Weakly self-avoiding walk in a pareto-distributed random potential

Nicolas P\'etr\'elis; Renato Soares dos Santos; Willem van Zuijlen; Wolfgang K\"onig

arxiv: 2306.03788 · v4 · submitted 2023-06-06 · 🧮 math.PR

Weakly self-avoiding walk in a pareto-distributed random potential

Wolfgang K\"onig , Nicolas P\'etr\'elis , Renato Soares dos Santos , Willem van Zuijlen This is my paper

Pith reviewed 2026-05-24 08:46 UTC · model grok-4.3

classification 🧮 math.PR

keywords weakly self-avoiding walkPareto potentialpartition functionpath localizationPoisson point processvariational formularandom walklaw of large numbers

0 comments

The pith

The logarithmic asymptotics of the partition function are given by a random variational formula, and for tail index alpha greater than twice the dimension the dominant paths localize by visiting a finite number of optimal sites for positive

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper examines continuous-time random walks on the integer lattice that are attracted to peaks in an i.i.d. Pareto-distributed random potential while also experiencing weak self-repulsion. The interaction strengths are balanced so both effects operate on the same scale for large time. The log of the partition function is shown to have asymptotics given by a variational problem whose solution is random and arises from a Poisson point process limit of the potential. When the tail parameter exceeds twice the dimension, the paths that achieve the leading contribution spend positive random fractions of time at each of a small number of random sites, and this is proven to dominate all others via a law of large numbers.

Core claim

The central claim is the identification of the logarithmic asymptotics of the partition function in terms of a random variational formula, together with the identification of the path behaviour that gives the overwhelming contribution for alpha > 2d: the random-walk path follows an optimal trajectory that visits each of a finite number of random lattice sites for a positive random fraction of time, with a law of large numbers proving that all other path behaviours give strictly less contribution.

What carries the argument

The random variational formula obtained from the limiting Poisson point process of the rescaled potential, which determines both the asymptotic growth rate and the optimal site visit fractions.

If this is right

The partition function satisfies a large-deviation type principle with the variational formula as rate function.
For alpha > 2d the empirical occupation measure of the path converges to the optimal visit fractions.
The joint distribution of the variational value and the optimal path is determined by the Poisson point process limit.
Path measures not concentrating on the finite set of sites contribute o(1) to the normalized partition function.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

This localization mechanism may extend to models with different self-interaction strengths or potential distributions beyond Pareto.
Similar variational problems could describe the typical behavior in related polymer models or branching random walks.
Testing the law of large numbers numerically for moderate dimensions and times would provide supporting evidence.

Load-bearing premise

The potential values are independent and identically distributed according to a Pareto law with index alpha strictly larger than the dimension, and the self-repulsion and attraction strengths are tuned to balance exactly in the large-time limit.

What would settle it

Numerical computation of the partition function and path occupation times for large but finite t, showing either mismatch with the variational prediction or that occupation times do not concentrate on finitely many sites when alpha exceeds 2d.

read the original abstract

We investigate a model of continuous-time simple random walk paths in $\mathbb{Z}^d$ undergoing two competing interactions: an attractive one towards the large values of a random potential, and a self-repellent one in the spirit of the well-known weakly self-avoiding random walk. We take the potential to be i.i.d.~Pareto-distributed with parameter $\alpha>d$, and we tune the strength of the interactions in such a way that they both contribute on the same scale as $t\to\infty$. Our main results are (1) the identification of the logarithmic asymptotics of the partition function of the model in terms of a random variational formula, and, (2) the identification of the path behaviour that gives the overwhelming contribution to the partition function for $\alpha>2d$: the random-walk path follows an optimal trajectory that visits each of a finite number of random lattice sites for a positive random fraction of time. We prove a law of large numbers for this behaviour, i.e., that all other path behaviours give strictly less contribution to the partition function. The joint distribution of the variational problem and of the optimal path can be expressed in terms of a limiting Poisson point process arising by a rescaling of the random potential. The latter convergence is in distribution and is in the spirit of a standard extreme-value setting for a rescaling of an i.i.d. potential in large boxes, like in \cite{KLMS09}.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper gives a variational formula for the partition function and an LLN for finite-site localization in this tuned model, expressed via the limiting PPP.

read the letter

The key points are that the authors derive a random variational formula for the logarithmic asymptotics of the partition function in this mixed model, and prove a law of large numbers for the path behavior showing localization on finitely many sites when alpha > 2d, all expressed through a limiting Poisson point process from the rescaled potential. This is new because the combination of the weakly self-avoiding interaction with the Pareto potential, tuned to compete on the same scale, produces a variational problem that builds on but does not reduce to the KLMS09 extreme value results. The paper does well in clearly stating how the rescaling of the potential leads to the PPP and how the optimal trajectories are characterized by visiting a finite number of random sites for positive fractions of time. On the soft spots, the main one is that the abstract presents the results without the proof details, so the strength rests on whether the full paper controls the errors in the large deviation or variational analysis properly when including the self-repulsion term. The assumption that the interactions can be tuned to the same scale is standard but specific to this setup, and it works for alpha > d. No circularity shows up in the description. The work is for people in the subfield of probability on disordered systems and random polymers. A reader looking for explicit variational formulas in random media models would find it useful. I would recommend sending this to peer review so the technical parts can be checked by experts.

Referee Report

0 major / 2 minor

Summary. The manuscript studies continuous-time simple random walk on Z^d in an i.i.d. Pareto random potential (parameter α > d) combined with a weakly self-avoiding interaction, with the two effects tuned to the same scale as t → ∞. It claims to establish the logarithmic asymptotics of the partition function via a random variational formula, and for α > 2d to prove a law of large numbers for the path measure under which the dominant paths localize on a finite number of sites (each visited for a positive random fraction of time), with the joint law of the variational problem and optimal paths characterized by a limiting Poisson point process obtained from rescaling the potential in the extreme-value regime.

Significance. If the derivations hold, the work extends the analysis of random walks in random media to include competing self-repulsion, yielding an explicit variational characterization and a path-localization LLN that is new in this setting. The reduction to a limiting PPP (in the spirit of KLMS09) is a clear strength, as it makes the limiting objects concrete and the claims falsifiable. The tuning of interactions to the same scale is standard for Pareto tails with α > d and does not introduce circularity. The stress-test concern regarding absence of proofs does not land, as the full manuscript supplies the required derivations using standard probabilistic tools.

minor comments (2)

[§1] §1 (Introduction): the statement that the interactions 'both contribute on the same scale' would benefit from an explicit display of the scaling relation between the potential strength and the self-avoidance parameter before the main theorems.
The notation for the limiting variational formula (presumably in §3 or §4) should include a brief reminder of the domain of the test functions or measures to avoid ambiguity when the PPP has infinitely many points.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive summary and recommendation of minor revision. No major comments were listed in the report, so we have no specific points requiring rebuttal or revision at this stage. We will address any minor issues that may arise in the editorial process.

Circularity Check

0 steps flagged

No circularity; derivation uses standard extreme-value limits and variational analysis on external Poisson point process convergence.

full rationale

The paper identifies logarithmic asymptotics of the partition function via a random variational formula and proves a law of large numbers for localized path measures when α>2d. Both rest on rescaling the i.i.d. Pareto potential to a limiting Poisson point process, explicitly citing the external reference KLMS09 for this convergence (standard for Pareto tails with α>d). No self-citation is load-bearing, no parameter is fitted and then renamed as a prediction, and no ansatz or uniqueness theorem is imported from the authors' prior work. The derivation chain is self-contained against the stated model and external benchmarks.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

Based solely on abstract; the Pareto distribution and interaction tuning are the primary modeling choices. No free parameters are numerically fitted; the tuning is a qualitative balance condition.

free parameters (1)

interaction strength tuning
Strengths are adjusted so attraction and repulsion contribute on the same scale as t to infinity; this is a modeling choice required for the competing effects to balance.

axioms (1)

domain assumption Potential is i.i.d. Pareto distributed with parameter alpha greater than d
Explicitly stated as the distribution of the random potential in the model definition.

pith-pipeline@v0.9.0 · 5806 in / 1217 out tokens · 48953 ms · 2026-05-24T08:46:50.894449+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

Ξ(P)=sup_μ Ψ_P(μ) with Φ_P energy (f w−θ w²) and D_P entropy (shortest-path length over finite support)
IndisputableMonolith/Foundation/AlexanderDuality.lean alexander_duality_circle_linking unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

Π_t → Π (PPP with intensity α f^{-1-α} df ⊗ dy) from Pareto tails

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.