Random Schrödinger operators and convolution on wreath products
Adam Arras · 2025
Read Random Schrödinger operators and convolution on wreath products by Adam Arras (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-22485.txt.
Introduction
Adam Arras links two objects that seem unrelated. One is the random Schrödinger operator, which models quantum particles in disordered media. The other is a convolution operator on a wreath product, a group built by stacking copies of one group along another. A classic case ties lamplighter groups to such operators with coin-flip disorder, and Arras extends that bridge in both directions.
Crossing the bridge one way gives a simple test for when convolutions on wreath products spread smoothly. Crossing back gives estimates for how wave energy dies out in disordered systems on graphs that grow in a polynomial way. He also derives an exact formula for a key Green function moment using a wreath product with a simple lamp group. The work serves pure mathematics on groups and quantum models alike.
Catalog record
| CREATOR | Adam Arras |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 39 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-22485
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Random Schrödinger operators and convolution on wreath products Adam Arras Abstract We establish a spectral correspondence between random Schrödinger operators and deterministic convolution operators on wreath products, generalizing previous results that relate Lamplighter groups to Schrödinger operators with Bernoulli potentials. Using this correspondence in both directions, we obtain an elementary criterion for the absolute continuity of convolutions on wreath product, Lifschitz tail estimates for Schrödinger operators on Cayley graph of polynomial growth, and an exact formula for the second moment of the Green function, expressed in terms of the wreath product with an Abelian group of lamps. 1 Introduction and main results The central theme of this paper is the spectral connection between deterministic convolution operators on wreath product groups and random Schrödinger operators. We will mostly be concerned with Hilbert spaces of the form \mathcal{H}=\ell^{2}(\Gamma) , \Gamma being a countable, discrete group with identity e_{\Gamma} . Given x\in\Gamma , the delta function \delta_{x} is defined by \delta_{x}(y)=1 if y=x , and \delta_{x}(y)=0 otherwise. The spectral representation [ 33 , Theorem VIII.6] of a self-adjoint operator will be denoted L=\int_{\Sigma(L)}\lambda E_{L}(\mathrm{d}\lambda),\qquad\mu_{L}^{\varphi,\psi}(\cdot)=\langle\varphi,E_{L}(\cdot)\psi\rangle. We shorten the notation of the spectral measures by setting \mu_{L}^{x}=\mu_{L}^{\delta_{x},\delta_{x}} . Convolution operators. The left regular representation (L_{x}f)(y)=f(x^{-1}y) , for any square summable function f\in\ell^{2}(\Gamma) , defines an unitary action of the group \Gamma . To any complex finite measure m\in\mathcal{M}_{\mathbb{C}}(\Gamma)\simeq\ell^{1}(\Gamma) , we consider the sum L_{m}:=\sum_{x}m(x)L_{x} . Any operators of this form will be referred as a convolution operators . When m is a probability measure, this correspond to the Markov operator associated with the simple random walk, where increments are i.i.d. and follow the distribution m . We refers to the book [ 14 ] for a modern introduction to group convolution. A convolution operator is self-adjoint if and only if the associated measure satisfies m(g^{-1})=\bar{m}(g) . In this case, the spectral measure \mu_{L_{m}}^{x} exists and is independent of x\in\Gamma . This measure is commonly known as the Plancherel measure [ 6 ] , or the Kesten-von Neumann-Serre spectral measure [ 20 ] . A case of particular of interest is when the measure takes value zero and one, corresponding to the adjacency operator of a Cayley graph. The spectral analysis of this algebraic structure, initiated by Kesten [ 25 ] , has led to some surprising recent developments [ 19 ] ; see also [ 6 , 18 , 9 , 31 , 29 , 11 ] . Random Schrödinger operator. In his seminal work [ 4 ] , the physicist P. W. Anderson studied the transport properties of disordered crystals. This quantum analogue of random walks in a random environment is modeled as follows. The Cayley graph \operatorname{Cay}(\Gamma,S) of an infinite, finitely generated group describes the hopping terms of a particle on the crystal lattice. The presence of small-scale impurities is represented by an independent and identically distributed scalar field (v(x,\omega))_{x\in\Gamma} . To avoid technicalities, we only consider distribution with bounded support. The system evolves according to the unitary group generated by the Random Schrödinger operator H(\omega)=A+V of form Laplacian plus potential 1 1 1 Instead of A=L_{1_{S}} where 1_{S}=\sum_{S}\delta_{s} , the true Laplacian of a Cayley graph \operatorname{Cay}(\Gamma,S) is the convolution operator \Delta=|S|I-A . , which acts on \varphi\in\ell^{2}(\Gamma) via (H(\omega)\varphi)(x)=\sum_{y\sim x}\varphi(y)+v(x,\omega)\varphi(x). (1) Since the distribution of the operator is invariant by the action of \Gamma on \ell^{2}(\Gamma) , it follows from ergodicity theory that global features of H(\omega) are almost sure. For example, the spectrum verifies \Sigma(H(\omega))\overset{a.s.}{=}\Sigma_{H} , for some closed set \Sigma_{H} , called the almost sure spectrum. The same applies to any subset in the Lebesgue decomposition of the spectral measures \mu_{H(\omega)}^{x} into absolutely continuous, singular continuous and pure point part. We refer to [ 3 , Theorem 3.10] for more details. In the one dimensional case \Gamma=\mathbb{Z} , it turns out that arbitrarily small randomness completely breaks down the transport properties of the medium. This translates into pure point spectrum and exponential decay of the eigenfunctions for the operator H(\omega) . This phenomenon, that also occurs for any group \Gamma for high enough disorder, is referred as Anderson’s localization [ 35 ] . It is by now reasonably understood from a mathematical viewpoint, since the work of Fröhlich and Spencer (via multi-scale analysis [ 16 ] ) and Aizenman and Molchanov (via fractionnal moment method [ 1 ] , see also [ 17 ] ). In contrary, the conjectured persistence at low disorder of transport in the case of \Gamma=\mathbb{Z}^{d},d\geq 3 , and therefore the stability of an absolutely continuous spectrum, remains a challenging open question despite considerable efforts of the community. Some results are known in the case where the potential decays at infinity [ 8 , 7 ] , but the presence of absolutely continuous spectrum for a stationary potential has only been proven for trees. This has been achieved on the d -regular tree \mathbb{T}_{d} by Klein [ 26 ] , and generalized by different approaches since then [ 15 , 2 , 24 , 5 ] . Wreath product of group. The wreath product \Lambda\wr\Gamma of two discrete groups is defined as the semidirect product (\bigoplus_{\Gamma}\Lambda)\rtimes\Gamma . The factor \bigoplus_{\Gamma}\Lambda denotes the direct sum, the subgroup of the direct product \prod_{\Gamma}\Lambda that consists of elements with finite support. The factor \Gamma acts on \bigoplus_{\Gamma}\Lambda by shifting indices via the regular action. Elements of \Lambda\wr\Gamma can thus be represented as pairs (f,g) , where f:\Gamma\to\Lambda has finite support 2 2 2 In our context, the support corresponds to the set f^{-1}(\Lambda\setminus\{e_{\Lambda}\}) . and g\in\Gamma . The group operations is given by (f,g)\cdot(f^{\prime},g^{\prime})=(f\cdot(f^{\prime}\circ g^{-1}),gg^{\prime}). Here, f^{\prime}\circ g^{-1} denotes the shifted function h\mapsto f^{\prime}(g^{-1}h) , reflecting the semidirect action. This algebraic structure has an intuitive geometric interpretation as automorphisms on a two-level rooted tree \mathcal{T}_{\Gamma,\Lambda} constructed as follows. The root o is connected to each element g of \Gamma and each g is attached to a copy of \Lambda , denoted gl,l\in\Lambda . {\lx@inpgf@ignorespaces o} {\lx@inpgf@ignorespaces g_{1}=e_{\Gamma}} {\lx@inpgf@ignorespaces g_{2}} {\lx@inpgf@ignorespaces\ldots} {\lx@inpgf@ignorespaces g_{n}} {\lx@inpgf@ignorespaces g_{1}l_{1}\ldots} {\lx@inpgf@ignorespaces g_{1}l_{p}} {\lx@inpgf@ignorespaces g_{2}l_{1}\ldots} {\lx@inpgf@ignorespaces g_{2}l_{p}} {\lx@inpgf@ignorespaces g_{n}l_{1}\ldots} {\lx@inpgf@ignorespaces g_{n}l_{p}} Figure 1: The tree \mathcal{T}_{\Gamma,\Lambda} An automorphism is a permutation of the vertices that preserves the adjacency relation. In particular, it must fix the root. Given an element h=(f,g)\in\Lambda\wr\Gamma , one associates the automorphism \varphi_{h} sending s\in\Gamma to \varphi_{h}(s)=gs and sl\in\Gamma\times\Lambda to \varphi_{h}(sl)=gsf(g)l . An important example is the wreath product (\mathbb{Z}/2\mathbb{Z})\wr\mathbb{Z} , often referred to as the lamplighter group . It can be visualized as follows: consider a lamplighter located on an infinite street indexed by the integer. Along the street, there are lamps at each integer that can be either on or off. The lamplighter moves along the line, switching the state of the lamps as they go. The group \mathbb{Z} corresponds to the position of the lamplighter, while the group \bigoplus_{\mathbb{Z}}\mathbb{Z}/2\mathbb{Z} tracks the configuration of the lamps. With this interpretation, we refer to \Gamma as the base group and \Lambda as the lamp group. A lamp configuration corresponds to a function f\colon\Gamma\to\Lambda with finite support. Both the base group and the lamp group embed into the wreath product via the maps \Gamma\ni g\hookrightarrow\hat{g}=(e,g),\qquad\Lambda\ni l\hookrightarrow\hat{l}=(1_{l},e_{\Gamma}), (2) where e=e_{{\underset{\Gamma}{\oplus}\Lambda}} and 1_{l}(g)=e_{\Lambda} , if g\not=e_{\Gamma} and 1_{l}(e_{\Gamma})=l otherwise. We extend these embeddings to measures