Math 2.0

Statistical mechanics, structure of matter Rigorous equilibrium theory Theorem

Dobrushin–Lanford–Ruelle equations

Builds on Gibbs measureProbability measureConditional expectationWeak convergence of measures

Statement

Let SS be a finite set and Ω=Sℤd\Omega = S^{\Z^d} with the product topology and Borel σ\sigma-algebra. Let Φ=(ΦA)\Phi = (\Phi_A) be an interaction: AA runs over finite subsets of ℤd\Z^d, and ΦA ⁣:Ω→ℝ\Phi_A \colon \Omega \to \R depends only on the spins in AA. Assume sup⁡x∈ℤd∑A∋x∥ΦA∥∞<∞\sup_{x \in \Z^d} \sum_{A \ni x} \norm{\Phi_A}_\infty < \infty, and let β>0\beta > 0. For finite Λ⊂ℤd\Lambda \subset \Z^d and ω∈Ω\omega \in \Omega, let σΛωΛc\sigma_\Lambda \omega_{\Lambda^c} be the configuration equal to σ\sigma on Λ\Lambda and to ω\omega off Λ\Lambda. Put HΛω(σ)=∑A∩Λ≠∅ΦA(σΛωΛc)H_\Lambda^\omega(\sigma) = \sum_{A \cap \Lambda \ne \emptyset} \Phi_A(\sigma_\Lambda \omega_{\Lambda^c}) and γΛ(σΛ∣ω)=e−βHΛω(σ)∑η∈SΛe−βHΛω(η).\gamma_\Lambda(\sigma_\Lambda \mid \omega) = \frac{e^{-\beta H_\Lambda^\omega(\sigma)}}{\sum_{\eta \in S^\Lambda} e^{-\beta H_\Lambda^\omega(\eta)}}. A probability measure μ\mu on Ω\Omega satisfies the DLR equations if, for every finite Λ\Lambda and every σΛ∈SΛ\sigma_\Lambda \in S^\Lambda, μ(σΛ∣ℱΛc)(ω)=γΛ(σΛ∣ω)\mu(\sigma_\Lambda \mid \mathcal{F}_{\Lambda^c})(\omega) = \gamma_\Lambda(\sigma_\Lambda \mid \omega) for μ\mu-almost every ω\omega. Here ℱΛc\mathcal{F}_{\Lambda^c} is the σ\sigma-algebra generated by the spins off Λ\Lambda, and the left side is a conditional probability. Let 𝒢\mathcal{G} be the set of these measures, the Gibbs measures for βΦ\beta \Phi.

  1. 𝒢\mathcal{G} is nonempty, convex and compact for the weak topology.
  2. For each ω∈Ω\omega \in \Omega, every weak limit point of γΛn(⋅∣ω)\gamma_{\Lambda_n}(\cdot \mid \omega), along cubes Λn↑ℤd\Lambda_n \uparrow \Z^d, lies in 𝒢\mathcal{G}.
  3. The extreme points of 𝒢\mathcal{G} are exactly the measures in 𝒢\mathcal{G} that are trivial on the tail σ\sigma-algebra.

Remarks

Proof idea: summability makes ω↦γΛ(σΛ∣ω)\omega \mapsto \gamma_\Lambda(\sigma_\Lambda \mid \omega) continuous, which lets the DLR equations pass to weak limits; compactness of Ω\Omega then gives limit points. The kernels are consistent: γΔγΛ=γΔ\gamma_\Delta \gamma_\Lambda = \gamma_\Delta for Λ⊆Δ\Lambda \subseteq \Delta, and this is why finite-volume measures with boundary conditions converge to solutions. Example: the Ising model on ℤ2\Z^2 with h=0h = 0 and large β\beta has two different limits, with all-plus and all-minus boundary conditions, so 𝒢\mathcal{G} has more than one element. In dimension one with a translation-invariant finite-range interaction, 𝒢\mathcal{G} has exactly one element for every β\beta.

Going deeper

For a student

The DLR approach, due to Dobrushin and to Lanford and Ruelle in the late 1960s, separates local rules from global behaviour. The local rules always have solutions, but there can be several, and their number is the mathematical meaning of phase coexistence. Extreme Gibbs measures are the pure phases, and every Gibbs measure is a unique mixture of them. The Dobrushin uniqueness criterion gives a condition for a unique solution, and the Gibbs variational principle characterizes the translation-invariant ones.

Reading path

To understand this page from the ground up, read these 5 pages in order. The foundations come first.

Show the path as a graph
Each page uses the pages above it. The colour of the bar on a box shows how strong the evidence is, and the box also gives it as text. The list has the same pages.
  • no evidence level
  1. Conditional expectation concept
  2. Gibbs measure concept
  3. Probability measure concept
  4. Weak convergence of measures concept
  5. Dobrushin–Lanford–Ruelle equations statement

Further reading

  1. Web https://en.wikipedia.org/wiki/Gibbs_measure
About this page
TagETPFR89C87
URIm2://ETPFR89C87
Kindstatement, theorem
WritingW0 stub
MSC82B05, 82B20, 60K35
StatusThis copy of the site has no build data. Run m2 status ETPFR89C87.
AboutGibbs measure
UsesGibbs measure, Probability measure, Conditional expectation, Weak convergence of measures
Changed3 November 2026

Markdown · JSON · Edit on GitHub

What this page needs for W1

  • an inventory row: the row of the inventory that the node accounts for (0 of 1)
  • a source with a locator (0 of 1)

The writing levels say how complete the text of a page is. They do not measure its truth.

What links here

No page links to this page yet.