In the paper, an equipment contribution degree evaluation method based on combat network of weapon system-of-systems (WSOS) is proposed. Firstly, weapons and armaments are abstracted as nodes while the complicated relationships between weapon and armaments are abstracted as edges of combat network. The nodes and edges are categorized and related mathematical models based on operational indicators are built.For a better evaluation of WSOS operational effectiveness, a comprehensive evaluation index based on operation loop is put forward. Besides, an equipment contribution degree evaluation model is established to weigh the equipment contribution within the WSOS. Finally, a theater missile defense system-of-systems is taken as a case study to verify the feasibility and effectiveness of the proposed methods. The results show that the decision organs play a most important role in the WSOS.

Aiming at the problem that information network with single factor pattern can not reflect effectively the complex structure and dynamic process for operational information flowing (OIF), this paper presents a new method of modeling and characteristic analyzing for OIF model with double layers and three networks based on super-network. Firstly, super-network structure characteristics of OIF are analyzed, multidimensional attributes for operational nodes and information relationships are abstracted, association mapping rules are defined, and then OIF super-network model is built. Secondly, from the perspective of super-degree distribution for nodes, degree distribution for super-edges, betweenness, convergence coefficient, sub-graph centrality and network resilience, topological characteristics of OIF super-network are studied. Finally, taking aviation aggressive operation as an example,and simulation results validate feasibility and rationality of the model and method.

The structure of social networks is of paramount importance in collective behaviors,e.g. information propagation, consensus and formation of social norms. In this paper, a special network is constructed by adding remote links among nodes over lattice graphs with total energy constraints.A power law distribution is used to model the relation between the link probability and the distance.We study the effect of geographic scaling property on the dynamics of Naming Game with a group interaction rule. We find that there exists an optimal parameter value which minimizes the time to converge to global consensus. When the total energy constraint is large enough the optimal parameter value is approximately 1.5. Numerical simulations indicate that the geographic scaling property in social network plays an important role in the emergence of social collective behavior and rules.

Based on the hypothesis that community structure determines the network connections, the connection probability matrix which describes the nodes’ community structure can be transfered into the form of product of matrices. The nonnegative matrix factorization is applied here to get the connection probability matrix and then obtain the reconstruction. Experiments on several real world datasets show that the proposed algorithm outperforms some other algorithm which are based on similarity indexes.

The node degrees, weighted node degrees, node hyperdegrees, hyperedge degrees, hyperedge hyperdegrees, average distance and clustering coefficient are proposed in the paper. TV programs are defined as nodes and broadcasting time periods are defined as hyperedges. By using hypernetwork analysis of television programs competitive relationships, we find that the cumulative probability distributions can be described by an exponential distribution. It shows that random factors result in the formation of the hypernetwork. The competition of the supernetwork can be better described by weighted node degrees. The average distance is small and the clustering coefficient is large. These parameters conform to the characteristics of small-world network. These topological characteristics may be useful for the studies of competitive hypernetworks. The methods proposed can also be used for other empirical studies.

This paper proposed a distributed and hierarchical scheme for formation control of multiple robotic fish. Firstly, each robotic fish takes consensus algorithm as a distributed data fusion modular to estimate the center of the formation, and then calculate its desired position in the formation at the next time. Such estimated results are the input variables of its path planning modular. Secondly, each robotic fish tracking its desired position by employing the fuzzy controller of linear speed and angular speed in order to form and maintain formation. The proposed method separates the distributed estimation and the path planning and behavior of robotic fish, such a property gives a new method for the cooperative control of multiple heterogeneous robotic fish. The results of experiment shows the validity of the proposed method.

Based on heterogenous threshold model and mean-field theory, we analyse the impact of neighbor effect, threshold distribution and degree distribution on the diffusion process and equilibrium. The result shows that the decrease in neighbor effect or the mean of threshold distribution would speed up the diffusion process and increase the equilibrium value. Besides, the decrease in the variance of threshold distribution would slow down the initial diffusion process, and would also speed up the end-stage diffusion when some conditions are given; if the neighbor effect is low, the increase in the heterogeneity of degree distribution would speed up the initial diffusion stage; the initial diffusion would experience a super-exponential increase stage when a certain relationship between the neighbor effect and the threshold distribution is met; the change in neighbor effect, threshold distribution or degree distribution may lead to a jump increase of the equilibrium value.

The Koch Fractal Island, which is starting from a regular polygon, is mapped to complex evolving Koch networks. The informative labels are given to nodes, the labels are based on the time and location when nodes are accessing to Koch networks. By the advantages of the informative labels, we get the exact solution of main structural properties of Koch networks, including degree distribution and cumulative degree distribution function, as well as the clustering coefficient, average shortest path length and the correlation function of degree, betweenness centrality and the shortest path routing and length. The results show that, Koch network is a scale-free and small-world network; its clustering coefficient tends to relatively large constant; average shortest path length is proportional to the logarithm of the size of networks; degree correlation function is exponential function relationship with node's degree.

Social network analysis has been used to analyze the collaborative network of Chinese film directors and actors for the past decade. We apply two-mode network analysis method to collaboration data of movie directors and actors and compared collaborative relationship between directors and actors in Hong Kong and Mainland China, as well as its changing. Some group characteristics are revealed from our data. The main findings of this paper are: (1) there exists significant difference in the structure of collaborative network between Hong Kong and Mainland China. HK directors are more likely to cooperate with the same actors repeatedly than mainland directors. HK directors show a trend of more and more cooperation with mainland actors in the past decade. (2) The collaborative network of directors and actors is very sparse, with density of only 0.026. Component analysis shows that HK directors and actors are at the core in the network while the mainland directors and actors are at the periphery in the network. (3) Block analysis shows that there indeed exists “fractions” in this network, that is directors and actors who are in the same area are more likely to cooperate with each other.

Based on improved Greenberg-Hastings cellular automaton model, this paper built a complex neural network model and studied the effects of reconnection probability p on the ability of penetrating defect of spiral wave under two kinds of classical small world network. It shows that when p is less than or equals a fixed value p_c, the ability of penetrating defect of spiral waves will markedly enhance with the increase of reconnection probability p; When p is greater than the p_c, the ability of penetrating defect of spiral waves will not enhance with the increase of p.

This paper first reports that the network topology of the cyber physical system of the nickel–cobalt acid leaching process is a scale-free, hierarchical, small-world network. The network topology is formed by mapping the processing units, valves, meters, actuators, and controllers as nodes and the communication cables and material stream pipes as edges. We speculate that this result may be valid for other complex processing system.

In this paper, the research combind with the PBO environmental characteristics to build multi-population dynamic evolutionary game model about knowledge transfer based on PMO as knowledge broker, to get the evolutionary stable strategy and evolutionary stable diagram. The Research shows that knowledge spillover effects, knowledge collaborative innovation effects, the cost of knowledge transfer and the rewards or punishment of knowledge transfer, are the main factors influencing the knowledge transfer behavior and determining the overall trend of knowledge transfer in PBO.

A planar rigid rod-spring pendulum model was constructed and dimensionless dynamic equation was given. We numerically simulated and analyzed the dynamical behavior of the two-time-scale system while the frequency ratio and length ratio and initial conditions vary. The dynamic equation is strongly nonlinear as the fast and slow variables couple each other. A cubic interpolation precise integration method was applied to solve nonlinear dynamic equation. We also employ the Poincaré maps and the maximum Lyapunov exponent methods. Numerical simulation results demonstrate that the system presents complex chaotic motion in different parameters conditions. It is also found that the fast variable may transform to chaos via the quasi-periodic torus breakdown.

Ratio and triangle ambiguities closure error (TACE) tests are two primary validation methods used in ambiguity resolution for long-range reference station network. once they are passed, it is considered the ambiguities could be fixed to integers with a high confidence. However, it is usually found that the ambiguities were wrongly fixed even both the two tests above are passed. To improve the correct rate of ambiguity solutions, we proposed a network failure probability (NFP) test, the failure probability of the reference station network is calculated under the conditions of that the Ratio test and TACE test have been passed. The experiment result shows that, with additional NFP test, the average correct rate of integer ambiguity solutions is raised up by 8.1% with a peak at 23.5%, the average initialization speed is increased by 5.6% as well when the number of the reference stations is smaller than 8.

In order to investigate possible distribution pattern of strategies of agents in EMG model distributed on a two-dimensional lattice, we put an evolutionary minority game model on 41x41 two dimensional square lattice network.We consider two situations which include unidirectional network and bidirectional network. Simulation results show no matter unidirectional or bidirectional network, the connectivity degree of nodes (for bidirectional network, it means out degree) is predominant at medium numbers. The distribution of agents' strategies and structure of network reach orderly or regular stable distribution structure after evolution. When network is unidirectional, some nodes form bulked community like structure, but there is no isolated node, degree distribution is relatively even. When network is bidirectional, there is no apparent community like structure, but there are some small communities or even isolated node, degree distribution region is somewhat larger, but still be fairly even.