Scaling properties of first-passage quantities on the fractal and transfractal scale free networks

TL;DR

This paper analyzes the scaling behavior of first-passage times on fractal and transfractal scale-free networks, deriving exact formulas and revealing how these quantities scale with network size and spectral properties.

Contribution

It introduces a method to exactly compute the probability generating function, mean, and variance of first-passage times on $(u,v)$ scale-free networks, including the novel transspectral dimension.

Findings

Mean GFPT scales as N_t^{2/d_s} for fractal networks.

Mean GFPT scales as N_t^{2/tilde{d}_s} for non-fractal networks.

Variance of GFPT scales quadratically with its mean.

Abstract

In this paper, we consider the random walk process on a kind of fractal (or transfractal) scale free networks, which also called as $(u,v)$ flowers, and we focus on the global first passage time (GFPT) and first return time (FRT). Here, we present method to derive exactly the probability generation function, mean and variance of the GFPT and FRT for a given hub (i.e., node with the highest degree) and then the scaling properties of the mean and the variance of the GFPT and FRT are disclosed. Our results show that, for the case of $u>1$, while the networks are fractals, the mean of the GFPT scales with the volume of the network as $\langle GMFPT_t\rangle \sim N_t^{{2}/{d_s}}$, where $\langle * \rangle$ denotes the mean of random variable $*$, $N_t$ is the volume of the network with generation $t$ and $d_s$ is the spectral dimension of the network; but, for the case of $u=1$, while the networks are nonfractals, the mean of the GFPT scales as $\langle GMFPT_t\rangle \sim N_t^{{2}/{\tilde{d}_s}}$, where $\tilde{d}_s$ is the transspectral dimension of the network, which is introduced in this paper. Results also show that, the variance of the GFPT scales as $Var(GFPT_t) \sim \langle GFPT_{t}\rangle^2$, where $Var(*)$ denotes variance of of random variable $*$; whereas the variance of the FRT scales as $Var(FRT_t) \sim \langle FRT_{t} \rangle \langle GFPT_{t}\rangle$. %$Var(GFPT_t) \sim \langle GFPT_{t}\rangle^2$ and $Var(FRT_t) \sim \langle FRT_{t} \rangle \langle GFPT_{t}\rangle$, where $Var(GFPT_t)$ and $Var(FRT_t)$ denote the variance of the GFPT and the FRT, $\langle GFPT_{t}\rangle$ and $\langle FRT_{t}\rangle$ denote the mean of the GFPT and the FRT respectively.