Road network vulnerability analysis with consideration of probability and consequences of disruptive events

Abstract

To assess the vulnerability of road networks, the commonly used analytical vulnerability analysis is to determine the global vulnerability of links in a road network by removing links one by one from the network and measuring the resulting increase in the total travel cost of the network. Such global vulnerability ranking might fail to identify the most critical links as it overlooks important factors that affect the vulnerability of road links. Instances of such factors include link specific geometry design, poor downstream traffic signal timing, and/or links that are prone to more collisions. Additionally, traditional techniques identify the critical links of a transportation network measuring only the consequences of the link closure with little consideration given to its closure probability. Consideration of the probability of link closure or failure is important as some of the links in a transportation network are more susceptible to disruptive events than others. To fill the void in the literature, I propose two data-driven vulnerability approaches: 1) vulnerability analysis by modeling monthly and seasonal extreme travel delay variations and 2) vulnerability analysis by measuring the spatio-temporal impact of incidents. In studying road network vulnerability by modeling monthly and seasonal variation of extreme travel delay, I propose a new class of extreme value distribution called compound generalized extreme value (CGEV) distribution for depicting the monthly and seasonal variation in extreme travel delays in road networks. Since the frequency and severity of extreme events are highly correlated to the variation in weather conditions as an extrinsic cause of incidents and long delays, monthly and seasonal changes in weather contribute to extreme travel time variability. The change in driving behavior, which itself varies according to road/weather conditions, also contributes to the monthly and seasonal variation in observed extreme travel times. Therefore, it is critical to model the effect of monthly and seasonal changes on observed extreme travel delays on road networks. Based on the empirically revealed linear relationship between mean and standard deviation (SD) of extreme travel delays for both monthly and seasonal levels, I formulated two multiplicative error models. I then obtain the CGEV distribution by linking the two multiplicative error models and formed a compound distribution that characterizes the overall variation in extreme travel delay. I calibrated the CGEV distribution parameters and validated the underlying assumptions that are used to derive the CGEV distribution using multi-year observed travel time data from the City of Calgary road network. The results indicate that accounting for the seasonality by identifying seasonal specific parameters provides a flexible and not too complex CGEV distribution that is shown to outperform the traditional GEV distribution. Finally, I evaluated the application of the proposed CGEV distribution in the context of road network vulnerability taking into account the stochastic nature of extreme event occurrences and the link importance. This derived data-driven vulnerability index incorporates a wealth of information related to both network topology in terms of connectivity and the dynamic interaction between travel demand and supply. In studying road network vulnerability by measuring the spatio-temporal impact of incidents, I propose a new data-driven, impact area, vulnerability analysis approach that takes into consideration both the probability of impact as well as the effects of incidents on the impact area. I employed multi-year observed travel time and incident data to investigate these underlying dynamics as the datasets contain important information that reflect the historical spatial and temporal occurrences of link closure and their network wide impacts. Rather than focusing solely on the travel time fluctuation of the link subject to incident, and to capture all aspects of incidents’ impact, I developed a new approach to identify both the resulting spatial and temporal impacts by monitoring the dynamic propagation and dissemination of congestion patterns in the set of links that are in the vicinity of the link subject to incident (i.e., impact area). I subsequently used these spatial and temporal dimensions of the impacts in the vulnerability analysis. I examined the performance of the developed approach, historical travel time and incident data of the City of Calgary. The results indicate that the recorded temporal impact of incidents is not representative enough of the true impact of incidents since it overlooks the dynamic spatial propagation of the effect of incidents on the impact area. Finally, I used the estimated spatio-temporal impact of incidents in Calgary road network to determine the vulnerability of the links considering both the probability that links in an impact area are affected by an incident as well as the spatio-temporal consequence of the incident on the impact area. This data-driven vulnerability measure could be used as a decision support tool for decision-makers in prioritizing improvements to critical links to enhance overall network vulnerability, reliability, and resilience.