Much recent research activity has been devoted to empirical study andtheoretical models of complex networks (random graphs) with three qualitativefeatures: power-law degree distribution, local clustering of edges, and smalldiameter. We point out a new (in this context) platform for such models -- thestochastic mean-field model of distance -- and within this platform study asimple two-parameter proportional attachment model. The model ismathematicallly natural, permits a wide variety of explicit calculations, hasthe desired three qualitative features, and fits the complete range of degreescaling exponents and clustering parameters; in these respects it comparesfavorably to existing models.
展开▼