Orthogonal polynomials and Laurent polynomials related to the Hahn-Exton \(q\)-Bessel function (Q1909583)

The (associated) Lommel polynomials arise when iterating the relation between Bessel functions of the first kind \(J_{\nu+ 1}\), \(J_{\nu}\), \(J_{\nu-1}\): \[ J_{\nu+ m} (z)= h_{m, \nu} \bigl( {\textstyle {1\over z}} \bigr) J_\nu (z)- h_{m-1, \nu+1} \bigl( {\textstyle {1\over z}} \bigr) J_{\nu-1} (z). \] They satisfy the three term recurrence relation \[ h_{m+1, \nu} (z)= 2z (m+ \nu) h_{m, \nu} (z)- h_{m-1, \nu} (z), \quad h_{-1, \nu} (z)=0, \quad h_{0, \nu} (z) =1, \] and are orthogonal with respect to the discrete measure with weight \(1/( j_k^{\nu-1} )^2\) in \(1/j_k^{\nu -1}\), with \(j_k^{\nu- 1}\) \((\nu> -1\), \(k= 1, 2, \dots\) the positive zeros of \(J_{\nu -1}\). The authors consider the well known Hahn-Exton \(q\)-Bessel function, given by \[ J_\nu (x; q)= {{(q^{\nu+ 1}; q)_\infty} \over {(q; q)_\infty}} x^\nu {}_1 \varphi_1 {0 \choose {q^{\nu+ 1};\;q, qx^2}} \] and the Laurent \(q\)-Lommel polynomials generated by the three therm recurrence relation (for the standard notations in \(q\)-theory concerning Pochhammer symbols and hypergeometric functions the reader is referred to the basic literature, to the paper under review or to \textit{H. T. Koelink} and \textit{R. F. Swarttouw} [J. Math. Anal. Appl. 186, No. 3, 690-710 (1994; Zbl 0811.33013)]. The polynomials satisfy: \[ h_{m+1, \nu} (x;q)= \bigl( {\textstyle {1\over x}}+ x(1-q^{\nu+m}) \bigr) h_{m,\nu} (x;q)- h_{m-1,\nu} (x; q), \;\;h_{-1, \nu} (x; q)= 0, \;\;h_{0, \nu} (x; q)=1. \] A second (independent) solution is given by \(h_{m-1, \nu+1} (x;q)\). It would go too far to give all subsequent formulae and proofs in detail here, the reader of this review has to glean his interest in the paper from the following highlights of subjects treated: 1. Explicit orthogonality relations for the Laurent \(q\)-Lommel polynomials; 2. Facts about the functional \({\mathcal L}\) arising (i.e. a functional on the space of Laurent polynomials such that \({\mathcal L} (x^n)\) exists for all \(n\in \mathbb{Z}\)); 3. The positive definite moment functionals \({\mathcal L}_\pm\) defined by \({\mathcal L}_+ (x^n)= {\mathcal L} (x^n)\), \({\mathcal L}_- (x^n)=- {\mathcal L} (x^{-2- n})\), \(n\in \mathbb{Z}_+\); 4. The support of \({\mathcal L}_+\) as one over the zeros of Hahn-Exton \(q\)-Bessel functions together with zero, where the mass at zero is strictly positive; 5. The orthogonal polynomials for \({\mathcal L}_-\) in terms of Al-Salam--Chihara polynomials; 6. Estimates using perturbation techniques for \(h_{m, \nu} (x; q)\), \(J_\nu (x; q)\) and the associated function \(j_\nu (x; q)\) [the recurrence relation and initial values of \(h_{m, \nu} (x; q)\); for \(q=0\), \(\nu\to \infty\) have a solution \(U_m (x+ x^{-1} )/2\), Chebyshev of the second kind, showing that indeed the \(q\)-Lommel polynomial can be seen as a perturbation of \(U_m\)]; 7. An existence theorem for the strong moment functional \({\mathcal L}\). This excellently written paper is the decisive work on the generalization of Lommel polynomials arising from the Hahn-Exton \(q\)-Bessel function.