Modeling highway networks with path-geographical transformations

Abstract

A model of highway networks is proposed which is based on the generalization of the concept of geographical networks to incorporate several of the intermediate towns found between two main localities. This model is validated with respect to the US highway network by comparing a large number of topological measurements extracted from that structure with respective measurements obtained from ensembles of networks produced by the proposed model as well as by more traditional theoretical models of complex networks. An optimal multivariate statistical method, namely canonical analysis, is applied in order to reduce the high dimensionality of the measurements space to allow visualization as well as redundancy reduction and enhanced stochastic sampling. Maximum likelihood decision theory is then applied over the reduced measurement space in order to identify the best models. The results corroborate that the currently proposed model allows the best adherence, among all the other considered models, to the original US highway network. © 2009 Springer-Verlag Berlin Heidelberg.