Leverage centrality analysis of infrastructure networks

Leverage centrality is a novel centrality measure proposed to identify the critical nodes that are highly influential within the network. Leverage centrality considers the extent of connectivity of a node relative to the connectivity of its neighbors. The leverage centrality of a node in a network is determined by the extent to which its direct neighbors rely on that node for information. In this paper, leverage centralities of the nodes of infrastructure networks are computed and critical nodes within the network are identified.     

Modeling centrality measures in social network analysis using bi-criteria network flow optimization problems

Centrality measures play an important role in the field of network analysis. In the particular case of social networks, the flow represents the way in which information passes through the network nodes. Freeman et al. (1991) were the first authors to relate centrality measures to network flow optimization problems in terms of betweenness, closeness, and the influence of one node over another one. Such measures are single dimensional and, in general, they amalgamate several heterogeneous dimensions into a single one, which is not suitable for dealing with most real-world problems. In this paper we extend the betweenness centrality measure (or concept) to take into account explicitly several dimensions (criteria). A new closeness centrality measure is defined to deal not only with the maximum flow between every ordered pair of nodes, but also with the cost associated with communications. We shall show how the classical measures can be enhanced when the problem is modeled as a bi-criteria network flow optimization problem.     

Measuring centrality by a generalization of degree

Network analysis has emerged as a key technique in communication studies, economics, geography, history and sociology, among others. A fundamental issue is how to identify key nodes, for which purpose a number of centrality measures have been developed. This paper proposes a new parametric family of centrality measures called generalized degree. It is based on the idea that a relationship to a more interconnected node contributes to centrality in a greater extent than a connection to a less central one. Generalized degree improves on degree by redistributing its sum over the network with the consideration of the global structure. Application of the measure is supported by a set of basic properties. A sufficient condition is given for generalized degree to be rank monotonic, excluding counter-intuitive changes in the centrality ranking after certain modifications of the network. The measure has a graph interpretation and can be calculated iteratively. Generalized degree is recommended to apply besides degree since it preserves most favourable attributes of degree, but better reflects the role of the nodes in the network and has an increased ability to distinguish among their importance.     

A new centrality measure of nodes in multilayer networks under the framework of tensor computation

One challenging issue in information science, biological systems and many other field is to determine the most central agents in multilayer networked systems characterized by different types of interrelationships. In this paper, using a fourth-order tensor to represent multilayer networks, we propose a new centrality measure, referred to as the Singular Vector of Tensor (SVT) centrality, which is used to quantitatively evaluate the importance of nodes connected by different types of links in multilayer networks. First, we present a novel iterative method to obtain four alternative metrics that can quantify the hub and authority scores of nodes and layers in multilayer networked systems. Moreover, we use the theory of multilinear algebra to prove that the four metrics converge to four singular vectors of the adjacency tensor of the multilayer network under reasonable conditions. Furthermore, a novel SVT centrality measure is obtained by integrating these four metrics. The experimental results demonstrate that the proposed method is a new centrality measure that significantly outperforms six other published centrality methods on two real-world multilayer networks related to complex diseases, i.e., gastric and colon cancers.     

Ranking hubs and authorities using matrix functions

The notions of subgraph centrality and communicability, based on the exponential of the adjacency matrix of the underlying graph, have been effectively used in the analysis of undirected networks. In this paper we propose an extension of these measures to directed networks, and we apply them to the problem of ranking hubs and authorities. The extension is achieved by bipartization, i.e., the directed network is mapped onto a bipartite undirected network with twice as many nodes in order to obtain a network with a symmetric adjacency matrix. We explicitly determine the exponential of this adjacency matrix in terms of the adjacency matrix of the original, directed network, and we give an interpretation of centrality and communicability in this new context, leading to a technique for ranking hubs and authorities. The matrix exponential method for computing hubs and authorities is compared to the well known HITS algorithm, both on small artificial examples and on more realistic real-world networks. A few other ranking algorithms are also discussed and compared with our technique. The use of Gaussian quadrature rules for calculating hub and authority scores is discussed.     

Some New Results on the Eigenvector Centrality

In this article we establish new results on the components of the principal eigenvector in an undirected graph. Those results are particularly significant in relation to the concept of centrality in social networks. In particular degree centrality and eigenvector centrality are compared. We find further conditions, based on the spectral radius, on which nodes with highest degree centrality are also the most eigencentral.     

Quasi-stationary distributions as centrality measures for the giant strongly connected component of a reducible graph

A random walk can be used as a centrality measure of a directed graph. However, if the graph is reducible the random walk will be absorbed in some subset of nodes and will never visit the rest of the graph. In Google PageRank the problem was solved by the introduction of uniform random jumps with some probability. Up to the present, there is no final answer to the question about the choice of this probability. We propose to use a parameter-free centrality measure which is based on the notion of a quasi-stationary distribution. Specifically, we suggest four quasi-stationary based centrality measures, analyze them and conclude that they produce approximately the same ranking.     

A genetic algorithm with local search strategy for improved detection of community structure

On the basis of modularity optimization, a genetic algorithm is proposed to detect community structure in networks by defining a local search operator. The local search operator emphasizes two features: one is that the connected nodes in a network should be located in the same community, while the other is “local selection” inspired by the mechanisms of efficient message delivery underlying the small‐world phenomenon. The results of community detection for some classic networks, such as Ucinet and Pajek networks, indicate that our algorithm achieves better community structure than other methodologies based on modularity optimization, such as the algorithms based on betweenness analysis, simulated annealing, or Tasgin and Bingol's genetic algorithm. © 2009 Wiley Periodicals, Inc. Complexity, 2010     

Scatter search for the cutwidth minimization problem

The cutwidth minimization problem consists of finding a linear layout of a graph so that the maximum linear cut of edges is minimized. This NP-hard problem has applications in network scheduling, automatic graph drawing and information retrieval. We propose a heuristic method based on the Scatter Search (SS) methodology for finding approximate solutions to this optimization problem. This metaheuristic explores solution space by evolving a set of reference points. Our SS method is based on a GRASP constructive algorithm, a local search strategy based on insertion moves and voting-based combination methods. We also introduce a new measure to control the diversity in the search process. Empirical results with 252 previously reported instances indicate that the proposed procedure compares favorably to previous metaheuristics for this problem, such as Simulated Annealing and Evolutionary Path Relinking.     

An efficient multi-objective PSO algorithm assisted by Kriging metamodel for expensive black-box problems

The huge computational overhead is the main challenge in the application of community based optimization methods, such as multi-objective particle swarm optimization and multi-objective genetic algorithm, to deal with the multi-objective optimization involving costly simulations. This paper proposes a Kriging metamodel assisted multi-objective particle swarm optimization method to solve this kind of expensively black-box multi-objective optimization problems. On the basis of crowding distance based multi-objective particle swarm optimization algorithm, the new proposed method constructs Kriging metamodel for each expensive objective function adaptively, and then the non-dominated solutions of the metamodels are utilized to guide the update of particle population. To reduce the computational cost, the generalized expected improvements of each particle predicted by metamodels are presented to determine which particles need to perform actual function evaluations. The suggested method is tested on 12 benchmark functions and compared with the original crowding distance based multi-objective particle swarm optimization algorithm and non-dominated sorting genetic algorithm-II algorithm. The test results show that the application of Kriging metamodel improves the search ability and reduces the number of evaluations. Additionally, the new proposed method is applied to the optimal design of a cycloid gear pump and achieves desirable results.     

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.     

基于仿射传播的道路网络聚类

对于道路网络聚类问题,提出了仿射传播算法。首先,将道路网络上的交叉路口和结点作为顶点,建立了无向图;然后,根据最短路径计算网络距离,进而得到图的相似度矩阵;并基于仿射传播算法对道路网络进行聚类;最后,试验结果证实了本文方法的有效性与稳定性。     

An algorithm for detecting community structure of social networks based on prior knowledge and modularity

An algorithm is proposed to detect community structure in social network. The algorithm begins with a community division based on prior knowledge of the degrees of the nodes, and then combines the communities until a clear partition is obtained. In applications such as a computer‐generated network, Ucinet networks, and Chinese rural‐urban migrants' social networks, the algorithm can achieve higher modularity and greater speed than others in the recent literature. © 2007 Wiley Periodicals, Inc. Complexity 12: 53–60, 2007     

A multicast routing algorithm based on searching in directed graph

Aiming at constructing a delay and delay variation bounded Steiner tree in the real-time streaming media communication, in this paper, we discuss a multicast routing algorithm based on searching a directed graph (MRASDH). During the process of the construction of the multicast tree, some nodes and links in the network topology do not affect the outcome of the constructed tree. Therefore, based on the thought of shrinking the search space through deleting these non-relative nodes and edges to the utmost, the ant algorithm is utilized to generate a directed sub-graph of the network topology for each destination node, in which each node owns a bounded out-degree. And all these sub-graphs can be merged into a new directed graph that serves as the new search space. In the new space, the simulated annealing algorithm is applied to obtain a multicast tree that satisfies the condition for the optimization. The performance analysis and simulation results demonstrate that this algorithm can effectively construct a delay and delay variation bounded multicast tree. They also show that the algorithm have lower time complexity than the current ones, which means a much better result would be achieved when the system scale rises greatly.     

Gaussian kernel approximation algorithm for feedforward neural network design

A Gaussian kernel approximation algorithm for a feedforward neural network is presented. The approach used by the algorithm, which is based on a constructive learning algorithm, is to create the hidden units directly so that automatic design of the architecture of neural networks can be carried out. The algorithm is defined using the linear summation of input patterns and their randomized input weights. Hidden-layer nodes are defined so as to partition the input space into homogeneous regions, where each region contains patterns belonging to the same class. The largest region is used to define the center of the corresponding Gaussian hidden nodes. The algorithm is tested on three benchmark data sets of different dimensionality and sample sizes to compare the approach presented here with other algorithms. Real medical diagnoses and a biological classification of mushrooms are used to illustrate the performance of the algorithm. These results confirm the effectiveness of the proposed algorithm.     

Handling equality constraints in evolutionary optimization

Over the last few decades several methods have been proposed for handling functional constraints while solving optimization problems using evolutionary algorithms (EAs). However, the presence of equality constraints makes the feasible space very small compared to the entire search space. As a consequence, the handling of equality constraints has long been a difficult issue for evolutionary optimization methods. This paper presents a Hybrid Evolutionary Algorithm (HEA) for solving optimization problems with both equality and inequality constraints. In HEA, we propose a new local search technique with special emphasis on equality constraints. The basic concept of the new technique is to reach a point on the equality constraint from the current position of an individual solution, and then explore on the constraint landscape. We believe this new concept will influence the future research direction for constrained optimization using population based algorithms. The proposed algorithm is tested on a set of standard benchmark problems. The results show that the proposed technique works very well on those benchmark problems.     

Aggregating nonnegative eigenvectors of the adjacency matrix as a measure of centrality for a directed graph

Eigenvector centrality is a popular measure that uses the principal eigenvector of the adjacency matrix to distinguish importance of nodes in a graph. To find the principal eigenvector, the power method iterating from a random initial vector is often adopted. In this article, we consider the adjacency matrix of a directed graph and choose suitable initial vectors according to strongly connected components of the graph instead so that nonnegative eigenvectors, including the principal one, can be found. Consequently, for aggregating nonnegative eigenvectors, we propose a weighted measure of centrality, called the aggregated-eigenvector centrality. Weighting each nonnegative eigenvector by the reachability of the associated strongly connected component, we can obtain a measure that follows a status hierarchy in a directed graph.     

一种连续的谱聚类优化模型

聚类与图的划分问题在大数据分析中有着重要的应用.这类问题一般被描述为组合优化问题,因此较难快速求解.本文设计了一种新的连续优化模型,并提出了一种块坐标下降算法,数值实验显示我们的新方法在求解聚类与图的划分问题上很有潜力.我们还更进一步分析了我们的连续优化模型和组合优化模型的关系.     

Mixed-Integer Linear Programming Formulations for the Software Clustering Problem

The clustering problem has an important application in software engineering, which usually deals with large software systems with complex structures. To facilitate the work of software maintainers, components of the system are divided into groups in such a way that the groups formed contain highly-interdependent modules and the independent modules are placed in different groups. The measure used to analyze the quality of the system partition is called Modularization Quality (MQ). Designers represent the software system as a graph where modules are represented by nodes and relationships between modules are represented by edges. This graph is referred in the literature as Module Dependency Graph (MDG). The Software Clustering Problem (SCP) consists in finding the partition of the MDG that maximizes the MQ. In this paper we present three new mathematical programming formulations for the SCP. Firstly, we formulate the SCP as a sum of linear fractional functions problem and then we apply two different linearization procedures to reformulate the problem as Mixed-Integer Linear Programming (MILP) problems. We discuss a preprocessing technique that reduces the size of the original problem and develop valid inequalities that have been shown to be very effective in tightening the formulations. We present numerical results that compare the formulations proposed and compare our results with the solutions obtained by the exhaustive algorithm supported by the freely available Bunch clustering tool, for benchmark problems.     

广播网络的最佳源点集

广播是研究通信网络的某个成员的消息如何尽快地传递给所有其它成员的消息传递问题,有两类常见的通信模式,一类是shouting模式,即在一个单位时间内,一个顶点能够和它的户斤有邻点通信;另一类是whispering模式,即在一个单位时间以内,一个顶点最多只能和它的一个邻点通信,通信网络通常用图来描述,最初贮存消息的网络成员称为源点。