OpenFrontiers
Join
← Back to the index
OP-2992OpenProbability

Metastability a potential theoretic approach · 2006-prob-207

Markov processes · ICM 2006: Metastability a potential theoretic approach

MSC 60 · Probability theory and stochastic processes

Inferred from the community field; finer classification has not been reviewed.

Source assessment · not reviewedAs of 2026-08-18Source & evidence ↓
Read original source ↗ 0Submit progressLean 4 ↓Discussion (0)

Problem statement

Let S\mathcal S be a finite spin space. For finite sets Λ⊂Zd\Lambda\subset\mathbb Z^d, let XΛ=(XtΛ)t≥0X^{\Lambda}=(X_t^{\Lambda})_{t\geq0} be an ergodic continuous-time Markov process on ΩΛ=SΛ\Omega_\Lambda=\mathcal S^\Lambda that is reversible with respect to a Gibbs measure μΛ\mu_\Lambda, and let LΛL_\Lambda denote its nonnegative self-adjoint generator on L2(μΛ)L^2(\mu_\Lambda). Also allow an infinite-volume reversible spin dynamics X=(Xt)t≥0X=(X_t)_{t\geq0} on Ω=SZd\Omega=\mathcal S^{\mathbb Z^d}, with invariant Gibbs measure μ\mu and nonnegative self-adjoint generator LL. Write

E(f,f)=⟨f,Lf⟩L2(μ) \mathcal E(f,f)=\langle f,Lf\rangle_{L^2(\mu)}

for the Dirichlet form, with the analogous notation EΛ\mathcal E_\Lambda in finite volume. For disjoint measurable sets A,DA,D of configurations, define the equilibrium potential hA,Dh_{A,D} as the weak solution of

LhA,D=0on (A∪D)c,hA,D∣A=1,hA,D∣D=0, Lh_{A,D}=0\quad\text{on }(A\cup D)^c,\qquad h_{A,D}|_A=1,\qquad h_{A,D}|_D=0,

and define the capacity

cap⁡(A,D)=inf⁡{E(f,f):f∣A=1, f∣D=0}, \operatorname{cap}(A,D)=\inf\bigl\{\mathcal E(f,f):f|_A=1,\ f|_D=0\bigr\},

whenever these objects are well defined; use the corresponding definitions with LΛL_\Lambda, EΛ\mathcal E_\Lambda, and μΛ\mu_\Lambda in finite volume.

Construct a potential-theoretic theory of metastability for reversible spin-system dynamics in the limit ∣Λ∣→∞|\Lambda|\to\infty and in infinite volume, including parameter regimes not restricted to the low-temperature limit. In particular, formulate verifiable hypotheses under which one can obtain effective lower bounds on cap⁡(A,D)\operatorname{cap}(A,D), uniform or quantitatively controlled as the volume grows, for competing metastable sets AA and DD. Establish a priori local regularity estimates, uniform or quantitatively controlled in the same limit, for equilibrium potentials and for weak solutions of the associated Poisson and Dirichlet problems Lu=gLu=g, analogous to the controlled local Hölder estimates available for finite-dimensional reversible diffusions. The capacity bounds and regularity estimates must be strong enough to yield sharp asymptotic formulas, including the relevant prefactors when they exist, for metastable transition times and for the low-lying spectrum of the generator.

Lean 4

Not yet attached

No Lean statement or proof is attached yet. Have a formalization? Share its source to add it to this problem.

Share a formalization →Verification guide ↗
Remaining questions in this record

Remaining questions recorded by the source

These are retained within the parent record and are not counted as additional index entries.

Question 1

Let S\mathcal S be a finite spin space, let (ΛN)N≥1(\Lambda_N)_{N\geq1} be finite subsets of Zd\mathbb Z^d with ∣ΛN∣→∞|\Lambda_N|\to\infty, and let XN=(XtN)t≥0X^N=(X_t^N)_{t\geq0} be a reversible local spin-system dynamics on ΩN=SΛN\Omega_N=\mathcal S^{\Lambda_N}. For disjoint sets AN,BN⊂ΩNA_N,B_N\subset\Omega_N, write τBN=inf⁡{t≥0:XtN∈BN}\tau_{B_N}=\inf\{t\geq0:X_t^N\in B_N\}, and suppose that ANA_N is a metastable well whose internal relaxation time is negligible compared with its typical exit time to BNB_N. Determine model-independent, verifiable conditions on the dynamics and on AN,BNA_N,B_N under which the mean exit time is asymptotically independent of the microscopic initial configuration throughout the metastable well, in the sense that

sup⁡σ,η∈AN∣Eσ[τBN]Eη[τBN]−1∣⟶0as N→∞. \sup_{\sigma,\eta\in A_N}\left|\frac{\mathbb E_\sigma[\tau_{B_N}]}{\mathbb E_\eta[\tau_{B_N}]}-1\right|\longrightarrow0 \qquad\text{as }N\to\infty.

Develop regularity or coupling estimates that establish this conclusion for spatial spin systems without relying on a model-specific contracting drift.

Source reference: Anton Bovier, “Metastability: From Mean Field Models to SPDEs,” in Jean-Dominique Deuschel, Barbara Gentz, Wolfgang König, Max von Renesse, Michael Scheutzow, and Uwe Schmock (eds.), Probability in Complex Physical Systems: In Honour of Erwin Bolthausen and Jürgen Gärtner, Springer Proceedings in Mathematics 11, Springer, 2012, pp. 443–462, Section 7.1 (“Initial Distributions and Regularity Theory”), https://doi.org/10.1007/978-3-642-23811-6_18.

Question 2

Let XNX^N be a sequence of finite reversible Markov chains describing microscopic spin or particle dynamics, with finite state spaces ΩN\Omega_N, invariant measures μN\mu_N, jump rates qN(x,y)q_N(x,y), and edge conductances cN(x,y)=μN(x)qN(x,y)=cN(y,x)c_N(x,y)=\mu_N(x)q_N(x,y)=c_N(y,x). For disjoint metastable sets AN,DN⊂ΩNA_N,D_N\subset\Omega_N, define

cap⁡N(AN,DN)=inf⁡{12∑x,y∈ΩNcN(x,y)[f(x)−f(y)]2:f∣AN=1, f∣DN=0}. \operatorname{cap}_N(A_N,D_N)=\inf\left\{\frac12\sum_{x,y\in\Omega_N}c_N(x,y)[f(x)-f(y)]^2:f|_{A_N}=1,\ f|_{D_N}=0\right\}.

A loop-free unit flow ΦN\Phi_N from ANA_N to DND_N is a nonnegative function on oriented edges having total outflow one from ANA_N, total inflow one into DND_N, zero divergence outside AN∪DNA_N\cup D_N, and no directed cycles. Let PΦN\mathbb P^{\Phi_N} be the law of the path γ\gamma obtained by following the transition probabilities induced by ΦN\Phi_N, starting from its unit source distribution on ANA_N and stopping upon reaching DND_N. Construct, under general verifiable coarse-graining and free-energy hypotheses, a model-independent procedure for producing loop-free unit flows ΦN\Phi_N such that the Berman–Konsowa lower bound

EΦN ⁣[(∑(x,y)∈γΦN(x,y)cN(x,y))−1] \mathbb E^{\Phi_N}\!\left[\left(\sum_{(x,y)\in\gamma}\frac{\Phi_N(x,y)}{c_N(x,y)}\right)^{-1}\right]

is asymptotic to cap⁡N(AN,DN)\operatorname{cap}_N(A_N,D_N) as N→∞N\to\infty. In particular, the construction should yield sharp capacity lower bounds, including the leading prefactor, that match corresponding Dirichlet-principle upper bounds for microscopic spatial spin and particle systems without requiring a new model-specific flow construction in each case.

Source reference: Anton Bovier, “Metastability: From Mean Field Models to SPDEs,” in Jean-Dominique Deuschel, Barbara Gentz, Wolfgang König, Max von Renesse, Michael Scheutzow, and Uwe Schmock (eds.), Probability in Complex Physical Systems: In Honour of Erwin Bolthausen and Jürgen Gärtner, Springer Proceedings in Mathematics 11, Springer, 2012, pp. 443–462, Section 7.2 (“Canonical Constructions of Flows”), https://doi.org/10.1007/978-3-642-23811-6_18.

Progress & source assessment

Source-reported assessment from the ICM catalogue snapshot. This is not an independent review by Frontier Science.

Verdict: unresolved.

No citable work was found that supplies all three requested ingredients for reversible spatial spin systems in growing or infinite volume beyond the low-temperature limit: generally verifiable sharp capacity lower bounds, volume-controlled local regularity for equilibrium potentials and Poisson/Dirichlet solutions, and the resulting sharp transition-time and low-lying spectral asymptotics. The literature contains strong but model-specific results.

  • Anton Bovier, “Metastability: A Potential Theoretic Approach,” Proceedings of the International Congress of Mathematicians, Madrid 2006, Vol. III, European Mathematical Society, 2006, pp. 499–518, DOI 10.4171/022-3/26. Section 9 explicitly identifies capacity lower bounds and diffusion-like a priori regularity in high or infinite-dimensional spin systems as the two open obstacles.
  • Alessandra Bianchi, Anton Bovier, and Dmitry Ioffe, “Sharp Asymptotics for Metastability in the Random Field Curie-Weiss Model,” Electronic Journal of Probability 14 (2009), 1541–1603, DOI 10.1214/EJP.v14-673. Theorems 1.2–1.3 give sharp large-volume capacity and mean-exit-time asymptotics at fixed inverse temperature for the random-field Curie-Weiss model, including continuously distributed fields. This is genuine progress beyond exact finite-dimensional lumping, but it is a mean-field model and the microscopic flow and harmonic-function estimates are model-specific.
  • Alessandra Bianchi, Anton Bovier, and Dmitry Ioffe, “Pointwise Estimates and Exponential Laws in Metastable Systems via Coupling Methods,” Annals of Probability 40 (2012), 339–371, DOI 10.1214/10-AOP622. The paper proves near constancy of mean exit times within metastable sets and an exponential law for the random-field Curie-Weiss model through coupling. It supplies a regularity substitute in that model, not a volume-uniform Hölder-type theory for general spin systems or Poisson problems.
  • Anton Bovier, Frank den Hollander, and Cristian Spitoni, “Homogeneous Nucleation for Glauber and Kawasaki Dynamics in Large Volumes at Low Temperatures,” Annals of Probability 38 (2010), 661–713, DOI 10.1214/09-AOP492. The main theorems compute the nucleation time, including its prefactor, for two-dimensional Glauber and Kawasaki dynamics in boxes growing with β\beta. This addresses large volume but retains the low-temperature limit and special starting distributions, and it does not establish the requested general regularity or spectral theory.
  • Anton Bovier, “Metastability: From Mean Field Models to SPDEs,” in Probability in Complex Physical Systems, Springer Proceedings in Mathematics 11, Springer, 2012, pp. 443–462, DOI 10.1007/978-3-642-23811-6_18. Sections 7.1–7.2 report that generally applicable regularity tools and a generic construction of microscopic flows for sharp capacity lower bounds are still missing; the successful constructions rely heavily on the model.
  • Anton Bovier and Frank den Hollander, Metastability: A Potential-Theoretic Approach, Grundlehren der mathematischen Wissenschaften 351, Springer, 2015, DOI 10.1007/978-3-319-24777-9. The monograph develops the abstract finite-process theory and treats positive-temperature large-volume Curie-Weiss models, low-temperature lattice systems, and an Allen-Cahn SPDE. Its scope separates these special regimes: it does not derive a general positive-temperature or infinite-volume theory for spatial spin dynamics, and its large-volume lattice results explicitly retain restrictive initial laws and model-dependent recurrence estimates.
  • Sabine Jansen and Frank den Hollander, “Berman-Konsowa Principle for Reversible Markov Jump Processes,” Markov Processes and Related Fields 22 (2016), 409–442, arXiv:1309.1305. The paper extends the path-and-flow variational formula for capacity to reversible jump processes on Polish spaces. It provides a general lower-bound framework, but not a generic near-optimal flow construction or the volume-uniform estimates required by the problem.
  • Anton Bovier, Frank den Hollander, Saeda Marello, Elena Pulvirenti, and Martin Slowik, “Metastability of Glauber Dynamics with Inhomogeneous Coupling Disorder,” ALEA, Latin American Journal of Probability and Mathematical Statistics 21 (2024), 1249–1273, DOI 10.30757/ALEA.v21-48. Theorems 2.10 and 2.12 compare metastability and hitting times for a broad class of dense mean-field-like random-coupling systems at fixed temperature. Remark 2.13(c) still describes the desired near-constancy in the starting configuration as known only for certain spin systems and says that adapting the coupling is highly nontrivial and beyond the paper. The results concern hitting-time comparisons, not a general local regularity theory or low-lying spectral asymptotics in infinite volume.

Thus, the lower-bound and regularity mechanisms have been developed in important special models, but the requested unified theory, especially for genuinely spatial finite-range spin systems at fixed positive temperature and for infinite-volume generators, remains absent.

References
  1. [1]ICM Conjectures · 2006-prob-207 link
  2. [2]ICM 2006: Metastability a potential theoretic approach link

Recorded progress

Add yours →

Nothing recorded yet. A pointer to a paper the entry is missing counts.

Machine attempts

0 recorded

Attempts run against the statement below and are recorded in full, including the failures. The point is to stop the next model — and the next person — from re-walking a road that is already known to end. Each attempt costs 25 credits.

No machine attempts recorded.

Sign in to run an attempt. New accounts start with enough credits for ten.

Discussion

0 notes

Notes are for clarifying the statement, flagging a missing reference, or recording why an approach fails. Claimed solutions belong in a submission, where they can be reviewed.

No notes yet.

Sign in to add a note.