Narrow Results

Interfirm credit guarantees serve as a critical mechanism for firms to secure low-interest bank loans by enhancing their creditworthiness. However, this practice also creates a channels through which financial risks can spread, particularly when the guaranteed firm defaults. This paper introduces an evolving network model that incorporates financial risk contagion within an interfirm credit guarantee network. We examine how the structure of the network evolves under the influence of financial risk contagion and investigate the role of firms’ behaviors and operational activities in exacerbating or mitigating risk contagion. Our simulation findings indicate that under financial risk contagion, interfirm credit guarantee networks exhibit core–periphery structures that emerge and evolve. However, these structures deviate from an idealized core–periphery configuration. Additionally, both the cumulative indegree and outdegree distributions of the network follow a power-law distribution and maintain dynamic stability throughout the risk contagion process. Moreover, we find that defaults of core firms lead to a higher failure rate compared to those of peripheral firms, indicating that core firms play a more significant role in enhancing network resilience than their peripheral counterparts. Further analysis reveals that a firm’s operational activities have a significant impact on the propagation of risk within the interfirm credit guarantee network. In contrast, the level of a firm’s risk aversion exerts minimal influence on the network’s resilience to risk contagion.

Previously many studies on online social networks simply analyze the static topology in which the friend relationship once established, then the links and nodes will not disappear, but this kind of static topology may not accurately reflect temporal interactions on online social services. In this study, we define four types of users and interactions in the interaction (dynamic) network. We found that active, disappeared, new and super nodes (users) have obviously different strength distribution properties and this result also can be revealed by the degree characteristics of the unweighted interaction and friendship (static) networks. However, the active, disappeared, new and super links (interactions) only can be reflected by the strength distribution in the weighted interaction network. This result indicates the limitation of the static topology data on analyzing social network evolutions. In addition, our study uncovers the approximately stable statistics for the dynamic social network in which there are a large variation for users and interaction intensity. Our findings not only verify the correctness of our definitions, but also helped to study the customer churn and evaluate the commercial value of valuable customers in online social networks.

We propose a special opinion model on Internet users' social platform selections where users only view those converging platforms as tools to maintain their communications with all of their friends or partners and one may use more than one platform at the same time. We construct the time evolution differential equations, seek the fixed points, and study their attractability and repellency by analyzing those equations. Then, we verify the analytical results and observe their accuracy by numerical simulation. The conclusion shows that in any practical system described by our model, one platform will completely eliminate its competitor sooner or later, and when the average degree of the interaction network is relatively low, the laggard may have more chance to turn the tide, but when that average degree is high, that chance is extremely limited.

Network structure will evolve over time, which will lead to changes in the spread of the epidemic. In this work, a network evolution model based on the principle of preferential attachment is proposed. The network will evolve into a scale-free network with a power-law exponent between 2 and 3 by our model, where the exponent is determined by the evolution parameters. We analyze the epidemic spreading process as the network evolves from a small-world one to a scale-free one, including the changes in epidemic threshold over time. The condition of epidemic threshold to increase is given with the evolution processes. The simulated results of real-world networks and synthetic networks show that as the network evolves at a low evolution rate, it is more conducive to preventing epidemic spreading.

An effective and reliable evolution model can provide strong support for the planning and design of transportation networks. As a network evolution mechanism, link prediction plays an important role in the expansion of transportation networks. Most of the previous algorithms mainly took node degree or common neighbors into account in calculating link probability between two nodes, and the structure characteristics which can enhance global network efficiency are rarely considered. To address these issues, we propose a new evolution mechanism of transportation networks from the aspect of link prediction. Specifically, node degree, distance, path, expected network structure, relevance, population and GDP are comprehensively considered according to the characteristics and requirements of the transportation networks. Numerical experiments are done with China’s high-speed railway network, China’s highway network and China’s inland civil aviation network. We compare receiver operating characteristic curve and network efficiency in different models and explore the degree and hubs of networks generated by the proposed model. The results show that the proposed model has better prediction performance and can effectively optimize the network structure compared with other baseline link prediction methods.

The power-law exponent γ describing the form of the degree distribution of networks is related to the mean degree of the networks. This relation provides a useful method to estimate this exponent because mean degrees can easily be obtained by the number of vertices and edges in the networks. We improved the relation obtained only by ordinary integral approximation, and examined its availability for number of vertices, minimum degree considered, and range of γ, taking also into consideration the number of vertices with minimum degree. As a result, we were able to estimate slight deviations of γ from 3, which are usually observed in numerical simulations of growing networks with linear preferential attachment. Furthermore, using this method, we were also able to predict to what extent γ changed by joining pre-existing vertices in growing networks or by imposing restrictions to prevent the gaining of new edges for certain vertices. For cases where γ < 2, we estimated the power-law exponent of degree distributions of networks formed by traces of random walkers from the increased rate of vertices with created edges.

Biological networks are structurally adaptive and take on non-random topological properties that influence system robustness. Studies are only beginning to reveal how these structural features emerge, however the influence of component fitness and community cohesion (modularity) have attracted interest from the scientific community. In this study, we apply these concepts to an evolutionary algorithm and allow its population to self-organize using information that the population receives as it moves over a fitness landscape. More precisely, we employ fitness and clustering based topological operators for guiding network structural dynamics, which in turn are guided by population changes taking place over evolutionary time. To investigate the effect on evolution, experiments are conducted on six engineering design problems and six artificial test functions and compared against cellular genetic algorithms and panmictic evolutionary algorithm designs. Our results suggest that a self-organizing topology evolutionary algorithm can exhibit robust search behavior with strong performance observed over short and long time scales. More generally, the coevolution between a population and its topology may constitute a promising new paradigm for designing adaptive search heuristics.

One of the notable trends in the software industry is that software vendors provide their software on a platform as a service. Software users consume those software services or compose new services by combining those existing software services. The software vendors, their services, software users, and the platform represent an open innovation system. Collective intelligence is the underlying mechanism for the cooperation between the users of the system, i.e. their continuous reuse of existing software services for the creation of new services. A successfully working software services system (i.e. collective intelligence system) is a system that is continuously adapted by its users to meet their needs. The evolution of this software-as-a-service (SaaS) innovation system and the behavior of SaaS users within this system are still unknown. In this paper, we describe the evolution of a SaaS network. The SaaS network consists of nodes (i.e. software services with open interfaces) and links (i.e. the co-development relationships of software services with open interfaces). The results suggest that the SaaS network has gradually grown into a scale-free network with a slight concavity in its cumulative degree distribution. The results also suggest that the topology characteristics are invariant over time except for the early time periods. Furthermore, the results suggest that the SaaS network is not as open (i.e. inter-operable) as its technology let us expect. Considering these results, we imply that SaaS innovation is achieved by platform providers striving to capture users with a few, leading SaaS services. That means, SaaS innovation is not achieved through the possibilities of potential combinations between any kind of SaaS services as could be expected theoretically. These findings are expected to stir further research on the actual structure of open innovation systems that are driven by collective intelligence.

We ask whether natural selection has shaped three biologically important features of 15 signal transduction networks and two genome-scale transcriptional regulation networks. These features are regulatory cycles, the lengths of the longest pathways through a network — a measure of network compactness — and the abundance of node pairs connected by many alternative regulatory pathways. We determine whether these features are significantly more or less abundant in biological networks than in randomized networks with the same distribution of incoming and outgoing connections per network node. We find that autoregulatory cycles are of exceptionally high abundance in transcriptional regulation networks. All other cycles, however, are significantly less abundant in several signal transduction networks. This suggests that the multistability caused by complex feedback loops in a network may interfere with the functioning of such networks. We also find that several of the networks we examine are more compact than expected by chance alone. This raises the possibility that the transmission of information through such networks, which is fastest in compact networks, is a biologically important characteristic of such networks.

This paper develops an analytical model of contagion risk in banking systems with tiered structure. It explores the respective effects of banking network structure and bank activity on contagion risk in banking network evolution. The findings suggest that increasing interbank connections is conducive to handling banking crisis and reducing the effect of contagion risk, but its positive effect is limited; raising bank reserve ratio will enhance the stability of individual banks to a certain extent, but it may immediately lead to liquidity problems for banks that have less excess reserves, causing the occurrence of contagion risk; an excessive drive for risk assets with high return may bring high risk to banks and lead to instability of banking systems; the bank risk preference is crucial to the stability of banking systems, and the radicalness of it may lead to greater systemic instability.

The interaction of information and the evolution of network structure are inseparable. In order to construct social network evolution and information propagation models that better fit real-world scenarios, this paper proposes a social network structure evolution model driven by changes in the strength of relationships between individuals through their information interactions with each other. During the evolution process of the network, information interaction between individuals is also influenced by the network structure. Therefore, we improve traditional propagation models and construct an information propagation model with dynamic propagation rates. The proposed model is used to simulate both the spread of information and the evolution of network structures in real social networks. Finally, simulation results are compared to real-world data, demonstrating the effectiveness and rationality of the proposed model.

Electric vehicle technology is a crucial technology for achieving sustainable energy transformation, which is of great significance to climate change and promotes sustainable development. This paper attempts to study the transnational R&D cooperation of electric vehicles. According to the authorized data of transnational co-patents from the United States Patent and Trademark Office (USPTO), a social network analysis method is employed, and a detailed study of transnational co-patent networks in electric vehicles is conducted, including the construction of network, the analysis of nationality distribution of co-patent inventors, the analysis of structural characteristics and important nodes of network in different stages from the perspective of inventors and countries. The research results show that the cooperative groups formed by inventors are independent of each other and have not yet formed a large network; the degree of transnational cooperation in developed countries far exceeds that in developing countries; the US and Germany are the dual-core in the transnational co-patent networks; the breadth and intensity of transnational cooperation are strengthening, and the regional borders are less and less restrictive. As for the existing problems, authoritative inventors could organize large international R&D cooperation institutions to gather dispersed inventors together and connect them into a large inventors’ network; developing countries are encouraged to seek partners through the network, actively participate in transnational R&D cooperation, and developed countries are encouraged to hold global technological innovation events.