Abstract
This research bridges algebraic topology and statistical mechanics by investigating the stochastic motion of cellular cycles within a finite CW complex. We introduce a homological observable termed the average current, which measures the flux of probability in the process. We prove that in the low-temperature, adiabatic limit—where the system is driven slowly with minimal noise—the average current undergoes fractional quantization. The denominators of these fractional values are shown to be combinatorial invariants of the underlying CW complex.
Introduction
Classical studies of stochastic processes often focus on random walks on graphs, which are one-dimensional structures with a finite number of states (vertices). However, many physical and biological systems involve higher-dimensional "cycles"—such as loops of string or membranes—evolving over time. A significant challenge in studying these systems is that a "random walk of cycles" on a higher-dimensional complex can have an infinite number of states, even if the complex itself is finite. This paper addresses this complexity by defining a rigorous framework for measuring the flux of these cycles and determining if they follow predictable, quantized patterns analogous to those found in 1D systems.
Methodology: Markov CW Chains and Dynamics
The authors define a Markov CW Chain where the states are $(k-1)$-cellular cycles. A transition occurs when a cycle "jumps" across a $k$-cell, effectively adding the boundary of that cell to the current state.
To analyze the evolution of probability distributions over these cycle states, the following tools are employed:
- Master Equations: A Fokker-Planck (or Kolmogorov) operator is used to describe the dynamics.
- Adiabatic and Low-Temperature Limits: The system is analyzed under conditions where the external driving protocol changes infinitely slowly ($\tau_D \to \infty$) and the noise level goes to zero ($\beta \to \infty$).
- Combinatorial Hodge Theory: Generalized versions of spanning trees and "co-trees" are used to decompose operators and solve the dynamical equations.
Key Insights and Results
The study reveals several fundamental properties of cycle fluctuations:
Expectation Dynamics (Theorem A)
While the full probability distribution of cycles is infinite-dimensional and difficult to track, the expected value (the first moment) follows a much simpler, finite-dimensional linear differential equation.
The Boltzmann Distribution as an Ergodic Limit (Theorem B)
In the slow (adiabatic) limit, the expected state of the system converges to a unique "harmonic form" known as the Boltzmann distribution. This acts as a bridge between the random process and the static topology of the space.
Fractional Quantization (Theorem C)
This is the central result of the paper. In 1D graphs, current typically quantizes to integers. In higher dimensions, the presence of torsion (topological twists in the space's homology) forces the average current $Q$ to take on fractional values. Specifically, $Q$ is a homology class that lies in a fractional lattice: $H_d(X; \mathbb{Z}[1/\delta])$.
Combinatorial Invariance
The denominator $\delta$ is not arbitrary; it is a product of orders of torsion subgroups and orders of relative homology groups associated with spanning trees and co-trees. For $d=1$, the theory reduces to known results for graphs where $\delta=1$ and the current is strictly integral.
Discussion and Future Directions
The main results rely on the driving protocol being "generic." If the protocol is "bad," the quantization might break down or become more complex. Additionally, while the expectation is well-behaved, the actual stochastic process can "blow up" (the cycle can grow infinitely large) in finite time under certain conditions.
Future work aims to extend these results to "robust" parameters and further explore "hypercurrents," which could have applications in studying topological insulators or complex biological networks.
Summary updated