In this paper, we collect developers′ logs and mailing list information of 10 versions from Tomcat 6 during the time ranges from 2006 to 2010.Then we perform an evolution analysis on developers′ activities. According to the trend of developers′ activities, we extract typical evolutional patterns. Finally, a comparison between developers′ contribution to the specific work and different evolution patterns is made. The results show that five typical evolution patterns exist in the developers network and those developers under different evolution patterns engage different work.

According to the fact that the conventional social network analysis methods did not consider the role of network property, such as social individuals′ position, the strength of social relations. We introduced a new social network model based on extended graph and hierarchical structure based on information flow. Also, we proposed a method for discovering the backbone nodes based on the Network Topology Potential Support and a method for discovering the network hierarchy based on the potential flow. The experimental results show the effectiveness of our method in mining the hierarchical structure of networks.

This paper constructs an artificial stock market which could exhibit the key statistical properties of the real market by agent-based modeling. By describing three types of investors (institutional investor, trend investor and noise investor) which are different from their information endowments and trade behaviors, the model exhibits the power-law distribution of investors′ wealth (in the high-wealth range), and we find that the Gini coefficient of the wealth distribution decreases as the initial share of institutional investors increases. This result implies that the institutional investors who are dominant in information gather market wealth, but this effect is reduced as more investors join their group. From the view of improving equality in resource allocation, we also give some targeting suggestions to the Chinese stock markets.

In this paper, we investigate the effect of the time window on the personalized recommendation algorithm based on ten similarity measures. The experimental results on the benchmark dataset MovieLens indicate that by only adapting approximately 12.56% recent rating records, the accuracy could be improved by an average of 27.17%, and the diversity could be improved by 3.28%. Our work is valuable in both theory and practice, and it could largely reduce the calculation complexity triggered by massive data.

To better explore the universal robustness against cascading failures on complex networks, closely focusing on the load which is the most important physical quantity that can affect the spread of cascading failure, and dynamic process after a node fails, a cascading failure model with tunable parameters is proposed based on the local characteristic of node. With this model we study the cascading failure condition of ER and BA networks, and obtain the formula of phase transition point theoretically. The relationship between the robustness against cascading failures on complex networks and parameters in the model, including the topology parameters, the initial load coefficient, and the redistribution coefficient, is discussed numerically. In addition, theoretical results also are verified by the simulation results of the ER and BA networks.

The topological characteristics and robustness of Chinese high-speed railway and civil aviation compound network are analyzed based on the complex network theory. The results include the following points: the high-speed railway and civil aviation compound network and its two subnetworks are scale-free and small-world networks; the robustness of high-speed railway and civil aviation compound network is better under the random attack than under the calculated attack; the robustness of compound network is better than its two subnetworks,whether under the random attack or under the calculated attack.

To investigate the influence of delayed decision and bounded rationality on complexity of airlines′ price competition, the airlines′ dynamic price competition model based on delayed bounded rationality was established by using the theory of bifurcation of dynamic systems. Theoretic analysis and numerical simulations were provided to study the complex dynamics of this model. The results show that the larger weight of delayed decision can enlarge the stability region of the model and shorten the time of the model to reach the equilibrium points. When the model is unstable, it is more beneficial for airlines to use delayed decision. And once an airline uses delayed decision, the free-rider problem can be shown from the competitors who will get extra profits. It is important to choose a reasonable weight of delayed decision to balance airlines′ market performance.

In order to describe collaborative decision-making process, the command and control model is put forward based on complex information relation integrtion and OODA. The characteristics of individual decision-making are analyzed, the complex information relation is considered as the cause of OODA difficulty to describe collaborative decision-making exactly. From the perspective of armament system of systems combat, the information, data, intelligence and knowledge is defined based on knowledge management and Shannon′s entropy theory. The basic element is confirmed and the information flow way is analyzed which in the physical, information, cognitive and social domains. From the perspective of information, the armament system of systems complex networks model framework is put forward based on the idea of using OODA to guide armament system of systems complex networks model design. The data, information, intelligence and knowledge flow is analyzed and the describing meta-model is put forward. Using the reticular structure of OODA to describe collaborative decision-making idea is proposed, the command and control model is put forward based on the idea, the simulating tool ExtendSim is used for simulation analysis and results show that our work offer a reference for innovate command and control thory.

For adapting to the changing needs of the combat capability model generating, the paper discusses the factal mechanism of combat elements interaction, and demonstrate the combat capability fractal model of morphological evolution including generation, maintenance, change and decay, reveals the fractal nature of combat capability morphological evolution, based on basic principles of the general systems theory and the basic definition of combat capability.Explains the fractal nature and the source of the classic model of combat capability morphological evolution based on statistics or experience, such as Lanchester equation, Osipov equation and the combat capability index model.Combining with an attack instance, quantitatively validated the force forms and degradation of the combat capability morphological evolution.

Aiming at two-level conflict situations with misperception, a modeling and analysis method is proposed based on two-level hierarchical hypergame. According to player′ perception information about strategy sets and outcome preference of other players, conflicts between two-level players from one side and the opponent player are respectively described as a hypergame and several bimatrix games. And then a two-level hierarchical hypergame model(TLHHM) for two-level conflict situations is constructed. The definition of hierarchical hyper Nash equilibrium(HHNE) for TLHHM is given, and the existence condition and solving method of HHNE are developed. An example illustrates the proposed model and method at the end.

The reasons why there are summer transmission peaks in tropical area and subtropical area in east Asia are still unclear. This paper uses an agent-based modeling (ABM), to model influenza transmission dynamics. Our influenza transmission ABM model considers two transmission routes and the model is illustrated by the analysis of Hong Kong influenza transmission pattern. Results show that: 1) the summer peak of influenza transmission in Hong Kong can be caused by virus transmission through the fomite-mediated route; 2) under the seasonality influence of the absolute humidity, aerosol route and fomite-mediated route dominant transmission in winter and summer repectively; 3) two routes transmission model can be better used for influenza transmission control for different seasons.

This research studies the impact of random bandwidth allocation on complex network transmission under the condition of limited bandwidth resources. Two types of heterogeneous random bandwidth allocation strategies are designed and a series of simulation tests are conducted within HK-BA network. The result reveals that the completely random bandwidth allocation has an uncertain influence on network load. Meanwhile, comparing to the shortest routing strategy, the G-L routing strategy has a better robustness in network load performance with the random bandwidth allocation. In addition, the network load capacity can be effectively improved by using the classification random bandwidth allocation with the reasonable combination of the edge weight and bandwidth in complex networks of betweenness heterogeneity.

For second-order multi-agent system with disturbance, a distributed control protocol with individual local information is proposed, and the sufficient condition is obtained to ensure the flocking motion and the speed synchronization when multi-agent system is disturbed by outside world. First, we design a disturbance observer for each agent, then, a distributed control protocol based on disturbance observer is presented. The protocol can guarantee multi-agent system achieve flocking motion and synchronous speed. Finally, simulation results show the validity of the conclusion.

This paper mainly studied the influence of the feedback strength on the dynamics of nonlinear optical feedback systems. We adopt Delay Partial Differential Equations as the theoretical model of nonlinear optical feedback systems, through the method of theoretical analysis and numerical simulation to study dynamics of the systems under the change of the feedback strength. We find that the system has a wide range of dynamical behavious as the feedback strength is increased gradually. It experiences periodic states, chaotic breather with slow and fast time scales, and then fully developed chaos.

Hopf bifurcation behavior and bifurcation control of a new Lorenz-like system are studied in this paper. Firstly, Hopf bifurcation type is determined by bifurcation stability norm. Then the linear controller and the non-linear controller are applied to control the original system respectively. In the section of linear control, the effect of linear parameter on the position of Hopf bifurcation is discussed by Routh-Hurwitz criterion; In the section of non-linear control, the Hopf bifurcation Normal Form of controlled system is obtained by using direct Normal Form method, and the effects of nonlinear parameter on amplitude of limit cycle and Hopf bifurcation type are discussed by coefficient of Normal Form. Discussions show that if non-linear parameter satisfies certain condition, bifurcation type of original system will be changed, and the periodic solution amplitude will increase with the parameter increasing.

In order to improve the output signal-to-noise ratio (SNR) of a hard-limiter circuit array, we study the mechanisms of vibrational stochastic resonance and array stochastic resonance, and then demonstrate the possible improvement via the electro-circuit experiments. Using both methods of stochastic resonance, we find that nonlinear system can cooperate with input signal and internal noise, resulting in the enhancement of output SNR. Moreover, upon increasing the array size, the SNR gain can be greater than unity for certain regions of noise intensity. It is also noted that, comparing with the method of array stochastic resonance, vibrational stochastic resonance can get better output SNR, and be easily implemented.