A Non-equilibrium Approach to Scale Free Networks

Many processes and systems in nature and society can be characterized as large numbers of discrete elements that are (usually non-uniformly) interrelated. These networks were long thought to be random, but in the late 1990s, Barabási and Albert found that an underlying structure did in fact exist in many natural and technological networks that are now referred to as scale free. Since then, researchers have gained a much deeper understanding of this particular form of complexity, largely by combining graph theory, statistical physics, and advances in computing technology. This dissertation focuses on out-of-equilibrium dynamic processes as they unfold on these complex networks. Diffusion in networks of non-interacting nodes is shown to be temporally complex, while equilibrium is represented by a stable state with Poissonian fluctuations. Scale free networks achieve equilibrium very quickly compared to regular networks, and the most efficient are those with the lowest inverse power law exponent. Temporally complex diffusion also occurs in networks with interacting nodes under a cooperative decision-making model. At a critical value of the cooperation parameter, the most efficient scale free network achieves consensus almost as quickly as the equivalent all-to-all network. This finding suggests that the ubiquity of scale free networks in nature is due to Zipf's principle of least effort. It also suggests that an efficient scale free network structure may be optimal for real networks that require high connectivity but are hampered by high link costs.