相关时间序列的可见性图违反Barthelemy的度和介数中心性猜想
The Visibility Graphs of Correlated Time Series Violate the Barthelemy's Conjecture for Degree and Betweenness Centralities
摘要:
介数中心性问题是复杂网络中一个尚未解决的基本问题。在Barthelemy的开创性工作之后,无标度(SF)网络的最大介数度($b$)指数是属于无标度树(SFT)的$η_{\text{max}}=2$已被广泛接受,在此基础上得出$δ\ge\frac{γ+1}{2}$,其中$γ$和$δ$分别是度和介数中心分布函数的标度指数。在这里,我们提出的证据违反这一猜想SF可见性图(VG)。为此,我们考虑了三种模型的VG:二维Bak-Tang-Weisenfeld(BTW)沙堆模型,一维分数布朗运动(FBM)和一维Levy行走,后两种情况分别由Hurst指数$H$和阶跃指数$α$控制.具体来说,对于BTW模型和具有$H\lesssim 0.5$的FBM,$η$大于$2$,并且对于BTW模型也大于$δ<\frac{γ+1}{2}$,而Barthelemy猜想对于Levy过程仍然有效。我们认为Barthelemy猜想的失败是由于$b$关系中的大的涨落导致超标度关系$η=\frac{γ-1}{δ-1}$的破坏和BTW模型和FBM的涌现反常行为。一个超普遍的行为被发现的分布函数的广义度函数相同的Barabasi-Albert网络模型。
英文摘要:
The problem of betweenness centrality remains a fundamental unsolved problem in complex networks. After a pioneering work by Barthelemy, it has been well-accepted that the maximal betweenness-degree ($b$-$k$) exponent for scale-free (SF) networks is $η_{\text{max}}=2$, belonging to scale-free trees (SFTs), based on which one concludes $δ\ge\frac{γ+1}{2}$, where $γ$ and $δ$ are the scaling exponents of the distribution functions of the degree and betweenness centrality, respectively. Here we present evidence for violation of this conjecture for SF visibility graphs (VGs). To this end, we consider the VG of three models: two-dimensional (2D) Bak-Tang-Weisenfeld (BTW) sandpile model, 1D fractional Brownian motion (FBM) and, 1D Levy walks, the two later cases are controlled by the Hurst exponent $H$ and step-index $α$, respectively. Specifically, for the BTW model and FBM with $H\lesssim 0.5$, $η$ is greater than $2$, and also $δ<\frac{γ+1}{2}$ for the BTW model, while Barthelemy's conjecture remains valid for the Levy process. We argue that this failure of Barthelemy's conjecture is due to large fluctuations in the scaling $b$-$k$ relation resulting in the violation of hyperscaling relation $η=\frac{γ-1}{δ-1}$ and emergent anomalous behaviors for the BTW model and FBM. A super-universal behavior is found for the distribution function for a generalized degree function identical to the Barabasi-Albert network model.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。