Robustness analysis of network controllability

Structural controllability, which is an interesting property of complex networks, attracts many researchers from various fields. The maximum matching algorithm was recently applied to explore the minimum number of driver nodes, where control signals are injected, for controlling the whole network. Here we study the controllability of directed Erdös–Rényi and scale-free networks under attacks and cascading failures. Results show that degree-based attacks are more efficient than random attacks on network structural controllability. Cascade failures also do great harm to network controllability even if they are triggered by a local node failure.     

Studies on controllability of directed networks with extremal optimization

Almost all natural, social and man-made-engineered systems can be represented by a complex network to describe their dynamic behaviors. To make a real-world complex network controllable with its desired topology, the study on network controllability has been one of the most critical and attractive subjects for both network and control communities. In this paper, based on a given directed–weighted network with both state and control nodes, a novel optimization tool with extremal dynamics to generate an optimal network topology with minimum control nodes and complete controllability under Kalman’s rank condition has been developed. The experimental results on a number of popular benchmark networks show the proposed tool is effective to identify the minimum control nodes which are sufficient to guide the whole network’s dynamics and provide the evolution of network topology during the optimization process. We also find the conclusion: “the sparse networks need more control nodes than the dense, and the homogeneous networks need fewer control nodes compared to the heterogeneous” (Liu et al., 2011  [18]), is also applicable to network complete controllability. These findings help us to understand the network dynamics and make a real-world network under the desired control. Moreover, compared with the relevant research results on structural controllability with minimum driver nodes, the proposed solution methodology may also be applied to other constrained network optimization problems beyond complete controllability with minimum control nodes.     

Robustness measurement of scale-free networks based on motif entropy

As a classical complex network model, scale-free network is widely used and studied. And motifs, as a high-order subgraph structure, frequently appear in scale-free networks, and have a great influence on the structural integrity, functional integrity and dynamics of the networks. In order to overcome the shortcomings in the existing work on the robustness of complex networks, only nodes or edges are considered, while the defects of high-order structure in the network are ignored. From the perspective of network motif, we propose an entropy of node degree distribution based on motif to measure the robustness of scale-free networks under random attacks. The effectiveness and superiority of our method are verified and analyzed in the BA scale-free networks.     

复杂网络可控性研究现状综述

控制复杂系统是人们对复杂系统模型结构及相关动力学进行研究的最终目标, 反映人们对复杂系统的认识能力. 近年来, 通过控制理论和复杂性科学相结合,复杂网络可控性的研究引起了人们的广泛关注. 在过去的几年内, 来自国内外不同领域的研究人员从不同的角度对复杂网络可控性进行了深入的分析研究, 取得了丰硕的成果. 本文重点讨论了复杂网络的结构可控性研究进展, 详细介绍了基于最大匹配方法的复杂网络结构可控性分析框架, 综述了自2011年以来复杂网络可控性的相关研究成果, 具体论述了不同类型的可控性、可控性与网络拓扑结构统计特征的关联、基于可控性的网络及节点度量、控制的鲁棒性和可控性的相关优化方法. 最后, 对网络可控性未来的研究动态进行了展望, 有助于国内同行开展网络可控性的相关研究.     

Complex networks: Dynamics and security

This paper presents a perspective in the study of complex networks by focusing on how dynamics may affect network security under attacks. In particular, we review two related problems: attack-induced cascading breakdown and range-based attacks on links. A cascade in a network means the failure of a substantial fraction of the entire network in a cascading manner, which can be induced by the failure of or attacks on only a few nodes. These have been reported for the internet and for the power grid (e.g., the August 10, 1996 failure of the western United States power grid). We study a mechanism for cascades in complex networks by constructing a model incorporating the flows of information and physical quantities in the network. Using this model we can also show that the cascading phenomenon can be understood as a phase transition in terms of the key parameter characterizing the node capacity. For a parameter value below the phase-transition point, cascading failures can cause the network to disintegrate almost entirely. We will show how to obtain a theoretical estimate for the phase-transition point. The second problem is motivated by the fact that most existing works on the security of complex networks consider attacks on nodes rather than on links. We address attacks on links. Our investigation leads to the finding that many scale-free networks are more sensitive to attacks on short-range than on long-range links. Considering that the small-world phenomenon in complex networks has been identified as being due to the presence of long-range links, i.e., links connecting nodes that would otherwise be separated by a long node-to-node distance, our result, besides its importance concerning network efficiency and security, has the striking implication that the small-world property of scale-free networks is mainly due to short-range links.     

基于节点负荷失效的网络可控性研究

Liu和Barabasi将现代控制理论应用到线性系统的网络可控性问题上, 提出了最小驱动节点集的计算方法, 解决了复杂网络控制的可计算问题. 针对现实网络中存在的节点因负荷过载而失效的问题, 本文提出了基于节点负荷失效的网络可控性模型. 通过对网络采用介数和Weibull失效模型, 在随机和目标失效机制下进行仿真, 研究结果表明: 维持无标度网络可控性的难度要明显大于随机网络; 在目标节点失效机制下, 即使对网络输入极少的失效信号, 也能极大地破坏网络的可控性; 使高介数节点失效要比使度高节点失效更能破坏网络的可控性, 说明高介数节点在维持网络可控性上发挥着重要作用; 对不同的负荷失效模型, 要合理采取措施, 防止网络发生阶跃性全不可控现象. 关键词： 网络可控性 结构可控性 节点失效     

利用邻域“结构洞”寻找社会网络中最具影响力节点

识别复杂网络中的关键节点对网络结构优化和鲁棒性增强具有十分重要的意义. 经典的关键节点测量方法在一定程度上能够辨识网络中影响力节点, 但存在一定局限性: 局部中心性测量方法仅考虑节点邻居的数目, 忽略了邻居间的拓扑关系, 不能在计算中反映邻居节点间的相互作用; 全局测量方法则由于算法本身的复杂性而不能应用于大规模社会网络的分析, 另外, 经典的关键节点测量方法也没有考虑社会网络特有的社区特征. 为高效、准确地辨识具有社区结构的社会网络中最具影响力节点, 提出了一种基于节点及其邻域结构洞的局部中心性测量方法, 该方法综合考虑了节点的邻居数量及其与邻居间的拓扑结构, 在节点约束系数的计算中同时体现了节点的度属性和“桥接”属性. 利用SIR(易感-感染-免疫)模型在真实社会网络数据上对节点传播能力进行评价后发现, 所提方法可以准确地评价节点的传播能力且具有强的鲁棒性.     

Error and attack tolerance of evolving networks with local preferential attachment

Networks generated by local-world evolving network model display a transition from exponential network to power-law network with respect to connectivity distribution. We investigate statistical properties of the evolving networks and the responses of these networks under random errors and intentional attacks. It has been found that local world size M has great effect on the network's heterogeneity, thus leading to transitional behaviors in network's robustness against errors and attacks. Numerical results show that networks constructed with local preferential attachment mechanism can maintain the robustness of scale-free networks under random errors and concurrently improve reliance against targeted attacks on highly connected nodes.     

Maximum entropy networks are more controllable than preferential attachment networks

A maximum entropy (ME) method to generate typical scale-free networks has been recently introduced. We investigate the controllability of ME networks and Barabási–Albert preferential attachment networks. Our experimental results show that ME networks are significantly more easily controlled than BA networks of the same size and the same degree distribution. Moreover, the control profiles are used to provide insight into control properties of both classes of network. We identify and classify the driver nodes and analyze the connectivity of their neighbors. We find that driver nodes in ME networks have fewer mutual neighbors and that their neighbors have lower average degree. We conclude that the properties of the neighbors of driver node sensitively affect the network controllability. Hence, subtle and important structural differences exist between BA networks and typical scale-free networks of the same degree distribution.     

Multiple scale analysis of complex networks using the empirical mode decomposition method

Empirical mode decomposition (EMD) method can decompose any complicated data into finite ‘intrinsic mode functions’ (IMFs). In this paper, we use EMD method to analyze and discuss the structural properties of complex networks. A random-walk method is used to collect the data series of network systems. Utilizing the EMD method, we decompose the obtained data into finite IMFs under different spatial scales. The analysis results show that EMD method is an effective tool for capturing the topological properties of network systems under different spatial scales, such as the modular structures of network systems and their energy densities.     

Network robustness to targeted attacks. The interplay of expansibility and degree distribution

We study the property of certain complex networks of being both sparse and highly connected, which is known as “good expansion” (GE). A network has GE properties if every subset S of nodes (up to 50% of the nodes) has a neighborhood that is larger than some “expansion factor” φ multiplied by the number of nodes in S. Using a graph spectral method we introduce here a new parameter measuring the good expansion character of a network. By means of this parameter we are able to classify 51 real-world complex networks — technological, biological, informational, biological and social — as GENs or non-GENs. Combining GE properties and node degree distribution (DD) we classify these complex networks in four different groups, which have different resilience to intentional attacks against their nodes. The simultaneous existence of GE properties and uniform degree distribution contribute significantly to the robustness in complex networks. These features appear solely in 14% of the 51 real-world networks studied here. At the other extreme we find that ∼40% of all networks are very vulnerable to targeted attacks. They lack GE properties, display skewed DD — exponential or power-law — and their topologies are changed more dramatically by targeted attacks directed at bottlenecks than by the removal of network hubs.     

Deterministic self-similar models of complex networks based on very symmetric graphs

Using very symmetric graphs we generalize several deterministic self-similar models of complex networks and we calculate the main network parameters of our generalization. More specifically, we calculate the order, size and the degree distribution, and we give an upper bound for the diameter and a lower bound for the clustering coefficient. These results yield conditions under which the network is a self-similar and scale-free small world network. We remark that all these conditions are posed on a small base graph which is used in the construction. As a consequence, we can construct complex networks having prescribed properties. We demonstrate this fact on the clustering coefficient. We propose eight new infinite classes of complex networks. One of these new classes is so rich that it is parametrized by three independent parameters.     

A last updating evolution model for online social networks

As information technology has advanced, people are turning to electronic media more frequently for communication, and social relationships are increasingly found on online channels. However, there is very limited knowledge about the actual evolution of the online social networks. In this paper, we propose and study a novel evolution network model with the new concept of “last updating time”, which exists in many real-life online social networks. The last updating evolution network model can maintain the robustness of scale-free networks and can improve the network reliance against intentional attacks. What is more, we also found that it has the “small-world effect”, which is the inherent property of most social networks. Simulation experiment based on this model show that the results and the real-life data are consistent, which means that our model is valid.     

面向结构洞的复杂网络关键节点排序

复杂网络中的结构洞节点对于信息传播具有重要作用, 现有关键节点排序方法多数没有兼顾结构洞节点和其他类型的关键节点进行排序. 本文根据结构洞理论与关键节点排序相关研究选取了网络约束系数、介数中心性、等级度、效率、网络规模、PageRank值以及聚类系数7个度量指标, 将基于ListNet的排序学习方法引入到复杂网络的关键节点排序问题中, 融合7个度量指标, 构建了一个能够综合评价面向结构洞节点的关键节点排序方法. 采用模拟网络和实际复杂网络进行了大量实验, 人工标准试验结果表明本文排序方法能够综合考虑结构洞节点和核心节点, 关键节点排序与人工排序结果具有较高的一致性. SIR传播模型评估实验结果表明由本文选择TOP-K节点发起的传播能够在较短的传播时间内达到最大的传播范围.     

基于传播免疫的复杂网络可控性研究

复杂网络控制反映了人类对复杂系统的认识深度和改造能力. 最新研究成果基于线性系统控制理论建立了复杂网络可控性的理论架构, 能够发现任意拓扑结构的线性时不变复杂网络中控制全部节点状态的最小驱动节点集, 但是该模型未考虑免疫节点或失效节点对控制信号传播的阻断.在继承该模型优点的前提下, 重新构建了基于传播免疫的复杂网络控制模型.在采用分属于随机免疫和目标免疫两种策略的 4个方法确定免疫节点的情况下,分析14个真实网络的可控性.结果表明:如果将网络中度数、 介数和紧密度指标较高的节点作为免疫节点,将极大地提高控制复杂网络的难度. 从而在一定程度上丰富了以往模型的结论.     

Link prediction in complex networks via modularity-based belief propagation

Link prediction aims at detecting missing, spurious or evolving links in a network, based on the topological information and/or nodes' attributes of the network. Under the assumption that the likelihood of the existence of a link between two nodes can be captured by nodes' similarity, several methods have been proposed to compute similarity directly or indirectly, with information on node degree. However, correctly predicting links is also crucial in revealing the link formation mechanisms and thus in providing more accurate modeling for networks. We here propose a novel method to predict links by incorporating stochastic-block-model link generating mechanisms with node degree. The proposed method first recovers the underlying block structure of a network by modularity-based belief propagation, and based on the recovered block structural information it models the link likelihood between two nodes to match the degree sequence of the network. Experiments on a set of real-world networks and synthetic networks generated by stochastic block model show that our proposed method is effective in detecting missing, spurious or evolving links of networks that can be well modeled by a stochastic block model. This approach efficiently complements the toolbox for complex network analysis, offering a novel tool to model links in stochastic block model networks that are fundamental in the modeling of real world complex networks.     

Bootstrapping Topological Properties and Systemic Risk of Complex Networks Using the Fitness Model

In this paper we present a novel method to reconstruct global topological properties of a complex network starting from limited information. We assume to know for all the nodes a non-topological quantity that we interpret as fitness. In contrast, we assume to know the degree, i.e. the number of connections, only for a subset of the nodes in the network. We then use a fitness model, calibrated on the subset of nodes for which degrees are known, in order to generate ensembles of networks. Here, we focus on topological properties that are relevant for processes of contagion and distress propagation in networks, i.e. network density and k-core structure, and we study how well these properties can be estimated as a function of the size of the subset of nodes utilized for the calibration. Finally, we also study how well the resilience to distress propagation in the network can be estimated using our method. We perform a first test on ensembles of synthetic networks generated with the Exponential Random Graph model, which allows to apply common tools from statistical mechanics. We then perform a second test on empirical networks taken from economic and financial contexts. In both cases, we find that a subset as small as 10 % of nodes can be enough to estimate the properties of the network along with its resilience with an error of 5 %.     

False Positive and False Negative Effects on Network Attacks

Robustness against attacks serves as evidence for complex network structures and failure mechanisms that lie behind them. Most often, due to detection capability limitation or good disguises, attacks on networks are subject to false positives and false negatives, meaning that functional nodes may be falsely regarded as compromised by the attacker and vice versa. In this work, we initiate a study of false positive/negative effects on network robustness against three fundamental types of attack strategies, namely, random attacks (RA), localized attacks (LA), and targeted attack (TA). By developing a general mathematical framework based upon the percolation model, we investigate analytically and by numerical simulations of attack robustness with false positive/negative rate (FPR/FNR) on three benchmark models including Erd?s-Rényi (ER) networks, random regular (RR) networks, and scale-free (SF) networks. We show that ER networks are equivalently robust against RA and LA only when FPR equals zero or the initial network is intact. We find several interesting crossovers in RR and SF networks when FPR is taken into consideration. By defining the cost of attack, we observe diminishing marginal attack efficiency for RA, LA, and TA. Our finding highlights the potential risk of underestimating or ignoring FPR in understanding attack robustness. The results may provide insights into ways of enhancing robustness of network architecture and improve the level of protection of critical infrastructures.     

一种基于文本互信息的金融复杂网络模型

复杂网络能够解决许多金融问题,能够发现金融市场的拓扑结构特征,反映不同金融主体之间的相互依赖关系.相关性度量在金融复杂网络构建中至关重要.通过将多元金融时间序列符号化,借鉴文本特征提取以及信息论的方法,定义了一种基于文本互信息的相关系数.为检验方法的有效性,分别构建了基于不同相关系数(Pearson和文本互信息)和不同网络缩减方法(阈值和最小生成树)的4个金融复杂网络模型.在阈值网络中提出了使用分位数来确定阈值的方法,将相关系数6等分,取第4部分的中点作为阈值,此时基于Pearson和文本互信息的阈值模型将会有相近的边数,有利于这两种模型的对比.数据使用了沪深两地证券市场地区指数收盘价,时间从2006年1月4日至2016年12月30日,共计2673个交易日.从网络节点相关性看,基于文本互信息的方法能够体现出大约20%的非线性相关关系;在网络整体拓扑指标上,本文计算了4种指标,结果显示能够使所保留的节点联系更为紧密,有效提高保留节点的重要性以及挖掘出更好的社区结构;最后,计算了阈值网络的动态指标,将数据按年分别构建网络,缩减方法只用了阈值方法,结果显示本文提出的方法在小世界动态和网络度中心性等指标上能够成功捕捉到样本区间内存在的两次异常波动.此外,本文构建的地区金融网络具有服从幂律分布、动态稳定性、一些经济欠发达地区在金融地区网络中占据重要地位等特性.     

一种基于网络最大可控子图的导航搜索模型

基于网络可控性模型提出了最大可控子图的概念,在此基础上提出了一种基于最大可控子图的导航搜索模型. 模型中基于最大可控子图的加边策略用最小的代价解决了有向网络搜索中存在的粒子因“无路可走”而终止搜索的问题;基于最大可控子图部署导航节点,仅用节点总数2%左右的导航点,就使全网搜索时间接近导航网络的平均最短路径. 通过在ER和SF 网络上的实验表明,全网搜索时间与网络的可控性有关,可控性越好,添加的边数量越少,同时会使网络中导航节点分布越多,越能提高网络的搜索效率. 关键词： 导航搜索 有向网 网络可控性