Non-compact integral operators with semi-separable kernels and their discrete analogues: Inversion and Fredholm properties (Q800598)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Non-compact integral operators with semi-separable kernels and their discrete analogues: Inversion and Fredholm properties
scientific article

    Statements

    Non-compact integral operators with semi-separable kernels and their discrete analogues: Inversion and Fredholm properties (English)
    0 references
    0 references
    0 references
    0 references
    0 references
    1984
    0 references
    The authors solve and give the Fredholm theory for the integral equations \[ (T\phi)(t)= \phi(t)-\int^{\infty}_{0} k(t,s)\phi(s)ds=f(t),\;0\leq t<\infty, \] and \[ \phi(t)+ \int^{\infty}_{-\infty} k(t,s)\phi(s)ds=f(t),\;-\infty <t<\infty. \] The kernel \(k(t,s)\) has the form \[ k(t,s)=C(t)U(t)(I-P)U(s)^{-1}B(s),\;s<t,\;k(t,s)=- C(t)U(t)PU(s)^{-1}B(s),\;s>t, \] where P is a projection of \({\mathbb{C}}^ n\), \(B(t)\) and \(C(t)\) are measurable, essentially bounded matrix functions of sizes \(m\times n\) and \(n\times m\) respectively, U(t) is the fundamental matrix (normalized to I at 0) of the differential equation \(\dot x(t)=A(t)x(t),\) where \(A(t)\) is a locally integrable \(n\times n\) matrix function. The proposed theory is a generalization of the theory of convolution integral equation with rational symbol (which was developed by the first two authors and H. Bart). The cases of a full line and of a half-line are considered respectively in first and second chapters. The discrete analogues of those cases are considered in the third chapter. The main results are formulated in the terms of the associate fundamental matrix \(U^ x(t)\) (the fundamental matrix of the equation \(\dot U^ x(t)=A^ x(t)U^ x(t)\), where \(A^ x(t)=A(t)-B(t)C(t))\) and of the exponential dichotomy (the projection I-P is an exponential dichotomy for the fundamental matrix U(t) if there exist positive constants L, M, \(\alpha\) and \(\beta\) such that \[ \| U(t)(I-P)U(s)^{-1}\|\leq Le^{-\alpha (t-s)},\;0\leq s<t<\infty, \] \[ \| U(t)PU(s)^{- 1}\|\leq Me^{-\beta (t-s)},\;0\leq t<s<\infty). \] For example: Let I-P be an exponential dichotomy for \(U(t)\). The operator \(T: L^ m_ p[0,\infty)\to L^ m_ p[0,\infty)\), \(p=1\) or \(p=\infty\), is invertible if and only if the associate fundamental matrix \(U^ x(t)\) has an exponential dichotomy \(I-P^ x\) and \({\mathbb{C}}^ n=\ker P^ x\oplus Im P\). In that case the solution of \(T_{\phi}=f\) is given by \[ \phi (t)=f(t)-\int^{\infty}_{0}\gamma (t,s)f(s)ds,\;0\leq t<\infty \] where \[ \gamma(t,s)= \begin{cases} C(t)U^ x(t)(I-\pi)U^ x(s)^{-1}B(s),\;& 0\leq s<t<\infty,\\ -C(t)U^ x(t)\pi U^ x(s)^{-1}B(s),\;& 0\leq t<s<\infty, \end{cases} \] and \(\pi\) is the projection of \({\mathbb{C}}^ n\) along ker\(P^ x\) onto Im P.
    0 references
    0 references
    half-line
    0 references
    full-line
    0 references
    non-compact kernels
    0 references
    semi-separable type
    0 references
    concept of dichotomy
    0 references
    invertibility
    0 references
    resolvent kernels
    0 references
    Fredholm characteristics
    0 references
    exponential dichotomy
    0 references