On the Rationality and Optimality of Transportation Networks Defense — a Network Centrality Approach

Transportation infrastructures have recently gained increasing attention in the context of homeland security. Being both a main target for attacks as well as a method for carrying out such attacks, much effort is being allocated these days towards increasing our understanding regarding transportation networks [14]. Specifically, measuring and predicting human mobility patterns along the links of a transportation network has been of a great importance to researchers in the field, as it contains the basic information needed in order to cope with transportation related threats more efficiently. Such threats can take for example the form of a group of terrorists trying to reach their target by car, or a truck filled with chemical or radioactive material. These threats require homeland security agencies to rapidly deploy monitoring or surveillance units in key junctions, dispatch air units to central locations etc. Clearly, carrying out this mission relies on the knowledge of what are those key traffic junctions, and how to deploy the existing (and always on shortage) resources most efficiently. Hitherto, producing the transportation data required for answering these questions was done off-line and relied heavily on expensive and time consuming surveying and on-field observational methods. Network Betweenness is known to be highly correlated with network load in communication and transportation networks. In this work we show how a specially designed Betweenness Centrality measure that can be useful for optimizing locations of static or mobile monitoring equipment. Furthermore, we show that the accuracy of the estimations produced using this approach can be further enhanced when additional (pre-defined and known) properties of the network are taken into account, generating an augmented Mobility Oriented Betweenness Centrality measure. We demonstrate the efficiency of the proposed method both analytically and experimentally, using real world transportation network constructed using cellular phones data, that contains a high resolution network of the Israeli transportation system. We show that the traffic flow level that were measured using this expensive and complicated method can be accurately estimated using our proposed Augmented Betweenness technique. As a result, we can generate an efficient deployment scheme of monitoring units for this specific network, and calculate the percentage of traffic it monitors.