A kinetically constrained model exhibiting non-linear diffusion and jamming

Antonello Scardicchio · 2024

Read A kinetically constrained model exhibiting non-linear diffusion and jamming by Antonello Scardicchio (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-05231.txt.

Introduction

Abhishek Raj and co-authors study a 2024 model of particles that can move only when their neighbors allow it. The particles sit on a triangular ladder, and the rules create a mix of spreading motion and stuck, jammed states that the team can also read as a system of interacting fermions.

In that reading, the rate of spread equals the inverse of the mass of the moving excitations, which the authors estimate with a mean-field method. At a set density, the system flips: a huge number of layouts jam for good while others keep flowing. The authors note the same setup can extend to flat two-dimensional grids. The work links traffic-like jamming in simple rules to tools from quantum theory.

Catalog record

CREATORAntonello Scardicchio
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE38 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-05231

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A kinetically constrained model exhibiting non-linear diffusion and jamming Antonello Scardicchio Abstract We present a classical kinetically constrained model of interacting particles on a triangular ladder, which displays diffusion and jamming and can be treated by means of a classical-quantum mapping. Interpreted as a theory of interacting fermions, the diffusion coefficient is the inverse of the effective mass of the quasiparticles which can be computed using mean-field theory. At a critical density ρ = 2 / 3 \rho=2/3 , the model undergoes a dynamical phase transition in which exponentially many configurations become jammed while others remain diffusive. The model can be generalized to two dimensions. 1 Introduction In recent years, we have witnessed a shift of attention in the field of statistical physics from the study of thermodynamic, equilibrium properties, which have been the core of the subject since its inception, to the characterization of the approach to equilibrium, or lack thereof. This is due to theoretical and experimental advances, both in quantum and classical systems. In the first realm, the current degree of control on cold atomic arrays [ 16 ] , superconducting qubits arrays [ 22 , 15 ] , and the larger availability of supercomputing resources has spurred a renewed interest in the dynamics of many-body quantum systems and in particular in their approach to equilibrium. The eigenstate thermalization hypothesis [ 9 , 44 , 6 ] , and its counterpart for strongly disordered systems, many-body localization [ 2 , 32 , 1 , 31 , 42 ] are two prominent examples of this work. In the classical world, the line of work which started with the study of equilibrium properties of spin glasses [ 12 , 26 ] turned out to apply well beyond the original scope [ 21 ] , and evolved into a theory of complexity [ 34 ] . The mean-field theory, which comprises replica symmetry breaking, has been successfully applied to computational complexity [ 27 , 28 ] , structural glasses [ 25 ] , and many other examples [ 35 ] . In this paper we turn our attention to a classical random process model which has slow dynamics and jamming, and whose solution is obtained by mapping to a model of interacting fermions, thereby connecting the two lines of work. Figure 1: Comparison of mean-field solution (dashed straight-line) against extrapolated numerical results (green diamonds) – see Figs. 7 and 8 ; Notice that the numerical data follow the mean-field prediction of Eq.( 28 ) quite closely until the jamming point \rho=2/3 . The disagreement at high density may be due to progressively less accurate extrapolation of numerics. To achieve this, we consider a model of interacting random walkers with simple constraints, or a kynetically constrained process (KCP). KCP’s are a widely used tool in statistical mechanics to describe the approach to equilibrium, the emergence of non-equilibrium states, and glassy dynamics [ 8 , 39 , 4 ] . Among other aspects, they are useful to study the emergence, in a controlled setting, of the hydrodynamic behavior and the statistics of large deviations from equilibrium. Such family of models include those introduced and treated by Katz, Lebowitz and Spohn [ 20 ] and many others [ 14 , 43 ] . The typical strategy to find the diffusion coefficient in these models is to use the Green-Kubo formula and thermodynamics to compute it, or if this is not possible, bound [ 24 , 23 ] . This is not the route we will follow in this paper. In this paper, we introduce and study a simple exclusion process in which particles can hop to their left or right conditioned to the fact that both the left and right side is empty. The model is easily defined on a triangular ladder , which will allow, in further work, to define it on a two-dimensional, triangular lattice . We employ a classical-to-quantum mapping to turn our KCP into a frustration-free [ 17 ] spin Hamiltonian whose groundstates may be obtained exactly [ 17 ] . To demonstrate the existence of a diffusive behavior in the original model we study the dispersion of elementary excitations of the quantum problem – the diffusion coefficient is proportional to the inverse mass of the quasiparticle. By construction, the diffusion coefficient is expected to decrease monotonically with increasing particle density \rho . We find the diffusion coefficient with a mean-field treatment of the fermionic Hamiltonian and compare against exact numerical results on finite chains, see Fig. 1 . The mean-field solution D=3(1-\rho) , where \rho is the density of particles, agrees surprisingly well with the numerics in a very large range of densities, see Fig. 1 , which makes us suspect it is an exact solution. This is more surprising if we consider that the model exhibits the presence of jammed configurations, which appear at a critical density \rho=2/3 , therefore showing an interesting connection with the study of jamming in granular materials [ 10 , 35 , 3 ] . 2 Preliminaries – model, mapping, definitions 2.1 The model and a classical-quantum mapping The random process studied describes random updates of 3 particles configurations where the only mobile configurations are those in which a particle is free on both sides. The rationale behind this choice [ 38 ] is to have a simple explicit model of random walkers with strongly dependent diffusion constant that produces strong nonlinear interactions among hydrodynamic modes. The rule is sketched in Figure 2 . It is also straightforward to implement in large scale simulations – see e.g. Ref. [ 38 ] . This work, by contrast, focuses on exact noise-averaged formulation of the dynamics as we explain now. Figure 2: The rules for hopping. In active triangles, where only one particle out of 3 is present, it can hop with the same probability along the green arrows. Where two or more particles are present in a triangle (red cross) such triangle is inactive. On the lower axis, the numeration convention we use. In order to write the rate matrix we choose the basis 000,001,...,111 , where 0 is an empty site and 1 a particle. On these 3 neighboring lattice sites, the matrix of the transition rates is h_{i,i+1,i+2}=\left(\begin{array}[]{cccccccc}0&0&0&0&0&0&0&0\\ 0&-1&\frac{1}{2}&0&\frac{1}{2}&0&0&0\\ 0&\frac{1}{2}&-1&0&\frac{1}{2}&0&0&0\\ 0&0&0&0&0&0&0&0\\ 0&\frac{1}{2}&\frac{1}{2}&0&-1&0&0&0\\ 0&0&0&0&0&0&0&0\\ 0&0&0&0&0&0&0&0\\ 0&0&0&0&0&0&0&0\\ \end{array}\right) (1) The total rate matrix is H=p\sum_{i=1}^{L}h_{i,i+1,i+2} (2) where we use periodic boundary conditions, and it is a positive matrix, whose steady state (later, quantum ground state energy) corresponds to the zero eigenvalues, E_{0}=0 . Master equation governs the evolution of probabilities of particle configurations \sigma=\{\sigma_{1},...,\sigma_{L}\}\in\{0,1\}^{L} \dot{P}(\sigma,t)=-\sum_{\sigma,\sigma^{\prime}}H_{\sigma,\sigma^{\prime}}P(\sigma^{\prime},t). (3) Time-dependent expectation values of observables O can be computed as \displaystyle\left\langle O(t)\right\rangle=\sum_{\sigma}O(\sigma)P(\sigma,t) \displaystyle= \displaystyle\sum_{\sigma,\sigma^{\prime}}O(\sigma)\left(e^{-Ht}\right)_{\sigma,\sigma^{\prime}}P(\sigma^{\prime},0), (4) \displaystyle= \displaystyle\sum_{a}e^{-E_{a}t}\sum_{\sigma,\sigma^{\prime}}O(\sigma)\left\langle\sigma\middle|E_{a}\right\rangle\left\langle E_{a}\middle|P(0)\right\rangle, (5) where \ket{E_{a}} are the eigenvectors of H . Note: while Eq.( 4 ) is a general solution of Eq.( 3 ), Eq.( 5 ) only follows for problems obeying detailed balance, i.e. the rate matrix is symmetric and therefore admits left-right-symmetric eigen-decomposition. Finally, we note the value of the diffusion constant of a single random walker on this lattice in discrete time is D_{0}=(p/2)(2\cdot(2a)^{2}+4\cdot a^{2})/(2\Delta t) , as each site belongs to 3 triangles with 6 possible hops. In what follows we set p=a=\Delta t=1 , for simplicity

Read the complete text