Keywords: infrastructure systems, adjacency matrix, Spearman, rank correlation, regression equations, critical infrastructures, transportation networks, transport infrastructure, network size, network disruptions, highways, bridges, roads, graph theory, network performance, attacks
Effect of transportation infrastructure network size on its performance during disruptions
This paper focuses on the effect of the size of a transportation infrastructure system network on its level of performance when disrupted. The average shortest path distance forms the basic property by which this study is carried out, since it also constitutes the basic unit around which the adjacency matrix is developed. In this paper, the main system studied comprises critical infrastructures (namely the highways and bridges in Newcastle County, Delaware), which is then developed into a network using graph theory. Two case studies are analysed, each selected from two different regions of the county, under the effect of two node removal strategies representing the real-life disruptions of natural and intentional attacks. Based on Spearman's rank correlation and the calculated average shortest path distances, it was shown that the smaller the size of a network, the shorter the path lengths and the more efficient the system. The other considered properties include the network density, average vertex degree, clustering coefficient, node betweenness, link betweenness and network connectivity to determine whether they showed a direct relationship with efficiency. Only the average shortest path distances and betweenness centralities showed a direct relationship with network efficiency. However, this work investigates how the presence of short path distances in a system network affects its performance and develops working equations for computing efficiency using regression equations.