For most cities in the world, bus service still acts as an essential part of public transport, though many metropolises has developed more efficient railway system to ensure the rapidly increased demand of travelling. The resilience of urban bus network in face of multifarious disruptions and natural disasters keeps being one of the major concerns of city transport authority and public in high-density cities. Different from the urban rail transport which rely on the exclusive railways in metropolitan area, the urban bus system depends on the road network which are shared by private and commercial vehicles. Such dependency, on the one hand, makes the bus system easily affected by the daily road events such as accidents, congestions and road works. On the other hand, it provides the bus network with extra robustness by enabling the ability of re-routing on the road network. Meanwhile, with the advantage of inherent flexibility of reorganization, the bus network performs better when responding to disruption events. The featured effects brought by the dependency help us develop a deeper understanding of the resilience of bus network from real-world complex network perspective. We investigate the bus network coupled with road network in Hong Kong to reveal the insights of the dependency between bus vehicles and roads in terms of resilience and develop a quantification framework based on four phases of resilience cycle (i.e. preparedness, robustness, recoverability and adaptability), so as to provide decision-makers with analytic assistance for building up resilience.

We build a coupled network model composed of two layers, that is, the bus network G_B=(S, L) with bus stops S and bus links L, and road network G_R=(V, E) with crossings V and road segments E. The dependency is defined as a function that the closure of road segment will cause the direct disruption of bus stops on this road and the blocking of bus route. The resilience of bus network is examined based on four phases through (1) an evaluation on the topological characteristics of network components; (2) simulations of bus network disruption under different modes of attack; (3) simulations of road disruption scenario, during which we adopt different reorganization strategies for recovery; (4) an evaluation of path redundancy on the road network for all bus links.