ABSTRACT
The central paradigm of network analysis relies on the fact that, the essential information to explain the functionality and the behavior of individuals within a complex network or the whole network itself is contained in its abstract structure, that is, the underlying graph model. Based on this assumption and reflecting the recent advances in graph theory and the development of specialized computer software, the contemporary analysis of complex networks (see [33] for a general introduction) uses a wide variety of quantitative methods for modeling, measuring their characteristics and visualization. Along with determining statistical properties, the global structure, or the detection of network communities, one of the basic objectives is the identification of the most important objects within a network. The common way to express the importance of network objects is to quantify it by evaluating a specific centrality index on the vertices of the graph representing a given network, where the vertices with the higher values of centrality are perceived as being the more important. The centrality index, as a real-valued function on the vertices of a graph, is a structural descriptor, that is, it is invariant under graph isomorphism.
