Robustness and resilience analysis of urban road networks

This thesis investigates the robustness and resilience of urban road networks (URNs) in the presence of either local or global disruptions. The problem is approached from the two perspectives of complex network theory and traffic network modelling. The former mainly considers the topological characteristics of the networks, while the latter explores the flow characteristics, demand distribution, user behaviour, control mechanisms, and their combined effect on the network’s performance under disruption. Building on complex network theory, the thesis conducts a topological analysis of URNs using topological indices such as clustering coefficient (CC) and weighted betweenness centrality (WBC). The relationships among indegree, outdegree, WBC and CC are also examined. Following this, the small-world property and community detections for URNs are conducted. The topological indices are then examined in detail in conjunction with operational indices that take into account link capacity, travel demand and driver behaviour. The potential correlations between these various indices are investigated based on their importance for the nodes/links in the network. The thesis also proposes a new relative area index (RAI) quantitatively to analyse the robustness of URNs afflicted by global capacity disruptions, and this index may shed light on the planning and management of URNs. In the second approach, the thesis employs day-to-day evolutionary dynamics to capture the transient state of URNs in the event of disruption. Both agent-based and continuum modelling approaches are considered for the network simulation, employing novel concepts such as percolation theory to quantify the network performance. Different network control measures, such as variable message signs (VMS) and adaptive signal control are incorporated into the day-to-day models in order to mitigate the congestion and delays caused by unexpected disruptions. Two key performance indicators (KPI), rapidity, and the new relative area index (RAI), are derived to represent and quantify the resilience and robustness respectively. Extensive simulation studies are conducted to assess the performance of networks equipped with various control mechanisms and under different levels of degradation.Open Acces