The conditions for the spontaneous emergence of cooperation under the complex human behavior model have become the focus of many disciplines. Exploring the conditions for cooperation has both important scientific significance and theoretical value for understanding the institutional arrangements in human society. The social dilemma games provide a theoretical prototype for studying cooperation issues between multiple individuals. As a dynamic analysis method that can describe individuals′ learning and strategy adjustment processes, evolutionary game theory has been one of the most effective frameworks for studying the evolution of cooperation. This review article systematically summarizes the research progress of using the evolutionary game method to study the issues of group cooperation in social dilemma games. Specifically, the following topics are included: research progress of (1) social dilemma game models, evolutionary game theory and equilibrium analysis methods, (2) social dilemma games and the evolution of cooperation under reward/punishment mechanism and reputation mechanism, (3) social dilemma games and the evolution of cooperation with separation strategy and extortion strategy, and (4) social dilemma games and the evolution of cooperation under the network reciprocity. Finally, prospects for further research issues in this area are presented.

The paper determines seven habitat patterns and plant growth rules, and propose ramets distribution network model to simulate the dynamic growth of digitally cloned plants. The effects of habitat heterogeneity and the structural characteristics of the plants on the diffusion process of ramets at different growth stages were revealed. Besides,we obtain ramet network and quantitatively describe the growth and diffusion process of ramet population under different treatments. The results show that the denser the patch distribution in the habitat, the faster the ramets population will grow. Different plants adaptability to the environment differently, and genotypes that produce longer spacers in resource-rich patchesprefer to explore the habitat while the genotypes with longer spacer in the resource-poor patches are more likely to obtain favorable resources in the habitat.

Different from the previous studies on the long memory and effectiveness of stock market returns and volatility, this paper focuses on the extreme fluctuation behavior of financial markets. Taking the most representative index of Chinese stock market, Shanghai stock index, as the sample, the financial market is divided into different time windows according to a certain period, and the extreme returns in each time window are formed into a time series. Taking extreme return series as the empirical research object, this paper studies the long memory of extreme return in Shanghai stock market by using multiple statistical methods such as rescaling range analysis and Detrended fluctuation analysis. The results show that both extreme return series and extreme volatility series have obvious long memory characteristics. The long memory characteristics of the extreme series is obviously stronger than that of the original full sample return series itself, which shows that there is a certain dependence between the extreme fluctuation behavior of the market, and the extreme fluctuation behavior of the market is measurable to a certain extent. In addition, the correlation between the maximum series and its corresponding volatility series, between the minimum series and its corresponding volatility series, and between the maximum series and the minimum series are further analyzed. The results show that the different extreme series show their own unique and different intensity interdependence.

In real social networks, there are multiple relationships between users. The existing social recommendation algorithms only consider the impact of one relationship on the recommendation results. Based on the multi-subnet composited complex network model, different social relationships among users are introduced into the user feature matrix. In this paper, matrix factorization social recommendation algorithm based on multi-relationship is proposed. By analyzing the experimental results on two real datasets, the social matrix factorization recommendation method with multi-relationship has a significant improvement in recommendation accuracy compared with the traditional matrix factorization algorithm.

User reputation measures user′s ability of rating accurate assessments of various objects,and is of great significance for ensuring the healthy development of social economy and people's livelihood.This paper empirically analyzed the user′s reputationin the MovieLens data setfrom two aspects of user′s activity and rating memory. Grouping by the user′s degree, the results showed that the user′s reputation increases with the user′s degree. When the data set was divided into 36 quarters according to the time, it was also found that the user′s reputation presents rising trend with the increase of the user′s degree. Moreover, the data set was divided into 9 years according to time, the results showed that the user's persistence rate is gradually reduced year by year. Furthermore, it presented a metric to measure the memory of user reputation. The results showed that the Kendall coefficients of user′s reputation ranking in 5 years are higher than those of user′s degree, indicating that the user′s reputation presents more memory than user′s activity. In addition, it proposed a null model to be compared with the empirical results. The results showed that the relationship between user′s reputation and degree as well as the memory of user reputation on the empirical data are significantly different from those of the null model.

Particle swarm optimization (PSO) is a new swarm intelligence algorithm.It has some advantages such as fewer parameters, easy implementation and good efficiency etc. such that obtains many applications. In order to improve the performance of PSO, based on adaptive particle swarm optimization and simulated annealing particle swarm optimization, a chaotic adaptive annealing particle swarm optimization based on chaotic mapping is proposed.A chaotic perturbation operator is added to near-optimal solutions to enhance the global search capability. A double selection strategy is adopted instead of the traditional inertia factor, which not only makes the inertia factor change with the change of objective function, but also with the change of distance between the current position and the previous one of the particle.The effect of self and group experience in iteration is dynamically adjusted by a linear decreasing accelerating factor.The performance of the improved algorithm is verified by numerical experiments. The results show that the improved algorithm is superior to adaptive PSO and simulated annealing PSO for different types of functions.

Aiming at the three-level closed-loop supply chain system composed of manufacturers, recyclers and consumers, chaos control theory is used to solve the chaotic control problem of closed-loop supply chains under three channel recovery modes. Firstly, the uncertainty of consumer's demand for remanufactured products is considered, over a period of time, data on the number of consumer resourcepurchases of remanufactured products are collected. The method of K-S test proves that the data obeys the uniform distribution, and further constructs the discrete dynamic model based on the decision variable. Secondly, using MATLAB software to simulate numerical simulation, study the recovery price sensitivity coefficient and the recovery price competition coefficient as fixed values, the manufacturer and the two recyclers draw a fork map, the largest Lyapunov index map, the initial value sensitive analysis map, and the system's chaotic characteristics and primary value sensitivity analysis. Finally, the chaotic system is controlled by means of state feedback control and parameter adjustment. The results show that this method can effectively improve or eliminate chaotic state, optimize decision-making behavior and improve the profit of decision makers.

Open innovation has become the main mode of enterprise innovation. The innovation community provides enterprises with a large number of externalinnovative resources. Effectively integrating various types of user′s knowledge in the innovation community can meet the needs of product and service innovation and promote the development of open innovation. Considering the knowledge characteristics of open innovation community and enterprise demand, this paper constructs the knowledge network model from three dimensions: innovation demand, innovation plan and innovation subject, then uses ontology to visualize and mine user-generated content in the innovation community and finally Huawei product custom community is taken as an example empirical research. Research shows that the knowledge network model based on ontology can realize multi-dimensional discovery of user′s knowledge in open innovation communities and help companies identify key users,innovative user demands and solutions to key technical problems, which can effectively overcome the complex problems of the knowledge discovery process in the open innovation community and provide new ideas for knowledge discovery for enterprises to develop open innovation.

With the in-depth analysis of space and time complexity of human behavior, a large number of empirical studies have found that human behavior is no longer considered as Poisson distribution, but shows a "clumpy" phenomenon that occurs frequently after a long silence. This paper proposes an improved method combining the characteristics of current online purchase behavior. By measuring and the purchase clumpy of customers on the YHD and the JD platform, this paper gives empirical research. The results show that the customer's recent purchase time R, the purchase frequency F and the customer's active odds are positively correlated; Clumpy has a moderating effect on the recent purchase time R, and the recent purchase time R is more related to the customer's active odds to the no clumpy customers.

In practice, the probability distribution model of the background noise is often unknown. Under this circumstance, a recursive least square adaptive algorithm is developed to estimate the random signal via the weighted stochastic pooling network. The analytical formula of the recursive least square adaptive algorithm is derived, and the convergence of the algorithm, the mean square error of the network outputs and the learning curve are analyzed. For non-stationary input signals, the proposed algorithm with the forgetting factor can effectively track the change of the signal. These theoretical results are demonstrated by the numerical experiments, and the phenomenon of suprathreshold stochastic resonance is also observed. The obtained results lay the fundamental framework for the application of the weighted stochastic pooling network in signal estimation.

This paper aims to analyze the force of infection of measles and the seasonality of measles transmission rate at different age groups in China, and to analyze the effect of seasonal children’s contact rate and seasonal population’s contacts on the seasonality of measles transmission rate. Base on the age-stratified reported cases of measles in China from 2013 to 2016, we used the force of infection analysis model to estimate the age-specific force of infection; we established an age-structured time series susceptible-infected-recovered (TSIR) model and analyzed the transmission dynamics of measles in China. The model contains two different driving factors for the measles seasonal transmission: the periodic contact rate among children, and the periodic contact rate of the overall population. Our results show that: 1) the force of infection of measles in China varied across ages with the highest for infants less than one years old (24%-44%) and the second highest for people aged 50-65 years;2) Seasonal contact rates that due to both school terms and Spring Festival travel rush will lead to seasonal transmission rate of measles in China; 3) school terms can increase transmission rate by 31%, and seasonal population contact rate caused by Spring Festival travel rush can increase the transmission rate by 23%.