An information theoretic approach to network tomography

In this article, we formulate an information theoretic approach to information recovery for a network flow transportation problem as an ill-posed inverse problem and use nonparametric information theoretic methods to recover the unknown adaptive-intelligent behaviour traffic flows. We indicate how, in general, information theoretic methods may provide a solution to the ill-posed inverse information flow problems, when a function must be inferred from insufficient sample information. As an application, we examine a data set which comprised traffic volumes at Bell Labs.