Potential theory for Schrödinger operators on finite networks (Q2493574)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Potential theory for Schrödinger operators on finite networks
scientific article

    Statements

    Potential theory for Schrödinger operators on finite networks (English)
    0 references
    0 references
    0 references
    0 references
    0 references
    21 June 2006
    0 references
    The authors of this interesting and carefully written paper consider discrete Schrödinger operators \(L_q(u)(x)=L(u)(x)+q(x)u(x)\) on finite networks, where \(L(u)\) is a combinatorial Laplacian \[ L(u)(x)=\int_V c(x,y)(u(x)-u(y))\,dy. \] A finite network is a finite simple graph with the vertex set \(V\), the integral is, of course, a sum, and the function (conductance) \(c(x,y)=c(y,x)\), \(c(x,y)\geq 0\), and \(c(x,y)=0\) unless the vertices \(x\) and \(y\) are adjacent. For any function \(\sigma (x)\neq 0, x\in V\), the authors define a potential \(q_{\sigma}=-\frac{1}{\sigma}L(\sigma)\) and prove that in turn \(q_{\sigma}\) determines \(\sigma\) up to a multiplicative constant. This allows the authors to develop a comprehensive theory of operators \(L_q\) with potentials \(q\geq q_{\sigma}\); since nontrivial potentials \(q_{\sigma}\) must change sign, this condition is more general than \(q\geq 0\). The authors study the selfadjointness and bilinear forms associated with \(L_q\) and develop the corresponding potential theory (monotonicity and comparison properties, capacity, Green and Poisson kernels). Moreover, they prove a version of the kernel theorem in this setting and derive an integral (i.e. sum) representation of \(L_q\).
    0 references
    0 references
    discrete Schrödinger operator
    0 references
    finite networks
    0 references