Quantitative homogenization of degenerate random environments (Q1635958)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Quantitative homogenization of degenerate random environments
scientific article

    Statements

    Quantitative homogenization of degenerate random environments (English)
    0 references
    0 references
    0 references
    1 June 2018
    0 references
    This paper deals with the homogenization of discrete divergence-form operators. Let \(u_{t}\) be a solution of the problem \(\partial_{t} u_{t} + D^{*} a D u_{t}=0\), \(u_{0}=g\), where \(D^{*}aD u_{t}:=\sum_{0\in e}a(e)D_{e}u_{t}(a)\), where \(e\) is an edge in the set of unoriented edges \(\mathbf B\) of graph structure \({\mathbb Z}^{d}\). For a random variable \(\xi:\;\Omega \to \mathbb R\) and a fixed edge \(e\in{\mathbf B}\) it is defined \(D_{e}\xi(a):=\xi(\tau_{\bar e}a)-\xi(\tau_{\underline{e}}a)\), where for a given edge \(e\in {\mathbf B}\), the author writes \(\bar e\) and \(\underline{e}\) to denote its two endpoints, and \(\tau_{x} a:=\left(a(x+e)\right)_{e\in{\mathbf B}}\), where for \(e\in{\mathbf B}\), \(x+e:=(x+\underline{e},x+\bar e)\). The main result of the paper is following. Under the condition \(\left\langle \left(\sup_{1\leq i \leq d} a(e_{i})\right)^{-q}\right\rangle<\infty\) for every \(q\geq 1\), where \((e_1,\ldots, e_{d})\) is the canonical basis of \({\mathbb Z}^{d}\), the following statements hold: (a) For every \(p\in[1,\infty)\), there exists a constant \(C=C(d,p)\) such that, if \(g:\;\Omega\to {\mathbb R}\) is local with support of size \(N\), bounded and centered, then \(\left\langle | u_{t}| ^{2p}\right\rangle^{1/p}\leq C N^{2}\| g\|^{2}_{L^{\infty}(\Omega)} t^{-d/2}\), where \(\langle\cdot\rangle\) denotes the expectation with respect to the law of \(a(e)\). (b) For every \(p\in [1,\infty)\) and \(\varepsilon>0\), there exists a constant \(C=C(d,p,\varepsilon)\) such that, if \(f:\;\Omega\to{\mathbb R}\) is local with support of size \(N\) and bounded, and if \(g=D^{*} f\), where \(D^{*}\xi:=\sum_{i=1}^{d}D_{i}^{*}\xi_{i}(a)\), \(D_{i}^{*}\xi_{i}:=\xi(\tau_{-e_{i}}a)-\xi(a)\), then \(\left\langle | u_{t}| ^{2p}\right\rangle^{1/p}\leq C N^{2}\| f\|^{2}_{L^{\infty}(\Omega)} t^{-(d/2+1-\varepsilon)}\).
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    discrete linear divergence-form operator
    0 references
    random coefficients
    0 references
    random walk
    0 references
    relaxation of the environment
    0 references
    corrector estimate
    0 references
    0 references
    0 references
    0 references