The enhancement of the robustness of complex networks has been a hotspot in the field of network science in recent years. It is both of great scientific significance and theoretical value to explore the strategy of enhancing the robustness for network structure design and function improvement. On the basis of extensive collation and systematic analysis of domestic and foreign literature, this paper summarizes comprehensively the key point and main ideasof the current research on the enhancement strategies of complex networks robustness from three aspects: pre-defense, in-process recovery and post-optimization. The advantages, disadvantages and applicability of different strategies are compared and analyzed. Then we look forward to future research direction in this field.

Combined with the topological characteristics and electrical characteristics, this paper proposes an integrated multi-attribute decision method for identifying key nodes in a power grid. Firstly, based on the complex network theory, several evaluation indicators are proposed and calculated considering the topology characteristics and the electrical characteristics of a power grid, an evaluation matrix is obtained. Then the final decision matrix is obtained by weighting the evaluation matrix combined with the analytic hierarchy process and the coefficient of variation method. Finally, the TOPSIS method combined with grey correlation degree is used to calculate the ranking of the important nodes in the power grid.In order to compare the advantages and disadvantages of different identification methods, the network efficiency and synchronization performance are adopted. An actual local power grid is used to further verify the effectiveness a feasibility of the proposed method.

The stable collaboration relationship between developers in open source software community can promote the emergence of community collective innovation. Taking the open source software project AngularJS as an example to study the impact of the stability on developer collaboration network in open source software community, which was caused by staff turnover. We construct the network by the project code-collaboration relationships. Then we analyze the structure and function of the developer collaboration network. According to the structure features and function features of the nodes, we classify them into different types, then explore the impact of the loss of different types of nodes on the network structure and functional stability. The study found that the developer's features are asymmetrical, which is the reason of the structure and function stability on the network are inconsistent; Different types of developers have different influences on network stability. A member owes large degree may not be the most important in the network; Coordination center developers who with high within-module degree and participation coefficient also have large knowledge contributions; Developers participating in multiple subprojects often occupy the center of the network. These two types of developers have a greater impact on network stability.

To solve the low-frequency short-sequence data and unbalanced classification problem in detecting the abnormal water consumption pattern of enterprises, this paper proposes a two-class prediction method based on Isolation Forest sampling. Firstly, the volatility and statistical features of water consumption are constructed. The Isolated Forest algorithm is used to calculate the degree of isolation of samples in the large class to measure the representation of each sample, and the samples are extracted according to their representation. Then the extracted samples are merged with the small class to form a balanced training dataset. Finally, the XGBoost classifier is trained with the balanced dataset and predicting the abnormal pattern. On the dataset of 7,604 enterprises' 13-month water consumption in a city, the AUC and recall ratio of the method proposed by this paper can reach 0.927 and 0.891, and those of XGBoost method based on random under sampling are 0.855 and 0.733, which are improved by 4.7% and 21.6% respectively.

Considering the influence of government guidance on the decision-making of farmers and supermarket chain enterprises during the development of agricultural super-docking, an evolutionary game model was constructed with government, enterprises and farmers, investigate the interaction mechanism of strategy choice between different subjects, and with the help of the simulation analysis of different members operation behavior, the influence of the farmers and stable operation. The result shows that the value perception of government policy support and subsidies by farmers and companies, as well as the level of participation costs and benefits, are important factors that affect the decision-making behavior of both parties; the low punishment intensity is benefit the government and companies in the direction of active operation evolution. Medium supervision intensity is helpful for both sides to reach an agreement on the business operation of "connecting farmers and supermarkets". Otherwise, the supervision cost will be too high, resulting in the dilemma of government policy guidance. Finally, it is pointed that the government should continue to play a leading role in the development of agricultural super docking. And put forward countermeasures and suggestions to promote its healthy and orderly development.

In this paper, the guided crowd evacuation dynamics model based on cellular automata is proposed, this model combines Gaussian mixture methodand fuzzy logic theory to study the influence of guides on pedestrian evacuation behaviors. Bayesian Information Criterion (BIC)determines the optimal number of guides and EM algorithmdetermines the optimal positions of the guides. The cellular automata model is used as the driven model of pedestrian motion, and fuzzy logic theory is adopted to simulate the pedestrians' selection behaviors for the guides. The influences of guide quantity, guide speed and exit width on the evacuation are explored, and it is concluded that the guides can improve the evacuation efficiency to a certain extent. However, the number of guides is not the more the better. Increasing the width of the exit within a certain range can increase the capacity of the exit and effectively reduce the evacuation time. When the speed of the guide is 75% of the pedestrian speed, the evacuation efficiency could be the highest. The model in this paper combines the advantages of small computation of cellular automata model and high clustering accuracy of Gaussian mixture model, which provides a new idea for pedestrian dynamics modeling and puts forward effective suggestions for pedestrian evacuation behaviors.

In order to explore the usage behavior and cause of emojis on social media, we analyzed the use of emojis in 1 800 958 microblogs under the topic of "Kunshan Case" on Sina Weibo. First, we analysis the frequency of emoji to study the phenomenon of repeated usage of emoji in the group, and then we classify popular emojis and micro-blog texts, we analyze the diversity of emoji usage of individual users. The result shows that: There are a lot of emojis in Weibo and the frequency of emojis is long-tailed and follow Zipf's Law; The evolution of popular emojis can reflect the public opinion of the event; Individual users are used to using 2~3 same emojis or different emojis with similar emotions. The emojis used by users are often related to the topics they express and are influenced by the psychology of herds, and the phenomenon of co-occurrence emojis is usually to strengthen the emotions expressed.

In order to provide effective decision-making basis for improvement of complex network invulnerability and protection of important nodes, this paper establishes a complex network invulnerability and node importance evaluation model based on redundancy. Firstly, the redundancy of complex networks is defined. At the same time, based on the redundancy, the invulnerability of the network is quantified. Then, this paper uses the global attribute of redundancy to evaluate the importance of each node in the network by means of node deletion. Finally, this paper uses actual networks for simulation experiments. The results show that the model and algorithm can provide a solution to the problem of high invulnerability network construction under some cost constraints, and at the same time they are effective and superior for evaluating the importance of nodes in larger networks.

In order to find the possible generation mechanism of double power law distribution, a spatial network model with dynamic constraints was proposed in this paper based on double-value confined areas. The results of theoretical analysis and simulation showed that the degree distribution of the model presented a single power law in the early stage of evolution, and then evolved into a stable double power law. Furthermore, we made an empirical analysis on Chinese aviation network to verify the validity of the model in reality. The empirical results not only gave the evolution patterns of degree distribution as predicted by theoretical model, but also enabled us to directly observe the evolution trend of the confined area required by model to match its degree distribution. Our theoretical model could provide a possible explanation for the emergence of double power law degree distribution.