Firstly we introduce emergent phenomenon in complex systems and the basic principle of emergent computation. Secondly an important emergent computation model—cellular automata (CA) is demonstrated, and the relationship between emergent computation and swarm intelligence is illustrated. Then we discuss some key themes such as the mapping from source problems to emergent computation models, self-organization and synchronization. At last we review the application of emergent computation in aspects of engineering computation and social computation.

This survey summarizes the mechanisms driving the formation and evolution of online social networks from the multidisciplinary perspective, with emphasis on homophily and preferential linking. The survey elaborates the advances made recently in the research of these mechanisms, reveals the reasons behind various mechanisms and the impact of these mechanisms on network evolution, points out the potential application of these mechanisms, and clarifies the common misunderstanding for the formation mechanisms of online social networks. Finally this survey summarizes the research significance, including theoretical and application value, and future research directions on the formation mechanisms of online social networks.

This paper introduces the concept of operation loop, and a network model of combat networks of Weapon System-of-systems based on operation loop. Then the concept of directed natural connectivity is proposed and is used to investigate the robustness of combat networks. To verify the feasibility and effectiveness of directed natural connectivity in evaluating the robustness of combat networks, three scenarios are designed: edge destruction, edge increasing and connection probability increasing. The robustness of combat networks measured by spectral radius of Laplacian matrix and directed natural connectivity are compared under the three scenarios. The results suggested that the directed natural connectivity allows a sensitive and precise quantitative analysis for the robustness of combat network. It is also showed that robust of combat networks will decrease dramatically if intelligence reconnaissance and acquisition capabilities of sensor nodes are restrained. Our results may lead to useful insights on designing a more robust combat networks, such as improving the accuracy of intelligence acquisition, ensuring instructions delivered accurately, strengthening the capability of precision strike, and enhancing communication capabilities of armaments etc.

We investigate the validation of the threshold predicated by the quenched mean-field theory. By using quenched mean-filed theory, we analyze SIR epidemic spreading models on ER networks, SF networks, and real networks. We find that when the eigenvector of the leading eigenvalue of the adjacent matrix is delocalized, the threshold predicated by this theoretical approach can basically fit its real threshold, and that once the leading eigenvector is localized, the theoretical threshold is lower than its actual threshold.

In this paper, we establish the evolution mechanism of the circuitness of the urban traffic network, and obtain the different topologies of networks. Then, we study the traffic carrying capacity under different network topology structure. As a result, with the increasing network traffic flow, it is much easier to be saturated for the smaller and larger cirucitness. On the contrary, the network has the largest carrying capacity when the circuitness is about 0.54.

In order to accurately describe the exit-selection of pedestrians, an exit-selection model is proposed and simulated. Three factors, i.e. distance, density and exit width, are studied in the model. A two-exit example is studied and the results show that the evacuation pedestrian number is decided by the pedestrain number if only the distance was considered, on the other hand, the evacuation pedestrain number is blanced effectively and the dynamic selection of pedestrain can be achieved if the density factor was considered. Under the condition of two unequal exit widths, the three factors must be considered in the model. Simulation results the provided model is also appropriate for multi-exit and other conditions as well as single exit condition.

Three types of connectivity models are designed from the perspective of Wechat information dissemination in this paper. It is shown by calculating propagation speed, cost, efficiency and privacy based on undirected network that the balance and optimization between the propagation efficiency and safety of Wechat can be achieved using the different connectivity models under different mechanism designs.

Based on the complex network theory, a complex network evolution model is set up on the basis of the distance and the attraction between ports. We select relevant data of 25 world main container ports in 2010, compared evolution network, practical network and the same size random network and verify the validity of the model. Then apply the model to the world shipping network under the arctic route, the results showed that the average path length decreases suggests that the open of arctic routes strengthen the contract between ports; degree value of part of northeast Asia and northwest Europe ports increase obviously; the value of k-core of ports tend to be more polarized. It suggests that the arctic route can enhance the network of gradation of shipping.

To verify avalanche dynamics of open source community, the paper offers an empirical study on SourceForge.net, which is the largest open source community in the world. It researches risk transfer between projects and the cascading phenomenon, finds there are two kinds of mechanisms, through which to transfer risk and trigger an avalanche, they include technical co-relation and cooperative co-relation. We measure the distribution of the avalanche size on the Sourceforge.net. The results show that the avalanche sizes follow power law distributions and the maximum size is only 18. However, based on the analysis of the dynamic evolution results of the cross section data, we find that the maximum massive avalanche size continues to get bigger, which means that Sourceforge.net is evolving into the growth stage.

In order to more truly reflect the competition and cooperation behavior in viral marketing, based on the classic propagation model SIS, a propagation model with partial cross-immunity is proposed. The model describes the propagation of two viruses which enjoy both cooperative and competitive relationship and have different propagation probability. Further, we in depth study two kinds of viral marketing information propagation characteristics with different network structure. Simulation results show that there is a phase transition: If the competition is harsh, then we can get the same results as classic ecology studies i.e. ‘winner takes all’; otherwise, the weaker information will survive. And the higher the probability of cooperation is the greater scale the weaker virus survives in. Simulation results also show that the weaker virus can survive only when it cooperates with the stronger one. Heterogeneous distribution of degree is conducive to cooperation.

Aiming at exploring the characteristics and internal laws of armament organization, taking the United States Army armament organization in a historic period for example and using the definition dimension in the fractal theory, this paper sets a δ-covering of major armaments as helicopters, tanks and infantry fighting vehicles of the United States Army during that period. Thus, with the information dimensions of the three types of armament organization determined and checked out respectively, this paper quantitatively verifies the conjecture that the armament organization is of fractal structure and suggests a mathematical method to achieve the figures of the information dimension. It shows that information dimension can quantitatively illustrate the complex armament organization’s similarities within itself. This paper provides an approach to improve the reasonability of armament organization.

Considering that public opinion is evolving in the spreading process, this paper proposes a new opinion spreading model, which takes opinion revision into consideration. We believe that the opinion of one node has some probability to be modified. And thus we study the influence of source node’s status and the factor of opinion revision to network opinion spreading. The experimental results indicate that: whether the nodes’ opinion of a network at the final state relate to source node’s opinion or not depends on the network nodes’ opinion revision rate and the opinion strength of source node. Besides, for the real Facebook network, this not only depends on the network nodes’ opinion revision rate and the opinion strength of source node, but also depends on the degree of source node. In addition, the experimental results also indicate that the opinion of source node with high degree is propagating faster than that of source node with low degree.

In order to further research on agricultural price volatility and the transmission in our country, an agricultural prices synchronous transmission complex network model is built based on the method of symbolic dynamics, which Nonglutinous rice, Japonica rice, Wheat, Maize and Soybean are selected as 5 kinds of commodities in bulk farm-products. Network fluctuation, strength distribution, cluster coefficient and the dependency between strength and cluster coefficient are studied in our paper. And empirical analysis is given based on price data during April 1997 to December 2012 in our country. At last, countermeasures and suggestions about price adjustment and control in our country are proposed.

In order to better predict failure rate of urban rail trains, the paper proposes a fault predict model based on PSO-BP neural network. Using Matlab with Newff function, failure rate prediction of urban rail train is modeled and simulated using the error back propagation (BP) neural network and BP neural network based on particle swarm algorithm optimization (PSO-BP).The simulation results show that the effect of failure rate predictive model with PSO-BP neural network is better than BP neural network. As a result, PSO-BP neural network is chosen as the failure rate predictive model and the theoretical basis for the decision of preventive maintenance.

For decoding information contained in the cochlea neural networks responses to speech signals, it is interesting to address which parts of input stimuli are more efficient. In this paper, the stimulus-specific information associated with a particular stimulus will be adopted to study the decrease of average uncertainties, and its calculation method is developed. We use a leaky integrate-and-fire model to capture the responses of cochlea neurons to the input speech signal, and calculate the stimulus-specific information caused by each speech signal part. It is shown that the weighted average of stimulus-specific information over the stimulus ensembles yields the mutual information, and the stimulus-specific information is also useful in clearly indentifying the stimuli that are significantly efficient to the cochlea neural network. Moreover, the stimulus-specific information can not only determine which signal component mainly contributes to the mutual information, but also confirms the availability of internal noise in the neural networks. There is a non-monotonic relationship between the noise intensity and the maximum stimulus-specific information. These results indicate that the applicability of the integrate-and-fire neuron model for current cochlear implant decoding technology deserves to be further investigated.