Understanding Resiliency and Vulnerability in Transport Network Design
C Constantinou, The University of Birmingham, UK; K Baughan, Ideas Network Ltd, UK; A Stepanenko, Aston University, UK
How resilient is a particular transport network design? How vulnerable is it to planned or unplanned disruption at different points in the network? How can you compare the positive benefits to its resiliency of two alternative proposals for its enhancement?
We all have an intuitive understanding of resiliency. It is straightforward to identify a transport network?s single point of failure - where disruption to the traffic flow would cause the complete loss of one or more routes across the network. Similarly, it is relatively easy to recognise the presence of alternative diverse paths through a network which would provide a route with a basic level of resiliency against single points of disruption. What is missing though is a mathematical framework with which we can understand the full depth of diversity present in a complex real world transport network made up of a great number and variety of paths for each possible route and a great number and variety in the number of points of potential disruption.
The authors of this paper propose an innovative new approach to network analysis that enables the full depth of diversity inherent in any bi-directional transportation network to be identified and studied on a rigorous, systematic basis. The beauty of the approach is that it creates both a 3D graphical visualisation of the transport network that provides an immediate qualitative view of its resiliency, as well as a powerful mathematical framework for detailed quantitative analysis. The techniques are as applicable to the flow of people across the London Underground as they are to the flow of cars through the road network of Birmingham or to the flow or airplanes over Europe. Armed with a more detailed understanding of network resiliency it is then possible to make much smarter choices whether you are a designer contrasting alternative expansion proposals, an operations manager planning routine maintenance activity across the network or a security specialist seeking a better understanding of a network?s vulnerability.
The techniques were originally validated through research activities funded by the US Air Force Research Labs looking at the resiliency of data communication networks. Since completing this original work the authors have looked at ways of expanding the analysis to other forms of network - exploring the similarities between data packets coping with outages in a communication network, drivers coping with traffic jams in a road network and commuters coping with line closures on a metro transport network.
The mathematical basis for the authors? approach to network resiliency analysis is based on two fundamental notions that can act as atomic descriptions of diverse connections. The first such notion is that of a cycle of vertices ? or more simply put, a ring of interconnected nodes in the network. The second notion is of a shared edge between two cycles ? or more simply put, a series of two or more interconnected nodes that are common to two neighbouring rings. The truly innovative step has then been to perform a recursive abstraction of a network in terms of these two atomic units in order to create a hierarchical representation of the full depth of diversity present in the network under study ? a process named logical network abridgement.
The paper will provide an easily understandable overview of the mathematical basis for the method of analysis and then move on to look at examples of how it might be applied to a range of problems in transport network design, operation & security.
Association for European Transport