Xiaofeng Gong

The development of generalized synchronization on complex networks

In this paper, we numerically investigate the development of generalized synchronization (GS) on typical complex networks, such as scale-free networks, small-world networks, random networks, and modular networks. By adopting the auxiliary-system approach to networks, we observe that GS generally takes place in oscillator networks with both heterogeneous and homogeneous degree distributions, regardless of whether the coupled chaotic oscillators are identical or nonidentical. We show that several factors, such as the network topology, the local dynamics, and the specific coupling strategies, can affect the development of GS on complex networks.

Optimal resource allocation for efficient transport on complex networks

The problem of efficient transport on a complex network is studied in this paper. We find that there exists an optimal way to allocate resources for information processing on each node to achieve the best transport capacity of the network, or the largest input information rate which does not cause jamming in network traffic, provided that the network structure and routing strategy are given. More interestingly, this achievable network capacity limit is closely related to the topological structure of the network, and is actually inversely proportional to the average distance of the network, measured according to the same routing rule.

A coevolving model based on preferential triadic closure for social media networks.

The dynamical origin of complex networks, i.e., the underlying principles governing network evolution, is a crucial issue in network study. In this paper, by carrying out analysis to the temporal data of Flickr and Epinions–two typical social media networks, we found that the dynamical pattern in neighborhood, especially the formation of triadic links, plays a dominant role in the evolution of networks. We thus proposed a coevolving dynamical model for such networks, in which the evolution is only driven by the local dynamics–the preferential triadic closure. Numerical experiments verified that the model can reproduce global properties which are qualitatively consistent with the empirical observations.

From sparse to dense and from assortative to disassortative in online social networks

Inspired by the analysis of several empirical online social networks, we propose a simple reaction-diffusion-like coevolving model, in which individuals are activated to create links based on their states, influenced by local dynamics and their own intention. It is shown that the model can reproduce the remarkable properties observed in empirical online social networks; in particular, the assortative coefficients are neutral or negative, and the power law exponents γ are smaller than 2. Moreover, we demonstrate that, under appropriate conditions, the model network naturally makes transition(s) from assortative to disassortative, and from sparse to dense in their characteristics. The model is useful in understanding the formation and evolution of online social networks.

ANALYSIS OF TRAFFIC FLOW ON COMPLEX NETWORKS

We propose a new routing strategy for controlling packet routing on complex networks. The delivery capability of each node is adopted as a piece of local information to be integrated with the load traffic dynamics to weight the next route. The efficiency of transport on complex network is measured by the network capacity, which is enhanced by distributing the traffic load over the whole network while nodes with high handling ability bear relative heavier traffic burden. By avoiding the packets through hubs and selecting next routes optimally, most travel times become shorter. The simulation results show that the new strategy is not only effective for scale-free networks but also for mixed networks in realistic networks.

A spectral algorithm of community identification

A novel spectral algorithm utilizing multiple eigenvectors is proposed to identify the communities in networks based on the modularity Q. We investigate the reduced modularity on low-rank approximations of the original modularity matrix consisting of leading eigenvectors. By exploiting the rotational invariance of the reduced modularity, near-optimal partitions of the network can be found. This approach generalizes the conventional spectral network partitioning algorithms which usually use only one eigenvector, and promises better results because more spectral information is used. The algorithm shows excellent performance on various real-world and computer-generated benchmark networks, and outperforms the most known community detection methods.