Matrix solutions of the functional equation of the gamma function (Q579547)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Matrix solutions of the functional equation of the gamma function
scientific article

    Statements

    Matrix solutions of the functional equation of the gamma function (English)
    0 references
    0 references
    0 references
    0 references
    1987
    0 references
    The general solution of the equation \(f(\lambda +1)=\lambda f(\lambda)\) is \(f(\lambda)=h(\lambda)\Gamma (\lambda),\) where \(\Gamma\) is the gamma function and h is an arbitrary function of period 1. The principal result in this paper is the determination of the matrix-valued functions f of the matrix variable A which satisfy the functional equations (1) \(f(A+1)=Af(A)=f(A)A\). The first half of the paper is a survey giving the needed background. In particular, it discusses Giorgi's definition of a matrix-valued function associated with a (sufficiently differentiable) scalar valued function f. If A has the canonical form \(P^{-1}AP=J_ 1\oplus J_ 2\oplus...\oplus J_ k\), then f(A) is given by \((2)\quad f(A)=P(f(J_ 1)\oplus...\oplus f(J_ k))P^{-1};\) if the Jordan block \(J_ i=\lambda_ iI+B_ i\) has size m(i), then \((3)\quad f(J_ i)=\sum (1/j!)f^{(j)}(\lambda_ i)B^ j_ i,\) the sum taken for \(0\leq j\leq m(i)-1\). The authors give an extension of Giorgi's definition; instead of a single function f and its derivatives, they have a finite sequence of functions \((f_ 0,...,f_{n-1})\) and \(f_ j(\lambda_ i)\) replaces \(f^{(j)}(\lambda_ i)\) in (3). With this machinery in place, the authors solve (1) for the case where \(A=J=\lambda I+B\) is a Jordan block of size m. This is equivalent to the solution of the system of difference equations \(f_ 0(\lambda +1)-\lambda f_ 0(\lambda)=0,\) \(f_ j(\lambda +1)-\lambda f_ j(\lambda)=f_{j-1}(\lambda),\) \(j=1,...,m-1\), and then \(f(J)=\sum f_ j(\lambda)B^ j,\) the sum taken for \(0\leq j\leq m(i)-1\). Finally, the solution of (1) for general A is obtained using (2).
    0 references
    0 references
    0 references
    0 references
    0 references
    system of difference equations
    0 references
    gamma function
    0 references
    matrix-valued functions
    0 references