Inequalities for the Hadamard weighted geometric mean of positive kernel operators on Banach function spaces (Q865020)

From MaRDI portal





scientific article; zbMATH DE number 5125400
Language Label Description Also known as
default for all languages
No label defined
    English
    Inequalities for the Hadamard weighted geometric mean of positive kernel operators on Banach function spaces
    scientific article; zbMATH DE number 5125400

      Statements

      Inequalities for the Hadamard weighted geometric mean of positive kernel operators on Banach function spaces (English)
      0 references
      0 references
      0 references
      0 references
      13 February 2007
      0 references
      Let \(u\) be a \(\sigma\)-finite positive measure on a \(\sigma\)-algebra of subsets of \(X\) and let \(M(u)\) denote the vector space of equivalence classes of almost everywhere equal complex measurable functions on \(X\). A Banach function space \(L\) is a Banach space consisting of elements of \(M(u)\) with the property that \(f\in L\), \(g\in M(u)\) and \(|g|\leq|f|\) imply that \(g\in L\) and \(\| g\|\leq\| f\|\). \(L\) is said to have the Fatou property if for every increasing net \(\{f_\alpha\}_\alpha\) with limit \(f\) in \(L\), we have \(\| f\|= \sup_{\alpha\| f_\alpha\|}\). A bounded linear operator \(K\) on such an \(L\) is called a kernel operator if \[ (Kf)(x) = \int_X k(x, y)f(y)\,du(y) \] for some measurable function \(k(x, y)\) on \(X\times X\) such that \[ \int_X|k(x, y)f (y)|\,du(y)<\infty \] for all \(f\) in \(L\) and for almost all \(x\) in \(X\). For kernel operators \(K_j\), \(1\leq j\leq n\), on \(L\) with kernels \(k_j\geq 0\) a.e.\ and positive numbers \(\alpha_j\), \(1\leq j\leq n\), satisfying \(\sum_j\alpha_j= 1\), the Hadamard weighted geometric mean \(K\) is defined as the kernel operator with kernel equal to \(k_1(x, y)^{\alpha_1}\cdots k_n(x, y)^{\alpha_n}\). The authors prove that in this case \(\| K\|\leq\| K_1\|^{\alpha_1}\cdots\| K_n\|^{\alpha_n}\), \(r(K)\leq r(K_1)^{\alpha_1}\cdots r(K_n)^{\alpha_n}\) and, for \(L= L^2(u)\), \(w(K)\leq w(K_1)^{\alpha_1}\cdots w(K_n)^{\alpha_n}\). Here, for an operator \(A\) on a Hilbert space \(H\) with inner product \(\langle\cdot,\cdot\rangle\), \(\| A\|\), \(r(A)= \sup\{|\lambda|: \lambda\in\sigma(A)\}\) and \(w(A)=\sup\{|\langle Ax,x\rangle|: x\in H\), \(\| x\|= 1\}\) denote its operator norm, spectral radius and numerical radius, respectively. Another result proved is an infinite-dimensional generalization to positive kernel operators on Banach function spaces over atomic measure spaces of a characterization of Elsner, Hershkowitz and Pinkus of functions \(f:\mathbb{R}^n_+\to \mathbb{R}_+\) satisfying \(r(f(A_1,\dots, A_n))\leq f(r(A_1),\dots, r(A_n))\) for all nonnegative matrices \(A_1,\dots, A_n\) of the same size.
      0 references
      Banach function spaces
      0 references
      kernel operators
      0 references
      operator norm
      0 references
      spectral radius
      0 references
      numerical radius
      0 references
      operator inequalities
      0 references
      Hadamard product
      0 references
      Schur product
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references