The weak limit of Ising models on locally tree-like graphs (Q664346)

The authors consider sequences of regular graphs \(G_n = (V_n, E_n)\) with increasing vertex sets \(V_n = [n] = \{1,\ldots, n\}\) that converge locally to trees, and they study the Ising model with inverse temperature \(\beta\) and without external field on sequences of graphs \(\{G_n\}_{n\in \mathbb{N}}\), where \(\{G_n\}_{n\in \mathbb{N}}\) is a sequence of regular graphs of degree \(k \geq 3\) with diverging girth. Local weak convergence for Ising measures is introduced. The authors examine whether the Ising measure converges to a probability measure over Gibbs measures. The main result is stated as follows: Let \(\{G_n\}_{n\in \mathbb{N}}\) be a sequence of \(k\)-regular graphs that converge locally to the tree \(\mathrm{T}_k\). For \((k - 1) \tanh \beta > 1\), define the sequences \(\{\mu_n\}_{n\in \mathbb{N}}\), \(\{\mu_{n,+}\}_{n\in \mathbb{N}}\) by \[ \begin{aligned} \mu_{n,+}(\underline{x}) =& \frac{1}{Z_{n,+}(\beta)}\exp\left\{\beta \sum_{(i,j)\in E_n} x_ix_j\right\} \mathbb{I} \left\{\beta \sum_{i \in V_n} x_i>0\right\},\\ \mu_{n}(\underline{x}) =& \frac{1}{Z_{n}(\beta)} \exp\left\{\beta \sum_{(i,j)\in E_n} x_ix_j \right\}.\end{aligned} \] Then (i) \(\mu_n\) converges locally in probability to \(\frac{1}{2} (\upsilon_+ +\upsilon_-)\) and (ii) if the graphs \(\{G_n\}\) are \((1/2, \lambda)\) edge expanders for some \(\lambda> 0\), then \(\mu_{n,+}\) converges locally in probability to the plus-boundary Gibbs measure on the infinite tree \(\nu_{+}\). They also prove a local characterization of the sequence of measures which corresponds to the phase transition phenomenon.