Discussion of “Network routing in a dynamic environment”
Andrew C. Thomas, Stephen E. Fienberg · The Annals of Applied Statistics · 2011
An earlier version of Professor Singpurwalla's paper (which we refer to as "Singpurwalla") has served as the springboard to our own investigation of the issue of deployments of Improvised Explosive Devices, or other obstacles with a large cost to overcome, which may be placed stochastically, or by an adversarial agent, or both.Rather than a decision-theoretic treatment, we consider a method based in part on social network analytical methods, namely, that the deployment pattern of IEDs induces a subgraph on a full road network, and that the deployment on any given road is unknown to anyone traversing the graph until arriving there, though there may be prior information on the likelihood of a deployment.The full treatment, as acknowledged in Singpurwalla, is illustrated by Thomas and Fienberg (2011); here, we give a brief overview of our method and how it compares with Singpurwalla's approach.