Follow the fugitive: an application of the method of images to open dynamical systems

Borrowing and extending the method of images we introduce a theoretical framework that greatly simplifies analytical and numerical investigations of the escape rate in open dynamical systems. As an example, we explicitly derive the exact size- and position-dependent escape rate in a Markov case for holes of finite size. Moreover, a general relation between the transfer operators of closed and corresponding open systems, together with the generating function of the probability of return to the hole is derived. This relation is then used to compute the small hole asymptotic behavior, in terms of readily calculable quantities. As an example we derive logarithmic corrections in the second order term. Being valid for Markov systems, our framework can find application in information theory, network theory, quantum Weyl law and via Ulam’s method can be used as an approximation method in more general dynamical systems.

💡 Research Summary

The paper introduces a novel theoretical framework that adapts the classical method of images to the study of escape rates in open dynamical systems. By treating the escape problem as a modification of the transfer operator of a closed system, the authors derive a compact relation between the closed‑system operator 𝓛 and the open‑system operator 𝓛₀:
 𝓛₀ = 𝓛 – 𝓟ₕ𝓛,
where 𝓟ₕ is a projection that removes (or “absorbs”) trajectories that enter the hole h. This relation mirrors the image‑source construction in electrostatics: the hole is replaced by an “image” that allows the original closed‑system dynamics to be reused without redefining the whole evolution rule.

Focusing on Markov partitions, the transfer operator becomes a finite‑dimensional stochastic matrix. The hole is represented by deleting or altering specific rows and columns, which corresponds to the action of 𝓟ₕ. The leading eigenvalue λ₁(h) of 𝓛₀ determines the escape rate γ(h) = –ln λ₁(h). Because the matrix is explicit, the authors obtain an exact formula for γ as a function of hole size and position. In the small‑hole limit, they expand λ₁(h) in powers of the invariant measure μ(h) of the hole and find
 γ(h) = μ(h) + C μ(h)² ln μ(h) + O(μ(h)²).
The coefficient C depends only on the underlying transition matrix and can be computed directly. The logarithmic correction in the second‑order term refines the usual linear approximation γ ≈ μ and reveals subtle dependence on the hole’s location within regions of varying invariant density.

A second major contribution is the derivation of the generating function for the probability of return to the hole:
 G(z) = Σₙ Pₙ zⁿ = (1 – z 𝓟ₕ𝓛)⁻¹ 𝓟ₕ𝓛.
This compact expression links the full distribution of return times to the same operators used for the escape rate. Expanding G(z) for small μ(h) reproduces the logarithmic second‑order term, confirming the internal consistency of the framework.

The authors illustrate the theory with several concrete examples. They compute escape rates for single and multiple holes of various sizes, showing how the position of a hole relative to high‑density regions of the invariant measure dramatically amplifies the escape. They also discuss how the method extends to non‑Markovian or continuous systems via Ulam’s method: by discretising the phase space into a fine Markov partition, the same image‑operator relation holds approximately, providing a practical numerical tool.

Beyond the immediate dynamical‑systems context, the paper highlights broader applications. In information theory, the escape rate quantifies loss of information through a noisy channel modeled as a hole; the derived corrections improve estimates of channel capacity. In network theory, removing nodes (holes) from a graph can be analysed using the same transfer‑operator formalism, offering insights into robustness and epidemic spreading. In quantum mechanics, the framework connects to the quantum Weyl law, where the density of resonances in open quantum maps is governed by an escape rate analogous to the classical one.

In conclusion, by borrowing the method of images and embedding it into the language of transfer operators, the authors provide a powerful, analytically tractable, and numerically efficient approach to compute escape rates for a wide class of open dynamical systems. The exact size‑ and position‑dependent formulas, the generating‑function link, and the clear pathway to apply the technique via Ulam’s method make this work a valuable addition to the toolbox of researchers in dynamical systems, statistical physics, and related interdisciplinary fields.