相依网络上基于相连边的择优恢复算法

如何有效地应对和控制故障在相依网络上的级联扩散避免系统发生结构性破碎,对于相依网络抗毁性研究具有十分重要的理论价值和现实意义.最新的研究提出一种基于相依网络的恢复模型,该模型的基本思想是通过定义共同边界节点,在每轮恢复阶段找出符合条件的共同边界节点并以一定比例实施恢复.当前的做法是按照随机概率进行选择.这种方法虽然简单直观,却没有考虑现实世界中资源成本的有限性和择优恢复的必然性.为此,针对相依网络的恢复模型,本文利用共同边界节点在极大连通网络内外的连接边数计算边界节点的重要性,提出一种基于相连边的择优恢复算法(preferential recovery based on connectivity link,PRCL)算法.利用渗流理论的随机故障模型,通过ER随机网络和无标度网络构建的不同结构相依网络上的级联仿真结果表明,相比随机方法和度数优先以及局域影响力优先的恢复算法,PRCL算法具备恢复能力强、起效时间早且迭代步数少的优势,能够更有效、更及时地遏制故障在网络间的级联扩散,极大地提高了相依网络遭受随机故障时的恢复能力.     

Non-consensus Opinion Models on Complex Networks

Social dynamic opinion models have been widely studied to understand how interactions among individuals cause opinions to evolve. Most opinion models that utilize spin interaction models usually produce a consensus steady state in which only one opinion exists. Because in reality different opinions usually coexist, we focus on non-consensus opinion models in which above a certain threshold two opinions coexist in a stable relationship. We revisit and extend the non-consensus opinion (NCO) model introduced by Shao et al. (Phys. Rev. Lett. 103:01870, 2009). The NCO model in random networks displays a second order phase transition that belongs to regular mean field percolation and is characterized by the appearance (above a certain threshold) of a large spanning cluster of the minority opinion. We generalize the NCO model by adding a weight factor W to each individual’s original opinion when determining their future opinion (NCOW model). We find that as W increases the minority opinion holders tend to form stable clusters with a smaller initial minority fraction than in the NCO model. We also revisit another non-consensus opinion model based on the NCO model, the inflexible contrarian opinion (ICO) model (Li et al. in Phys. Rev. E 84:066101, 2011), which introduces inflexible contrarians to model the competition between two opinions in a steady state. Inflexible contrarians are individuals that never change their original opinion but may influence the opinions of others. To place the inflexible contrarians in the ICO model we use two different strategies, random placement and one in which high-degree nodes are targeted. The inflexible contrarians effectively decrease the size of the largest rival-opinion cluster in both strategies, but the effect is more pronounced under the targeted method. All of the above models have previously been explored in terms of a single network, but human communities are usually interconnected, not isolated. Because opinions propagate not only within single networks but also between networks, and because the rules of opinion formation within a network may differ from those between networks, we study here the opinion dynamics in coupled networks. Each network represents a social group or community and the interdependent links joining individuals from different networks may be social ties that are unusually strong, e.g., married couples. We apply the non-consensus opinion (NCO) rule on each individual network and the global majority rule on interdependent pairs such that two interdependent agents with different opinions will, due to the influence of mass media, follow the majority opinion of the entire population. The opinion interactions within each network and the interdependent links across networks interlace periodically until a steady state is reached. We find that the interdependent links effectively force the system from a second order phase transition, which is characteristic of the NCO model on a single network, to a hybrid phase transition, i.e., a mix of second-order and abrupt jump-like transitions that ultimately becomes, as we increase the percentage of interdependent agents, a pure abrupt transition. We conclude that for the NCO model on coupled networks, interactions through interdependent links could push the non-consensus opinion model to a consensus opinion model, which mimics the reality that increased mass communication causes people to hold opinions that are increasingly similar. We also find that the effect of interdependent links is more pronounced in interdependent scale free networks than in interdependent Erd?s Rényi networks.     

Percolation Phase Transitions from Second Order to First Order in Random Networks

We investigate a percolation process where an additional parameter q is used to interpolate between the classical Erd¨os–R′enyi(ER) network model and the smallest cluster(SC) model. This model becomes the ER network at q = 1, which is characterized by a robust second order phase transition. When q = 0, this model recovers to the SC model which exhibits a first order phase transition. To study how the percolation phase transition changes from second order to first order with the decrease of the value of q from 1 to 0, the numerical simulations study the final vanishing moment of the each existing cluster except the N-cluster in the percolation process. For the continuous phase transition,it is shown that the tail of the graph of the final vanishing moment has the characteristic of the convexity. While for the discontinuous phase transition, the graph of the final vanishing moment possesses the characteristic of the concavity.Just before the critical point, it is found that the ratio between the maximum of the sequential vanishing clusters sizes and the network size N can be used to decide the phase transition type. We show that when the ratio is larger than or equal to zero in the thermodynamic limit, the percolation phase transition is first or second order respectively. For our model, the numerical simulations indicate that there exists a tricritical point qcwhich is estimated to be between0.2 qc 0.25 separating the two phase transition types.     

Weighted percolation on directed networks

We present and numerically test an analysis of the percolation transition for general node removal strategies valid for locally treelike directed networks. On the basis of heuristic arguments we predict that, if the probability of removing node i is p(i), the network disintegrates if p(i) is such that the largest eigenvalue of the matrix with entries A(ij)(1-p(i)) is less than 1, where A is the adjacency matrix of the network. The knowledge or applicability of a Markov network model is not required by our theory, thus making it applicable to situations not covered by previous works.     

Percolation transitions are not always sharpened by making networks interdependent

We study a model for coupled networks introduced recently by Buldyrev et al., [Nature (London) 464, 1025 (2010)], where each node has to be connected to others via two types of links to be viable. Removing a critical fraction of nodes leads to a percolation transition that has been claimed to be more abrupt than that for uncoupled networks. Indeed, it was found to be discontinuous in all cases studied. Using an efficient new algorithm we verify that the transition is discontinuous for coupled Erd?s-Rényi networks, but find it to be continuous for fully interdependent diluted lattices. In 2 and 3 dimensions, the order parameter exponent β is larger than in ordinary percolation, showing that the transition is less sharp, i.e., further from discontinuity, than for isolated networks. Possible consequences for spatially embedded networks are discussed.     

Percolation on Interdependent Networks with a Fraction of Antagonistic Interactions

Recently, the percolation transition has been characterized on interacting networks both in presence of interdependent interactions and in presence of antagonistic interactions. Here we characterize the phase diagram of the percolation transition in two Poisson interdependent networks with a percentage q of antagonistic nodes. We show that this system can present a bistability of the steady state solutions, and both discontinuous and continuous phase transitions. In particular, we observe a bistability of the solutions in some regions of the phase space also for a small fraction of antagonistic interactions 0<q<0.4. Moreover, we show that a fraction q>q _c =2/3 of antagonistic interactions is necessary to strongly reduce the region in phase-space in which both networks are percolating. This last result suggests that interdependent networks are robust to the presence of antagonistic interactions. Our approach can be extended to multiple networks, and to complex boolean rules for regulating the percolation phase transition.     

The Combined Effect of Connectivity and Dependency Links on Percolation of Networks

Percolation theory is extensively studied in statistical physics and mathematics with applications in diverse fields. However, the research is focused on systems with only one type of links, connectivity links. We review a recently developed mathematical framework for analyzing percolation properties of realistic scenarios of networks having links of two types, connectivity and dependency links. This formalism was applied to study Erdős-Rényi (ER) networks that include also dependency links. For an ER network with average degree [`(k)]\bar{k} that is composed of dependency clusters of size s, the fraction of nodes that belong to the giant component, P <sub>∞</sub>, is given by P<sub>￥</sub>=p<sup>s-1</sup>[1-exp(-[`(k)]pP_￥) ]^sP_{\infty}=p^{s-1}[1-\exp{(-\bar{k}pP_{\infty})} ]^{s} where 1−p is the initial fraction of randomly removed nodes. Here, we apply the formalism to the study of random-regular (RR) networks and find a formula for the size of the giant component in the percolation process: P _∞=p ^s−1(1−r ^k ) ^s where r is the solution of r=p ^s (r ^k−1−1)(1−r ^k )+1, and k is the degree of the nodes. These general results coincide, for s=1, with the known equations for percolation in ER and RR networks respectively without dependency links. In contrast to s=1, where the percolation transition is second order, for s>1 it is of first order. Comparing the percolation behavior of ER and RR networks we find a remarkable difference regarding their resilience. We show, analytically and numerically, that in ER networks with low connectivity degree or large dependency clusters, removal of even a finite number (zero fraction) of the infinite network nodes will trigger a cascade of failures that fragments the whole network. Specifically, for any given s there exists a critical degree value, [`(k)]<sub>min</sub>\bar{k}_{\min}, such that an ER network with [`(k)] ￡ [`(k)]_min\bar{k}\leq \bar{k}_{\min} is unstable and collapse when removing even a single node. This result is in contrast to RR networks where such cascades and full fragmentation can be triggered only by removal of a finite fraction of nodes in the network.     

Aging dynamics and the topology of inhomogenous networks

We study phase ordering on networks and we establish a relation between the exponent a(x) of the aging part of the integrated auto-response function and the topology of the underlying structures. We show that a(x) > 0 in full generality on networks which are above the lower critical dimension d(L), i.e., where the corresponding statistical model has a phase transition at finite temperature. For discrete symmetry models on finite ramified structures with T(c) = 0, which are at the lower critical dimension d(L), we show that a(x) is expected to vanish. We provide numerical results for the physically interesting case of the 2 - d percolation cluster at or above the percolation threshold, i.e., at or above d(L), and for other networks, showing that the value of a(x) changes according to our hypothesis. For O(N) models we find that the same picture holds in the large-N limit and that a(x) only depends on the spectral dimension of the network.     

An anti-attack model based on complex network theory in P2P networks

Complex network theory is a useful way to study many real systems. In this paper, an anti-attack model based on complex network theory is introduced. The mechanism of this model is based on a dynamic compensation process and a reverse percolation process in P2P networks. The main purpose of the paper is: (i) a dynamic compensation process can turn an attacked P2P network into a power-law (PL) network with exponential cutoff; (ii) a local healing process can restore the maximum degree of peers in an attacked P2P network to a normal level; (iii) a restoring process based on reverse percolation theory connects the fragmentary peers of an attacked P2P network together into a giant connected component. In this way, the model based on complex network theory can be effectively utilized for anti-attack and protection purposes in P2P networks.     

相依网络的条件依赖群逾渗

相依网络鲁棒性研究多集中于满足无反馈条件的一对一依赖,但现实网络节点往往依赖于多节点构成的依赖群,即使群内部分节点失效也不会导致依赖节点失效.针对此现象提出了一种相依网络的条件依赖群逾渗模型,该模型允许依赖群内节点失效比例不超过容忍度γ时,依赖节点仍可正常工作.通过理论分析给出了基于生成函数方法的模型巨分量方程,仿真结果表明方程理论解与相依网络模拟逾渗值相吻合,增大γ值和依赖群规模可提高相依网络鲁棒性.本文模型有助于更好地理解现实网络逾渗现象,对如何增强相依网络鲁棒性有一定指导作用.     

Robustness of spatial networks and networks of networks

Many complex networks have recently been recognized to involve significant interdependence between different systems. Motivation comes primarily from infrastructures like power grids and communications networks, but also includes areas such as the human brain and finance. Interdependence implies that when components in one system fail, they lead to failures in the same system or other systems. This can then lead to additional failures finally resulting in a long cascade that can cripple the entire system. Furthermore, many of these networks, in particular infrastructure networks, are embedded in space and thus have unique spatial properties that significantly decrease their resilience to failures. Here we present a review of novel results on interdependent spatial networks and how cascading processes are affected by spatial embedding. We include various aspects of spatial embedding such as cases where dependencies are spatially restricted and localized attacks on nodes contained in some spatial region of the network. In general, we find that spatial networks are more vulnerable when they are interdependent and that they are more likely to undergo abrupt failure transitions than interdependent non-embedded networks. We also present results on recovery in spatial networks, the nature of cascades due to overload failures in these networks, and some examples of percolation features found in real-world traffic networks. Finally, we conclude with an outlook on future possible research directions in this area.     

Resilience of the internet to random breakdowns

A common property of many large networks, including the Internet, is that the connectivity of the various nodes follows a scale-free power-law distribution, P(k) = ck(-alpha). We study the stability of such networks with respect to crashes, such as random removal of sites. Our approach, based on percolation theory, leads to a general condition for the critical fraction of nodes, p(c), that needs to be removed before the network disintegrates. We show analytically and numerically that for alpha0.99.     

一种全局同质化相依网络耦合模式

相依网络的相依模式(耦合模式)是影响其鲁棒性的重要因素之一.本文针对具有无标度特性的两个子网络提出一种全局同质化相依网络耦合模式.该模式以子网络的总度分布均匀化为原则建立相依网络的相依边,一方面压缩度分布宽度,提高其对随机失效的抗毁性,另一方面避开对度大节点(关键节点)的相依,提高其对蓄意攻击的抗毁性.论文将其与常见的节点一对一的同配、异配及随机相依模式以及一对多随机相依模式作了对比分析,仿真研究其在随机失效和蓄意攻击下的鲁棒性能.研究结果表明,本文所提全局同质化相依网络耦合模式能大大提高无标度子网络所构成的相依网络抗级联失效能力.本文研究成果能够为相依网络的安全设计等提供指导意义.     

Connectivity and percolation behaviour of grain boundary networks in three dimensions

Grain boundary networks are subject to crystallographic constraints at both triple junctions (first-order constraints) and quadruple nodes (second-order constraints). First-order constraints are known to influence the connectivity and percolation behaviour in two-dimensional grain boundary networks, and here we extend these considerations to fully three-dimensional microstructures. Defining a quadruple node distribution (QND) to quantify both the composition and isomerism of quadruple nodes, we explore how the QNDs for crystallographically consistent networks differ from that expected in a randomly assembled network. Configurational entropy is used to quantify the relative strength of each type of constraint (i.e., first- and second-order), with first-order triple junction constraints accounting for at least 75% of the non-random correlations in the network. As the dominant effects of constraint are captured by considering the triple junctions alone, a new analytical model is presented which allows the 3-D network connectivity to be estimated from data on 2-D microstructural sections. Finally, we show that the percolation thresholds for 3-D crystallographically consistent networks differ by as much as ±0.07 from those of standard percolation theory.     

面向级联失效的相依网络鲁棒性研究

针对相依网络耦合强度、子网络边以及耦合边对网络鲁棒性影响的问题,基于三种典型网络模型,建立对称相依网络和不对称相依网络模型.针对六种不同的相依网络模型,计算其网络临界成本,比较耦合边权值和子网络边权值对相依网络成本的贡献程度,发现耦合边对网络的贡献更大.采用仿真和理论证明的方法,获得使网络具有最小网络成本时的子网络负载参数α值和耦合强度参数β值,并证明了网络成本变化趋势与该参数对有关.以网络成本作为鲁棒性测度的变量,通过对六种相依网络模型进行级联失效仿真,给出了网络具有最强鲁棒性时参数对的取值,以及网络鲁棒性与耦合强度之间的关系,发现网络鲁棒性并不是随着耦合强度单调地增加或减少.     

Impurity diffusion in a harmonic potential and nonhomogeneous temperature

Classical blockmodel is known as the simplest among models of networks with community structure. The model can be also seen as an extremely simply example of interconnected networks. For this reason, it is surprising that the percolation transition in the classical blockmodel has not been examined so far, although the phenomenon has been studied in a variety of much more complicated models of interconnected and multiplex networks. In this paper we derive the self-consistent equation for the size the global percolation cluster in the classical blockmodel. We also find the condition for percolation threshold which characterizes the emergence of the giant component. We show that the discussed percolation phenomenon may cause unexpected problems in a simple optimization process of the multilevel network construction. Numerical simulations confirm the correctness of our theoretical derivations.     

Viscoelasticity of single wall carbon nanotube suspensions

We investigate the viscoelastic properties of an associating rigid rod network: aqueous suspensions of surfactant stabilized single wall carbon nanotubes (SWNTs). The SWNT suspensions exhibit a rigidity percolation transition with an onset of solidlike elasticity at a volume fraction of 0.0026; the percolation exponent is 2.3+/-0.1. At large strain, the solidlike samples show volume fraction dependent yielding. We develop a simple model to understand these rheological responses and show that the shear dependent stresses can be scaled onto a single master curve to obtain an internanotube interaction energy per bond approximately 40k(B)T. Our experimental observations suggest SWNTs in suspension form interconnected networks with bonds that freely rotate and resist stretching. Suspension elasticity originates from bonds between SWNTs rather than from the stiffness or stretching of individual SWNTs.     

Limited path percolation in complex networks

We study the stability of network communication after removal of a fraction q=1-p of links under the assumption that communication is effective only if the shortest path between nodes i and j after removal is shorter than al(ij)(a> or =1) where l(ij) is the shortest path before removal. For a large class of networks, we find analytically and numerically a new percolation transition at p(c)=(kappa(0)-1)((1-a)/a), where kappa(0) [triple bond] / and k is the node degree. Above p(c), order N nodes can communicate within the limited path length al(ij), while below p(c), N(delta) (delta<1) nodes can communicate. We expect our results to influence network design, routing algorithms, and immunization strategies, where short paths are most relevant.     

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.     

具有双峰特性的双层超网络模型

随着社会经济的快速发展,社会成员及群体之间的关系呈现出了更复杂、更多元化的特点.超网络作为一种描述复杂多元关系的网络,已在不同领域中得到了广泛的应用.服从泊松度分布的随机网络是研究复杂网络的开创性模型之一,而在现有的超网络研究中,基于ER随机图的超网络模型尚属空白.本文首先在基于超图的超网络结构中引入ER随机图理论,提出了一种ER随机超网络模型,对超网络中的节点超度分布进行了理论分析,并通过计算机仿真了在不同超边连接概率条件下的节点超度分布情况,结果表明节点超度分布服从泊松分布,符合随机网络特征并且与理论推导相一致.进一步,为更准确有效地描述现实生活中的多层、异质关系,本文构建了节点超度分布具有双峰特性,层间采用随机方式连接,层内分别为ER-ER,BA-BA和BA-ER三种不同类型的双层超网络模型,理论分析得到了三种双层超网络节点超度分布的解析表达式,三种双层超网络在仿真实验中的节点超度分布均具有双峰特性.