Network routing in a dynamic environment

Abstract

Recently, there has been an explosion of work on network routing in hostile environments. Hostile environments tend to be dynamic, and the motivation for this work stems from the scenario of IED placements by insurgents in a logistical network. For discussion, we consider here a sub-network abstracted from a real network, and propose a framework for route selection. What distinguishes our work from related work is its decision theoretic foundation, and statistical considerations pertaining to probability assessments. The latter entails the fusion of data from diverse sources, modeling the socio-psychological behavior of adversaries, and likelihood functions that are induced by simulation. This paper demonstrates the role of statistical inference and data analysis on problems that have traditionally belonged in the domain of computer science, communications, transportation science, and operations research.