conditions, greater or equal to the demand. In the second travel time reliability has been taken into consideration, defined as the probability that a trip between an origin-destinazion pair is done within a determinate time interval (Bell e Iida, 1997). Several measure methodologies have been proposed (Asakura and Kashiwadani, 1991, Inouye, 2003) to evaluate travel time reliability. Starting from the definition of network weakness (D'Este and Taylor, 2003), in this paper it has been developed a model that, taking advantage of the properties of a recently proposed assignment model (Russo e Vitetta, 2003), spreads out the concept of connective robustness recently proposed by the authors (Di Gangi and Luongo, 2004) and allows the definition of quantitative measures able both to evaluate the reliability of connections among the several origin-destination pairs and to quantify the level of potential exposition due to the vulnerability of the network. The proposed methodology can be adopted, once defined a risk scenario, both to select where to plan necessary adjustments on transport system and to define adequate operations for evacuation and to come to the aid of population, in order to reduce exposition. Results of an experimentation concerning the application of the proposed methodology at a regional scale on a site located in southern Italy are also reported. For the covering abstract please see ITRD E135207.
Samenvatting