Loewner's differential equation and spidernets (Q2279037)

In the paper, the Loewner differential equation is regarded from a quantum probabilistic point of view. The author finds discrete approximations via comb products of spidernets for the Loewner equation \[\frac{\partial}{\partial t}g_t(z)=\frac{1}{g_t(z)-U(t)},\;\;g_0(z)=z\in\mathbb C^+:=\{z\in\mathbb C:\text{Im}\,z>0\},\;t\geq0,\] with a continuous non-negative driving function \(U\). Also, the author considers the general Loewner equation \[\frac{\partial}{\partial t}f_t(z)=-\frac{\partial}{\partial z}f_t(z)\int_{\mathbb R}\frac{\nu_t(du)}{z-u},\;\;f_0(z)=z\in\mathbb C^+,\] where \((\nu_t)_{t\geq0}\) is a family of probability measures on \(\mathbb R\), \(\text{supp}\,\nu_t\subset[0,M]\) for all \(t\geq0\). Let \((f_t)_{t\geq0}\) be the solution to the Loewner equation. Denote by \((\mu_t)_{t\geq0}\) the probability measures with \(F_{\mu_t}=f_t\) where the \(F\)-transform \(F_{\mu}\) of \(\mu\) is the mapping \[F_\mu:\mathbb C^+\to\mathbb C^+,\;\;F_{\mu}(z):=\left(\int_{\mathbb R}\frac{\mu(du)}{z-u}\right)^{-1}.\] Let \((X_t)_{t\geq0}\) be the corresponding self-adjoint operator-valued additive monotone increment process. Denote by \(\lfloor x\rfloor\) the largest \(m\in\mathbb N_0\) with \(m\leq x\). The author approximates \((X_t)_{t\in[0,T]}\) by graphs. Theorem 5.3. Choose \(n_0\in\mathbb N\) such that \(M\leq\sqrt{T/2}(2\sqrt n-1/\sqrt{n^3})\) for all \(n\geq n_0\). There exists a family \((\mathcal D_{n,k})_{n\geq n_0,k=1,\dots,n}\) of loop-free and unweighted rooted graphs such that: \newline (a) For all \(n\geq n_0\), \((\mathcal D_{n,k})_{k=1,\dots,n}\) can be considered as graphs with common vertex set \(V_n\) and common root \(o_n\). Let \(A_{n,k}\) be the adjacency matrix of \(\mathcal D_{n,k}\). Then the increments \((A_{n,1},A_{n,2}-A_{n,1},\dots,A_{n,n}-A_{n,n-1})\) are monotonically independent with respect to the quantum probability space \((l^2(V_n),\delta_{o_n})\). \newline (b) Denote by \(\mu_{n,k}\) the distribution of \(A_{n,k}\). Then \[\lim_{n\to\infty}\mu_{n,\lfloor tn/T\rfloor}(\sqrt{2n^3/T})=\mu_t(\cdot)\] with respect to weak convergence for all \(t\in[0,T]\). The limit also holds true with respect to the convergence of all moments.