基于加权复杂网络的文本关键词提取

通过分析基于复杂网络的网络提取的文本关键词新算法.首先根据文本特征词之间的关系构建文本的加权复杂网络模型,其次通过节点的加权聚类系数和节点的介数计算节点的综合特征值,最后根据综合特征值提取出文本关键词.实验结果表明,该算法提取的关键词能够较好地体现文本主题,提取关键词的准确率比已有算法有明显提高.     

基于全局可达点的中国教育网平均最短路径近似算法

平均最短路径长度是复杂网络的一个重要特性,但由于计算时间的限制,求解大规模网络的平均最短路径长度很困难.以中国教育网数据为例,分析了中国教育网的拓扑结构,提出了全局可达点和局部可达点的概念,发现整个网络的平均最短路径长度由全局可达点决定.通过分析全局可达点的平均单源最短路径长度分布,发现整个网络的平均最短路径长度可由少数随机选取的点的平均最短路径长度来近似.通过三个网络验证了近似计算方法的有效性,并通过随机选取的数百个点,计算得到了含49041472个点的中国教育网的平均最短路径长度在14-15之间.     

检测网络社团结构的基于最短路径的$k$-均值算法

我们考虑复杂网络社团结构的检测问题,即检测出那些具有高于平均密度的边所连接的节点的集合.本文我们利用模拟退火策略来极大化可表示为稳定效益函数的模量(modularity),并结合基于最短路径的$k$-均值迭代过程来对网络进行分区.该算法不仅能检测出社团,而且能够识别出在最短路径度量下,该社团中位于中心位置的节点.社团的最优数目可以在无需任何关于网络结构的先验信息下自动确定.对人工生成网络和真实世界中的网络的成功应用表明了算法的有效性.     

非平衡交通分配的拟Frank-Wolfe迭代算法

提出了基于最短路动态生成的一种新的非平衡交通分配迭代算法.在每轮迭代中,将按全有全无方法在当前最短路上分配的交通量与前一轮迭代所得到的交通量加权组合,而各O-D对的加权系数则依据Logit原则来确定.和Frank-Wolfe算法不同,不必通过一维搜索确定加权系数.同时又避免了Logit方法要求枚举所有路径的困难.本文还证明了算法的收敛性,而计算实例显示,由本算法所得结果与平衡交通分配非常接近,因而它是一个高效而可靠的交通分配算法,适用于大、中型道路交通网络的交通分配计算.     

复杂网络的可控性及算法

复杂网络可控性问题的核心在于解决两个最少问题:最少需要多少输入和最少需要控制多少节点.Liu和Barabasi将现代控制理论应用到线性系统的结构可控性问题上,提出了最少需要多少输入的计算方法,解决了复杂有向网络结构能控性的可计算问题.针对现实网络中存在的无向图或者加权图,文章引入了节点控制能力的概念,利用添加输入或者连边的方法,给出了求系统精确可控的最小驱动点集合和极小被控点集合的算法.     

一种基于Dijkstra算法的机器人避障问题路径规划

移动机器人的避障问题是移动机器人控制领域的研究热点.针对给定的移动机器人避障问题,探讨了最短路径及最短时间路径的路径规划问题.对于最短路径问题,建立了简化的路径网格模型,将其抽象为由节点及边构成的两维图,再使用经典的Dijkstra算法获得可行的最短路径.对于最短时间路径问题,通过分析移动机器人弯道运行的速度曲线,基于几何方法得出了移动时间与过渡圆弧圆心之间严格的数学关系,此后借助MATLAB优化函数获得最佳的移动路径.算法可为类似机器人避障问题的解决提供借鉴.     

复杂网络在双曲空间的节点动态择优路径研究

本文研究复杂网络双曲嵌套模型.利用改进克林伯格和克莱尔科夫网络拓扑模型的方法,得到了复杂网络在双曲空间的动态择优路径,推广和发展了复杂网络节点间最优路径的算法.     

基于网络拓扑结构的重要节点发现算法

复杂网络中的重要节点发现在现实生活中有着广泛的应用价值。传统重要节点发现方法可分为局部发现和全局发现两类算法,全局发现算法中最具代表性的是特征向量中心性算法(Eigenvector Centrality, EC),EC算法将所有节点归为一个社区并利用邻居节点重要性反馈计算节点的影响力大小,具有较高的计算效率和识别精度。但是,EC算法忽略了网络的拓扑结构,未考虑到真实网络中节点所在社区的结构特征。为此,本文提出一种基于网络拓扑结构的可达中心性算法(Accessibility Centrality, AC),首先利用邻接矩阵作为反馈路径,在反馈过程中计算不同路径下的节点整体影响力。同时,利用影响力传递过程中的噪音干扰特性,修正每一路径长度下节点整体影响力大小,最后利用修正结果得到AC值。为评估AC算法,本文利用两种传染病模型模拟节点影响力在四组真实网络中的传播过程,并引入其他四种算法进行对比验证。实验结果表明,与其他算法相比,AC算法可以更准确、有效地识别出有具有影响力的重要节点。     

基于Erdos1合著网络和论文引用网络的复杂网络节点特征研究

首先,研究了Erdos1合著网络的特征属性,一方面使用节点的度、介数、接近中心性来描述Erdos1合著网络节点重要性,另一方面使用特征向量中心性和本文提出的高阶度参数来描述Erdos1合著网络节点影响力;然后,分别用逼近理想解的排序算法（TOPSIS算法）和主成份分析（PCA）对节点重要性和影响力排序;最后,利用修改的网页排名算法（PageRank算法）讨论了网络科学原创性论文中最具影响力的论文。     

有向循环图寻径控制

有向循环图 G(N ;1 ,s)作为有向双环网的图论模型备受关注 .本文将图的点集分划为几个不交子集 ,找到任意节点对之间路径沿跳长为 1和跳长为 s的边数的上确界 .找到了判断节点对间最短路径的充要条件 ,利用点集的分布特征设计了一个最优寻径算法 .对双环网络的容错路径进行了深入研究 ,给出了容错直径公式 ,提出了一个最优容错路径算法 .     

A new ranking algorithm and an evolutionary model based on comprehensive weighted clique degree for complex network

This paper proposes a new ranking algorithm based on comprehensive weighted clique degree (CWCDR) for ranking importance of nodes in complex network. Simulation results show that CWCDR algorithm can overcome the limitation of degree ranking algorithm and find important nodes in complex networks more precisely and effectively. To the shortage of small-world model and BA model, this paper proposes an evolutionary model of complex network based on CWCDR algorithms, named CWCDR model. Simulation results show that the CWCDR model accords with power-law distribution. Compared with BA model, this model has better average shortest path length and clustering coefficient. Therefore, the CWCDR model is more consistent with the real network.     

Communicability graph and community structures in complex networks

We use the concept of the network communicability [E. Estrada, N. Hatano, Communicability in complex networks, Phys. Rev. E 77 (2008) 036111] to define communities in a complex network. The communities are defined as the cliques of a “communicability graph”, which has the same set of nodes as the complex network and links determined by the communicability function. Then, the problem of finding the network communities is transformed to an all-clique problem of the communicability graph. We discuss the efficiency of this algorithm of community detection. In addition, we extend here the concept of the communicability to account for the strength of the interactions between the nodes by using the concept of inverse temperature of the network. Finally, we develop an algorithm to manage the different degrees of overlapping between the communities in a complex network. We then analyze the USA airport network, for which we successfully detect two big communities of the eastern airports and of the western/central airports as well as two bridging central communities. In striking contrast, a well-known algorithm groups all but two of the continental airports into one community.     

Generalized function projective (lag, anticipated and complete) synchronization between two different complex networks with nonidentical nodes

Generalized function projective (lag, anticipated and complete) synchronization between two different complex networks with nonidentical nodes is investigated in this paper. Based on Barbalat’s lemma, some sufficient synchronization criteria are derived by applying the nonlinear feedback control. Although previous work studied function projective synchronization on complex dynamical networks, the dynamics of the nodes are coupled partially linear chaotic systems. In our work, the dynamics of the nodes of the complex networks are any chaotic systems without the limitation of the partial linearity. In addition, each network can be undirected or directed, connected or disconnected, and nodes in either network may have identical or different dynamics. The proposed strategy is applicable to almost all kinds of complex networks. Numerical simulations further verify the effectiveness and feasibility of the proposed synchronization method. Numeric evidence shows that the synchronization rate is sensitively influenced by the feedback strength, the time delay, the network size and the network topological structure.     

Adaptive synchronization of two nonlinearly coupled complex dynamical networks with delayed coupling

This paper investigates the adaptive synchronization between two nonlinearly delay-coupled complex networks with the bidirectional actions and nonidentical topological structures. Based on LaSalle’s invariance principle, some criteria for the synchronization between two coupled complex networks are achieved via adaptive control. To validate the proposed methods, the unified chaotic system as the nodes of the networks are analyzed in detail, and numerical simulations are given to illustrate the theoretical results.     

Robust synchronization of singular complex switched networks with parametric uncertainties and unknown coupling topologies via impulsive control

This paper presents a general model of singular complex switched networks, in which the nodes can be singular dynamic systems and switching behaviors act on both nodes and edges. The parametric uncertainties and unknown coupling topologies are also considered in this model. Two robust synchronization schemes are discussed respectively. In one scheme, the network is synchronized to a homogeneous orbit and in the other one the network is synchronized to a weighted average of all the nodes. Based on the Lyapunov stability theory, different robust synchronization conditions for the two schemes are obtained for this singular complex switched network model via impulsive control. The similarities and differences between these synchronization conditions for the two schemes are discussed. In addition, three useful robust results for the special cases of the singular complex switched networks are presented. Two systematic-design procedures are presented for the two schemes, and three numerical examples are provided for illustrations.     

Projective lag synchronization of the general complex dynamical networks with distinct nodes

In this paper, projective lag synchronization of the general complex dynamical networks with different nodes is investigated. Combining Barbalat’s lemma with adaptive control technique, the adaptive feedback controllers are constructed to achieve projective lag synchronization between the dynamical network with diverse nodes and arbitrary desired trajectory. The presented synchronization method can be applied to any complex networks. It is discovered that the update gains, the time delay, the network size and the network topology have influence on the synchronization effect. Furthermore, projective lag synchronization of the dynamical networks can still be efficiently realized in presence of noise and parameter perturbations. Corresponding numerical simulations are performed to validate the effectiveness and robustness of the proposed synchronization scheme.     

Finite‐time stabilization of complex dynamical networks via optimal control

An optimal control scheme is proposed to stabilize complex networks in finite time. Furthermore, since it is costly and impractical to control a network by applying controllers to all the nodes, an algorithm inspired by Kalmans controllability rank condition is presented for local stabilization by locating pinned components. Numerical examples are provided to illustrate the effectiveness of the proposed method as well as its superiority over a traditional pinning control technique. This work offers a theoretical framework for designing optimal controllers to stabilize networks in finite time with reduced control cost. © 2016 Wiley Periodicals, Inc. Complexity 21: 417–425, 2016     

Stabilization and set stabilization of switched Boolean control networks via flipping mechanism

Stabilization and set stabilization of switched Boolean control networks are investigated by using flipping mechanism in this paper. Firstly, with the help of Warshall algorithm, an explicit criterion for the stabilization of switched Boolean control networks is derived. Secondly, the necessary and sufficient condition for the solvability of stabilization of switched Boolean control networks, by flipping some elements of perturbation set once, is presented. Thirdly, a search algorithm is proposed to calculate the minimum number of stabilization flipped nodes and what exactly they are. Furthermore, a necessary and sufficient condition is established for the solvability of set stabilization of switched Boolean control networks by flipping some elements of perturbation set once. Analogously, an algorithm is given to find the minimum number of set stabilization flipped nodes. Finally, examples are shown to demonstrate the feasibility of the above results.