The China Network Science Forum 2014 i.e.the 10th anniversary were reviewed. It retrospects and summarizes several academic progress in our country, involving D＜3 super small world network, a complex network of command and synergy mechanism, synthetic biology network engineering,the cognition of the era of large data calculation, the controllabity and observability of the complex network, cancer dynamics theory, social network research, network science interdisciplinary exploration, generalized synchronization network structure identification, and network science in mathematical research, etc.Looking to the future, it puts forward top 10 challenging issues.

In this paper, we constructed the super network model of SoS operation according to varied types and interactions of agents in the perspective of physics, information and cognition domains, and a triple-layer-super-network model is composed of command network, information network and belligerent network. In order to investigate the capability of the command and coordination for SoS operation, we discussed the influence of the factors, such as coupling intensity between different domains, topological structure of networks, and then proposed synchronization model based on operation loop. Some total order parameters and local order parameters were used to evaluate the capability of synchronization. At last, experiment has been made by utilizing the simulating data collected from the manoeuvre.

Following the Broadcast-Storage concept, this paper puts forward a dual-structural future Internet scheme that integrates the current Internet architecture with a complementary secondary structure characterized by the combination of broadcast transmission and ubiquitous content storage. Design principles and architectural model of the dual-structural future Internet are introduced, and the main features, physical infrastructures and key implementation mechanisms of the secondary structure are also discussed emphatically. This paper provides a cost-efficient innovative approach to future Internet research.

With reviewing the existing studies on the structural controllability of static networks, we explore the approach to understand and analyze the structural controllability of temporal networks. We firstly reveal the characteristics of the structural controllability of temporal networks by the improved maximum matching method; then, by mapping a temporal into a LTV system, we give a framework to study the structural controllability of temporal networks; finally, on the basis of this framework, we further analyze the upper and lower bounds of network nodes′ controlling centrality in the case of single controller.

Networks are usually used to describe the structural relationships of the componentsof a system. However, for a real system, individuals could interact through many different channels. So, the conception of multilayer networks has been proposed. Generally speaking,multilayer networks explicitly incorporate multiple channels of connectivity, and each channel is represented by a layer and the same nodes may have different kinds of interactions. In this paper, we will give a brief introduction to the structure and robustness of the multilayer networks.

We attempt to discuss the issue from the perspective of socio-technique congruence. To construct the developer network and software network, we collect the source codes and developers′ historical data of 10 consecutive versions from the open-source software project Vuze. Then, we compare the co-evolution of the two networks by introducing some typical parameters, i.e., the scale, average degree, path lengths, clustering coefficient, and community structure. The results show that: the growth of software network is cyclical, about 3 to 4 versions a cycle, whereas the developer network shows a linear increase. Their average degree parameters meet the approximately linear increase with time as a whole. Although the local collaboration in developer network is more significant, the common “small world” characteristic becomes more obvious and their average path lengths decrease regularly. Both networks have a decreased trend of the modularity, but the fluctuation of modularity in developer network is sharper.

This paper reviews the main research literatures which note that the previous collaborative decision-making limited just with ATC authorities and airlines and new solution should be proposed to involve airport managers into the decision-making. The paper presents the recent research using of Deferred Acceptance algorithm from matching technology of the game theory to consider three parts: ATC unit, airlines and airport managers in airport CDM.

In this article, we analyzed users′ mobility patterns and proposed several predictors to solve the next location prediction problem with users′ mobile GPRS data. It is found that individual′s mobile behaviors are bursting, heterogeneous, weakly regular in temporal-spacial aspect, and individuals tend to stay in the same location in a short interval time. Furthermore, based on these empirical results, a blending model is developed to improve the prediction accuracy, overcomingall models with standalone feature.

This paper starts from the seven bridges problem solved by Euler, and extends the principles of topology and graph theory. Then we give a review of the network sciences development from regular networks to random networks and then to complex networks, expounding the connotation of network sciences and discussing the cross with other disciplines. Finally, we point out four aspects of mathematical challenges including the new topological representation theory of complexity, computing technology of million level of sparse networks, dynamic behavior of large scale network structure, and aggregation mechanism of different levels hybrid networks.

In order to reveal some characteristics of supernetworks,in this paper we proposed and established four kinds of three-layer network evolution models based on small-world and scale-free model, and defined two kinds of the layer intersectant degree between the levels, which are used to measure the cooperation-competition relation between network nodes. The numerical analytical results show that for the supernetwork theory model and empirical analysis, the layer intersectant degrees can not only be used to analyze the relationship between different layers of nodes, but also describe and quantify supernetwork robustness, From the theory our research results further deep perfect the multi-layered supernetwork evolution models, and lay a certain foundation for applications.

The structure of social group deeply influences the development and evolution of human society, but studies on this subject are relatively rare. Focusing on QQ friendship network, this paper proposes a percolation-like diffusion model which is based on users′ common interest to simulate and analyze the clustering behaviors of users and the growing process of social groups. Numerical simulation on the real QQ friendship network of Tencent shows that the statistical features generated by our model accord with the real empirical properties of the group network. It indicates that this mechanism is an important driven-factor for the growth of real social group. This work provides vital theoretical evidence for the further studies on the prediction of social group growth.

In this paper, we analyze the committed data of developers in Git repository to investigate the distribution of the numbers of files in different commits and the dependency among the files in each commit. We find that the distribution of the number of files in a commit for each project follows approximately a power-law distribution, and files committed by a developer at the same time are likely to dependent on each other. Such analysis of the dependency between the files committed at the same time can help to design better recommender system to predict the next focus of a developer and thus may improve her work efficiency.

Division of network community contributes to a better understanding of community structure and predict the behavior of complex network, and has great application value in social network, information recommendation and precision marketing, and so on. This paper builds a LR-S hierarchical network, proposed a SA-LEN evaluation function of similarity degree and a SA-LEN community detecting algorithm, based on the two factors that shortest path length of L-IRS networkand activeness of R-RTCS network. This algorithm proved to be effective in nature networks and computer generated networks.

In view of the existing community partition algorithms too granular and the limitation of optimization based on the modularity,this paper proposes a community partition algorithm based on gravity,which the relation between the adjacent vertices is considered as attraction while the relation between the non-adjacent vertices is repulsion. The community structure is formed by the self-organization of vertices which are influenced by the virtual force from their neighbors. Compared with GN and CNM algorithm in the reality of the network which known community structure,the algorithm in this paper has high reliability and nearly linear time complexity.

In this paper, we present the advanced progress of the online social interaction, rating behaviors, hypergraph model construction, node importance identification and the information diffusion of online social systems, as well as the possible future directions, which could shed some light on the corresponding progress.

A brief review about multilayer networks research is given in this paper. The concept, theoretical model and data analysis of multilayer networks are introduced, which help us to draw scientific essence as well as pinpoint practical bottleneck and open questions in the future. As one of very important directions in network science, future studies of multilayer networks should focus on three aspects. First, for concept system, form a standardized and universal research paradigm. Second, develop the theoretical model to utilize the dependence and interaction of multilayer networks rather than sum of exiting models of single network. At last, in empirical research it is necessary to transit from verifying current model based on real data to improve real system via the theoretical model.