---
title: Random Walks on Random Conductances
url: https://www.emergentmind.com/topics/random-walk-among-random-conductances
type: topic
---

# Random Walks on Random Conductances

Random walk among random conductances (RWRC), often called the random conductance model, is a reversible Markov process on a random weighted graph in which each edge \(e\) carries a conductance \(\omega_e\), and transitions are taken with probabilities proportional to the conductances adjacent to the current vertex. On nearest-neighbor lattices this framework interpolates between uniformly elliptic reversible walks, percolation-type degeneracies with zero edges, and heavy-tailed environments with very large or very small weights. Its core problems are the asymptotic speed, recurrence and transience, quenched invariance principles, heat-kernel and Green-kernel asymptotics, and trap-driven anomalous scaling under degeneracy or bias [2504.06690]. In two dimensions, an especially sharp result is that a single-edge log-moment condition with exponent strictly larger than one forces zero speed in stationary ergodic symmetric environments, while below that threshold one can construct ballistic walks or walks whose limiting speed does not exist [1205.5449].

## 1. Probabilistic formulation and main variants

In the standard lattice formulation, one works on \(\mathbb{Z}^d\) with nearest-neighbor edge set \(E_d=\{\{x,y\}:|x-y|=1\}\), and an environment is a field \(\omega=(\omega(e))_{e\in E_d}\) with \(\omega(e)\ge 0\) or \(\omega(e)>0\), depending on the model. Stationarity and ergodicity under lattice shifts are the basic structural assumptions, and conductances are symmetric on undirected edges. For the discrete-time walk, the quenched transition kernel is
\[
p^\omega(x,y)=\frac{\omega(x,y)}{\mu^\omega(x)},\qquad \mu^\omega(x)=\sum_{z\sim x}\omega(x,z),
\]
so the walk is reversible with respect to \(\mu^\omega\) or, in equivalent notation, \(\pi^\omega(x)=\sum_{y\sim x}\omega(x,y)\) [1205.5449]. Continuous-time versions are equally central. The variable-speed random walk (VSRW) has generator
\[
Lf(x)=\sum_{y\sim x}\omega(x,y)\bigl(f(y)-f(x)\bigr),
\]
while the constant-speed random walk (CSRW) normalizes by \(\mu^\omega(x)\) and waits mean \(1\) at each site [1902.05793].

This reversible structure is retained in biased models by tilting conductances rather than transition probabilities. A standard choice is
\[
c^\omega(x,y)=c_*^\omega([x,y])\,e^{(x+y)\cdot \ell},
\]
with \(\ell=\lambda \hat \ell\), so that the biased walk remains reversible with respect to \(\pi^\omega(x)=\sum_{y\sim x}c^\omega(x,y)\) while favoring motion in direction \(\hat \ell\) [1107.0706]. This formulation is technically important because many estimates—Dirichlet-form methods, electrical-network arguments, heat-kernel bounds, and regeneration constructions—continue to operate in a reversible setting. The resulting model is therefore broad enough to cover unbiased diffusion, degenerate trapping, directional transience, and heavy-tailed sub-ballistic motion within a single conductance-based formalism [2507.17583].

## 2. Reversibility, correctors, and homogenization

A central organizing principle is the harmonic-coordinate decomposition. For VSRW in stationary ergodic environments, one constructs a corrector \(\chi\) and harmonic coordinate \(\Phi(\omega,x)=x-\chi(\omega,x)\) such that
\[
\sum_{y\sim x}\omega(x,y)\bigl(\Phi(\omega,y)-\Phi(\omega,x)\bigr)=0.
\]
Then \(M_t=\Phi(\omega,X_t)=X_t-\chi(\omega,X_t)\) is a quenched martingale, and its quadratic variation identifies the homogenized covariance matrix \(\Sigma\) [1902.05793]. The quenched invariance principle follows once \(\chi\) is shown to be sublinear on diffusive scales. In dimension \(d\ge 3\), Bella and Schaffner prove the QIP for degenerate positive conductances under the moment assumptions
\[
\mathbb{E}[\omega(e)^p]<\infty,\qquad \mathbb{E}[\omega(e)^{-q}]<\infty,\qquad \frac1p+\frac1q<\frac{2}{d-1},
\]
and identify this relation as the minimal requirement ensuring everywhere sublinear correctors [1902.05793]. The same paper emphasizes that this improves the earlier threshold \(2/d\).

On random subgraphs with zero conductances, the same corrector-and-martingale mechanism survives but requires geometric input on the infinite cluster. In particular, a quenched invariance principle for the VSRW on a unique infinite cluster of positive edges is obtained under stationarity, ergodicity, mixed moment bounds on \(\omega\) and \(\omega^{-1}\), and \(\theta\)-very regular balls; an anchored relative isoperimetric inequality is the key new ingredient [1602.08428]. The survey literature places these results within a broader homogenization theory that also includes local limit theorems and heat-kernel estimates for degenerate conductances [2504.06690]. In two dimensions, the potential kernel and Green kernel of the walk killed outside balls admit precise quenched logarithmic asymptotics, with the leading constant determined by the covariance matrix from the invariance principle [1808.08126]. Under i.i.d. uniformly elliptic conductances, quantitative annealed CLTs are also available: Mourrat proves Berry–Esseen bounds with speed \(t^{-1/10}\) for \(d<3\) and \(t^{-1/5}\) for \(d\ge 3\), up to logarithmic corrections [1105.4485].

## 3. Speed, zero-speed criteria, and sharp thresholds in the symmetric model

For discrete-time RWRC, the speed is the almost sure limit
\[
v=\lim_{n\to\infty}\frac{X_n}{n},
\]
when it exists. In the two-dimensional stationary ergodic symmetric setting, a sharp criterion in terms of the upper tail of \(\log \omega_e\) is available: if there exists \(\alpha>1\) such that
\[
\mathbb{E}\big[(\log \omega_e)^\alpha\big]<\infty,
\]
then the walk has quenched, hence annealed, speed zero [1205.5449]. The assumptions are notably weak: stationarity and ergodicity under shifts, symmetric conductances on undirected edges, and the single-edge log-moment condition. Uniform ellipticity is not required, and the proof does not use mixing [1205.5449].

The same work shows that the threshold is sharp in dimension two. For every \(\alpha<1\), there exist stationary, ergodic laws with equal horizontal and vertical marginals such that \(\mathbb{E}[(\log \omega_e)^\alpha]<\infty\) but speed zero fails; one may arrange either ballisticity or almost sure non-existence of \(\lim_{n\to\infty}X_n/n\) [1205.5449]. These constructions do not contradict the zero-speed theorem because their \(\alpha\)-log-moments diverge for every \(\alpha\ge 1\). The paper also remarks that the same methods should extend to \(\mathbb{Z}^d\) with critical exponent \(\alpha_c=d-1\), which would identify a higher-dimensional threshold for purely moment-based speed-zero criteria [1205.5449].

This sharpness result isolates an important asymmetry between upper and lower tails. Very small conductances can create traps and slow the walk, but the speed-zero proof in dimension two requires no lower-tail control at all. What matters is that very large conductances inflate the reversible measure
\[
\pi(x)=\sum_{y\sim x}\omega_{x,y}
\]
at distant sites; once that inflation is controlled by an \(\alpha>1\) log-moment, sublinear displacement follows [1205.5449]. The same source also notes that for symmetric reversible RWRC with i.i.d. conductances, speed zero holds regardless of the single-edge law.

## 4. Heat-kernel control, tree highways, and failure of a limiting velocity

The zero-speed proof in dimension two is based on the Varopoulos–Carne inequality. For any irreducible reversible kernel \(L\) with reversible measure \(\pi\),
\[
L^n(x,y)\le 2\sqrt{\frac{\pi(y)}{\pi(x)}}\,
\exp\!\Big(-\frac{d(x,y)^2}{2n}\Big).
\]
Under the assumption \(\mathbb{E}[(\log \omega_e)^\alpha]<\infty\) with \(\alpha>1\), Markov’s inequality and a Borel–Cantelli summability argument show that conductances adjacent to far \(L^\infty\)-boundaries are eventually bounded by stretched exponentials, hence \(\pi(y)\) is controlled on those boundaries. Summing the Varopoulos–Carne bound over boundary sites and times then yields a summable estimate for reaching distance \(m\) by time \(Mm\), from which \(\limsup_{n\to\infty}\|X_n\|_\infty/n\le 2/M\) follows; since \(M\) is arbitrary, the speed is zero [1205.5449].

Below the threshold, the obstruction is not mere trapping but structured transport along sparse high-conductance subgraphs. The counterexamples are built from directed ancestral functions \(a:\mathbb{Z}^2\to\mathbb{Z}^2\) that define spanning trees, together with “tree-conductances”
\[
\omega_{\{z,a(z)\}}=2^{(h(z)+1)^A},\qquad \omega_e=1\ \text{otherwise},
\]
where \(h(z)\) is the height of the descendant subtree and \(A>1\) [1205.5449]. Because the preferred edge \(\{z,a(z)\}\) dominates neighboring weights, the random walk eventually follows the tree deterministically after a finite random time. The geometry of the tree then determines the macroscopic behavior.

Two constructions are decisive. In the Bramson–Zeitouni–Zerner umbrella tree, nested straight umbrellas generate alternating “rushes” of horizontal and vertical motion; the walk almost surely leaves each rush in finite time and enters new ones infinitely often, producing infinitely many macroscopic changes in the empirical velocity, so \(X_n/n\) does not converge [1205.5449]. In the diagonal-tree construction, umbrellas have slopes \(\pi/4\pm 1/\log t\), so sufficiently long edges are asymptotically parallel to \(v=(1/2,1/2)\); eventual alignment with the tree then yields
\[
\lim_{n\to\infty}\frac{X_n}{n}=\bigl(\tfrac12,\tfrac12\bigr)
\]
almost surely [1205.5449]. These examples show that reversible, stationary, ergodic conductance environments can sustain macroscopic directional effects without violating any symmetry or reversibility principle, provided uniform ellipticity is absent.

## 5. Biased conductance models and anomalous transport

Bias fundamentally changes the phase diagram. For positive i.i.d. conductances on \(\mathbb{Z}^d\), \(d\ge 2\), the biased reversible walk is always transient in the direction of the bias, but its speed is ballistic if and only if the conductances have finite mean:
\[
\mathbb{E}_*[c_*]<\infty \Longleftrightarrow v\cdot \hat\ell>0,\qquad
\mathbb{E}_*[c_*]=\infty \Longrightarrow v=0.
\]
When the upper tail satisfies \(\log P_*(c_*>n)/\log n\to -\gamma\) with \(\gamma<1\), the walk is sub-ballistic and
\[
\lim_{n\to\infty}\frac{\log (X_n\cdot \hat\ell)}{\log n}=\gamma,\qquad
\lim_{n\to\infty}\frac{\log \Delta_n}{\log n}=\frac1\gamma
\]
almost surely under the annealed law [1107.0706]. The mechanism is trap formation by single high-conductance edges, analyzed through regeneration times and electrical-network estimates.

The dependence of the speed on the bias parameter can itself be delicate. In uniformly elliptic i.i.d. conductances, the velocity is eventually increasing for large bias, and under sufficiently small disorder it is strictly increasing for all \(\lambda\ge 0\); the key identity is a derivative formula expressing \(v'(\lambda)\) as a covariance under the stationary environment seen from the particle [1704.08844]. Yet monotonicity fails in general: in \(d=2\), a two-valued uniformly elliptic environment with conductances \(1\) and \(\kappa\) can exhibit \(v_1(\lambda_1)>v_1(\lambda_2)\) for \(0<\lambda_1<\lambda_2\), because stronger bias amplifies time spent in dead ends [1704.08844].

In the heavy-tailed biased regime, the asymptotic process is no longer ballistic or diffusive but fractional kinetic. For i.i.d. positive conductances with regularly varying tail \(\mathbb{P}(c_*\ge u)=L(u)u^{-\gamma}\), \(\gamma\in(0,1)\), the quenched scaling limit of the position is governed by the inverse of a \(\gamma\)-stable subordinator in every dimension \(d\ge 2\), with longitudinal motion along a deterministic direction \(v_0\) and transversal Brownian fluctuations run on the same random clock [2507.17583]. In one dimension, heavy tails at \(0\) and \(\infty\) lead to a complementary “wells and walls” picture: very small conductances act as walls, very large ones as traps, and the scaling limit of sub-ballistic biased motion is again the inverse of an \(\alpha\)-stable subordinator, with aging described by the generalized arcsine law [1904.05283]. Related one-dimensional continuous-time models exhibit aging for the maximum when only the lower tail is heavy, and aging or sub-aging for the position when both lower and upper tails are heavy [2308.02230].

## 6. Large deviations, kernels, mean-field variants, and open problems

RWRC also supports a substantial local-time and spectral theory. For continuous-time walks in a fixed finite domain \(D\subset \mathbb{Z}^d\) with i.i.d. positive conductances whose lower tail satisfies
\[
\log \mathbb{P}(c_e\le \varepsilon)\sim -D\varepsilon^{-\eta}\qquad (\varepsilon\downarrow 0),
\]
the annealed normalized local times satisfy a large deviation principle with speed \(t^{\eta+1}\) and rate
\[
J(g^2)=K_{\eta,D}\sum_{\langle x,y\rangle\in E_D}\frac{|g(y)-g(x)|^{\eta+1}}{2\eta},
\qquad
K_{\eta,D}=(1+\eta)(D\eta)^{\eta/(\eta+1)},
\]
reflecting a joint strategy of small conductance values and large holding times [1104.1548]. In growing boxes \(B_t=a_tG\cap \mathbb{Z}^d\), the annealed problem displays a phase transition at \(\nu=d/2\): for \(\nu>d/2\), the walk spreads across the entire growing box and the rate becomes a \(p\)-energy with \(p=2\nu/(\nu+1)\in(2d/(d+2),2)\); for \(\nu\le d/2\), the optimal strategy confines the walk to a bounded region [1308.4520].

At the level of kernels, two-dimensional recurrent RWRC admits precise potential-kernel and killed Green-kernel asymptotics [1808.08126]. At the opposite end, mean-field complete-graph models have a Poisson Weighted Infinite Tree limit; in that regime the empirical spectral distribution of the random transition matrix converges to a symmetric deterministic measure on \([-1,1]\), and the RWRC on the limiting tree has positive speed or weakly zero speed according to whether the mean of the largest outgoing conductance is finite or infinite [1702.04087]. These results show that conductance disorder interacts with geometry at every scale, from finite-box spectral tails to infinite-volume transport and mean-field local weak limits.

Several open directions recur across the literature. For symmetric speed questions, the higher-dimensional critical exponent suggested by the two-dimensional theory is \(\alpha_c=d-1\), but the exact threshold has not been established [1205.5449]. The same paper asks whether the ballistic and oscillatory counterexamples below the threshold can be made uniformly elliptic. In homogenization theory, the near-optimal moment relation \(1/p+1/q<2/(d-1)\) is known for everywhere sublinear correctors in \(d\ge 3\), but sharp quenched local limit theorems, parabolic Harnack inequalities, and Gaussian heat-kernel bounds under equally weak assumptions remain incomplete [1902.05793]. Correlated environments, dynamic weights, percolation-type zeros under mere ergodicity, and extensions beyond \(\mathbb{Z}^d\) all remain active problems [1602.08428]. The unifying lesson is that reversibility alone does not determine the macroscopic regime: the decisive inputs are the geometry induced by the conductances, the relative influence of upper and lower tails, and the analytic control available on correctors, kernels, and regeneration structures [2504.06690].

Source: https://www.emergentmind.com/topics/random-walk-among-random-conductances