Geographically-sensitive network centrality and survivability assessment

Network survivability deals with theories and practices involved in preserving essential services of a network after accidents and/or attacks. Major catastrophes such as hurricanes and weapons of mass destruction are expected to introduce patterns of faults on a network that are strongly geographically correlated. In this work we emphasize the interplay between geographic information about the network and the established graph-theoretic measures of network connectivity and performance. In particular, we propose a method of randomly generating networks that permit geometric constraints, and we update classical graph centrality measures to reflect geographic proximity. Moreover, we simulate geographically guided fault patterns in the network and we study the effect of such faults on the structural as well as the steady-state functional performance of the network.