Convolution structures associated with orthogonal polynomials (Q1206908)

The author studies a sequence of polynomials \(\{P_ n\}_ 0^{\infty}\) orthogonal w.r.t. a measure \(\mu\) on the real line and the convolution structure (following from the non-negativity of the coefficients \(c(n,m,k)\) in the formula \(P_ nP_ m=\sum c(n,m,k)P_ k\)) which makes \(\ell^ 1({\mathcal N})\) a Banach algebra. The first theorem asserts that the support of \(\mu\) is contained in the maximal ideal space \(\mathcal M\) of the algebra and the second theorem states that \({\mathcal M}=\{z:| P_ n(z) | \leq 1,n\geq 0\}=\{z:\sup_ n | P_ n(z) | < \infty\}= \hbox{supp} \mu\) when the coefficients in the three term recurrence relation (normalised to sum equal \(1\)) for the \(P_ n\) satisfy \(\alpha_ n \rightarrow {1\over 2},\;\beta_ n\rightarrow 0, \gamma_ n \rightarrow {1\over 2}\). The result can be applied to Jacobi polynomials \(P_ n^{(\alpha , \beta)}\) with \(\alpha\geq \beta,\;\alpha+\beta+1\geq 0\). Finally two results are given along the lines of a maximum principle: the maximal absolute value of orthogonal polynomials leading to a convolution structure is attained on the right end of the major axis under rather simple conditions on the coefficients \(\alpha_ n,\;\beta_ n, \gamma_ n\).