Mark Ashbaugh (with Rafael D. Benguria), Properties of Bessel Function Zeros Related to Isoperimetric Inequalities for Eigenvalues of Membranes and Plates

Abstract:
This talk will center on properties of zeros of Bessel functions that are related to various isoperimetric inequalities for the low characteristic frequencies of homogeneous membranes and plates. These correspond to optimal inequalities for the eigenvalues of the Laplacian and biharmonic operators under suitable boundary conditions. For example, by means of a simple identity we recover Lorch's result that $j_{p,1}^2-(p+1)(p+5)$ increases from $0$ to infinity as $p$ runs from $-1$ to infinity (here $j_{p,1}$ denotes the first positive zero of the Bessel function $J_p(x)$). Several other inequalities for Bessel function zeros which are suggested by variational theory or conjectured eigenvalue inequalities will also be discussed.

Richard Askey (University of Wisconsin) "Bessel functions and how to use them when considering more general classes of functions"

Abstract:
From the time of Sturm's work on comparison theorems, one role Bessel functions have played is to suggest results for more general classes of functions. Some examples will be given, both of cases where more general results have been found and ones where we are still looking to see how to extend known results about Bessel functions.

James A. Donaldson (Howard University), ``The Linear Shallow Water Theory and Dirichlet-Neumann Operators''

Abstract
The ``Shallow Water'' Theory provides an example of the approximation of the solution of an initial-boundary value problem for an elliptic partial differential equation by the solution of an initial-value problem for a hyperbolic partial differential equation. Formal and rigorous derivations of the this theory are given and a mathematical justification for the theory is obtained.

In obtaining a mathematical justification for the shallow water theory for the case where the bottom of the region of the water is horizontal, the Fourier transformation plays a primary role; for the case where the bottom of the region of the water is non-horizontal, a generalized Fourier transformation plays a key role in the two-dimensional theory, and spectral theory of self-adjoint operators plays a similar role in the three-dimensional theory. We show that Dirichlet-Neumann operators appear naturally in the treatment of these cases, and discuss some of their properties.

This is joint work with Daniel A. Williams.

Jean-Pierre Kahane (Universite de Paris - Orsay)
"Summability, order and products of Dirichlet series".
Abstract:
This talk will be mainly historical, but also contains a few new results on some questions left open by Stieltjes, Landau and H.Bohr. In particular, results and problems on the Lindeloef function for ordinary Dirichlet series will be considered.

Amram Meir (York University), Degree distribution in random trees

 

Abstract:
\documentstyle[12pt]{article} \pagestyle{empty} \hoffset=-2cm \textwidth=17cm \topmargin=-1cm \textheight=24cm \begin{document} \begin{center} {\large Degree distribution in random trees} \\ \normalsize \bigskip Amram Meir\\ \bigskip \end{center} Let ${\cal F}$ be a simply generated family of random rooted trees characterized by a functional relation $y = x\varphi(y)$, where $\varphi(t) = \sum^\infty_0 c_k t^k$ satisfies certain analyticity conditions. Let $N_m(T_n)$ denote the (weighted) number of nodes of out-degree $m$ in $T_n \in {\cal F}$. It was established in an earlier paper by Meir and Moon that $$ E\{ N_m(T_n) : T_n \in {\cal F}\} \sim \frac{c_m\tau^m}{\varphi(\tau)} n , $$ where $\tau$ is the solution of $\tau\varphi'(\tau) = \varphi(\tau)$. V.T. S\'os and I raised the question whether the proportion of nodes of out-degree $m$ {\em on any level} $k$ is asymptotically the same as in the tree itself. This is indeed the case for all $k = {\cal O}(n^{1/2})$ and $m = o(k)$. This covers almost all of the relevant cases. \end{document}

A. McD. Mercer (University of Guelph), On "deleting a row and column" from a differential operator.

Abstract:
If we delete a row and its corresponding column from a real symmetric matrix then the eigenvalues of the resulting matrix interlace those of the original one. There is an analogous result for a certain class of differential operators. Furthermore, in this more general context, large families of complete orthogonal systems of functions make their appearance.

Mark Pinsky (Northwestern University), "Pointwise convergence of the Fourier integral and related orthogonal expansions in several variables"

Abstract:
We review our recent work [AMS Notices, vol 42, March 1995, pp. 330-334.] on the convergence of Fourier expansions at a pre-assigned point, for piecewise smooth functions on Euclidean space. The neccessary and sufficient condition for convergence of the spherical partial sums is expressed in terms of the smoothness of the radial averages of the function w.r.t. the pre-assigned point. Extensions to Jacobi polynomial expansions and Hermite polynomial expansions in several variables are formulated and proved.

P. G. Rooney (University of Toronto), "On the Hankel transformation".

Abstract:

   
\documentstyle[12pt]{article}
\pagestyle{empty}
\hoffset=-2cm
\textwidth=17cm
\topmargin=-1cm
\textheight=24cm
\begin{document}
\begin{center}
{\large
On the Hankel transformation
}\\
\normalsize
\bigskip
P. G. Rooney\\
\bigskip
\end{center}

The Hankel transformation of order $\nu > -1$, $H_\nu$ is defined for
suitable functions $f$ by
$$(H_\nu f)(x) = \int_0^\infty (xt)^\frac12 J_\nu(xt)f(t)dt.$$

Its boundedness and range will be reviewed on the spaces ${\cal
L}_{\mu,p}$, defined for $\mu \in {\bf R},\; 1~<~p~<~\infty$, by the norm
$$ \parallel f \parallel_{\mu,p} \;= \left\{ \int_0^\infty |x^\mu f(x)|
\frac{dx}{x} \right\}^{\frac{1}{p}}.$$
It will be seen that the transforms of functions in suitable ${\cal
L}_{\mu,p}$ spaces satisfy Lipschitz and integral Lipschitz conditions. 
This is joint work with P. Heywood.

 \end{document}

Cora Sadosky (Howard University), "Restricted BMO in product spaces".

Abstract:
The class BMOr, "restricted BMO," appears naturally in several problems of harmonic analysis in product spaces, as those of interpolation and approximation by analytic functions in the polydisk, or of boundedness of the product Hilbert transform in weighted Lebesgue spaces. First introduced to characterize the symbols of bounded Hankel operators acting in the multidimensional torus, BMOr provides necessary and sufficient conditions for solutions to polydisk versions of Pick-Nevanlinna and Caratheodory- Fejer problems. The characterization of weights in terms of logarithmic BMOr provides new insights into function theory in product spaces. This is joint work with Mischa Cotlar.

Walter Van Assche, Zeros of orthogonal polynomials and eigenvalues of matrices

%Lee Lorch meeting, June 9 -10, 1995
% plain TeX 
%
\magnification=1200
\centerline{\bf Zeros of orthogonal polynomials and eigenvalues of matrices}
\centerline{by Walter Van Assche\footnote{$^*$}{Senior Research Associate
of the Belgian National Fund for Scientific Research}}
\centerline{Katholieke Universiteit Leuven}
\bigskip

Orthonormal polynomials on the real line always satisfy a three-term recurrence relation $$ x p_n(x) = a_{n+1}p_{n+1}(x) + b_n p_n(x) + a_n p_{n-1}(x), $$ where $a_n > 0$ and $b_n$ is real. This can be written as $$ \left( \matrix{ b_0 & a_1 & & & \cr a_1 & b_1 & a_2 & & \cr & a_2 & b_2 & \ddots & \cr & & \ddots & \ddots & \cr & & & a_{n-1} & b_{n-1}} \right) \left( \matrix{p_0(x) \cr p_1(x) \cr p_2(x) \cr \vdots \cr p_{n-1}(x)} \right) = x \left( \matrix{p_0(x) \cr p_1(x) \cr p_2(x) \cr \vdots \cr p_{n-1}(x)} \right) - a_n p_n(x) \left( \matrix{0 \cr 0 \cr \vdots \cr 0 \cr 1} \right) , $$

which shows that a zero $x_{j,n}$ of $p_n$ is an eigenvalue of the $n\times n$ Jacobi matrix containing the recurrence coefficients, with eigenvector $\left( p_0(x_{j,n}) , p_1(x_{j,n}) , \cdots \, p_{n-1}(x_{j,n}) \right)$. A similar connection exists between zeros of orthogonal polynomials on the unit circle and certain Hessenberg matrices which are almost unitary. Moreover this connection between zeros and eigenvalues can also be extended to matrix valued polynomials.

We will show how one can obtain properties of zeros of orthogonal polynomials by means of results from linear algebra. In particular interlacing properties, Gauss quadrature (on the real line and on the unit citcle), and monotonicity properties of the zeros as a function of the parameters. By considering the infinite Jacobi matrix as an operator acting on $\ell_2$, we will also show how results from operator theory give interesting results for the corresponding orthogonal polynomials. \bye

revised May 28, 1995