Ratios of the Gauss hypergeometric functions with parameters shifted by integers: more on integral representations (Q2062072)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Ratios of the Gauss hypergeometric functions with parameters shifted by integers: more on integral representations |
scientific article |
Statements
Ratios of the Gauss hypergeometric functions with parameters shifted by integers: more on integral representations (English)
0 references
22 December 2021
0 references
The paper is concerned with the study of the ratio of two Gauss hypergeometric functions, in which the parameters of the numerator function differ from the respective parameters of the denominator function by integers. The authors derive explicit integral representations for this ratio based on a formula for its imaginary part. The new representations are illustrated with a few examples and an application to products of proportions. The authors recently published a paper [Ratios of the Gauss hypergeometric functions with parameters shifted by integers. I, Preprint, \url{arXiv:2103.13312}] in which they study the ratios of two Gauss hypergeometric functions \[ R_{n_1,n_2,m}(z)=\frac{{2}_{}F_{1}(a+n_1,b+n_2;c+m;z)}{{2}_{}F_{1}(a,b;c;z)}, \] where \(n_1,n_2,m\) are arbitrary integer. The particular case \(R_{0,1,1}\) was derived by Gauss who found a continued fraction representation for this ratio which, under additional restrictions on parameters, becomes a Stieltjes or \(S\)-fraction convergent to a Stieltjes transform of a positive measure. In [loc. cit.], the authors extended this integral representation to the general \(R_{n_1,n_2,m}(z)\) under the assumptions that \(a,b,c\in\mathbb{R}~; c,c+m\neq0,-1,-2,-3,...\) and \( R_{n_1,n_2,m}(z)\) has no poles in a complex plan \(\mathbb{C}\backslash [1,+\infty)\). Also, the authors assumed that the behavior of \(R_{n_1,n_2,m}(z)\) near \(z=1\) is mild, so that the singularity at this point is integrable. The main purpose of the present paper is to drop these restrictions and extend the representations given in [loc. cit.] to a more general setting. Firstly, the authors will get rid of the assumption of integrability near the point \(z =1\) and allow arbitrary behaviour in the neighbourhood of this point. Secondly, the authors will remove the assumption that \(R_{n_1,n_2,m}(z)\) is free of poles in \(\mathbb{C}\backslash [1,+\infty)\). However, the representation for ratios with poles obtained by authors will depend on rational functions whose numerators and denominators have explicit degrees but unknown coefficients. Calculation of these coefficients requires the knowledge of the zeros of the Gauss hypergeometric function and residues of \(R_{n_1,n_2,m}(z)\) at these zeros (for generic values of its parameters). These, of course, cannot be given explicitly, but may, in principle, be computed numerically.
0 references
Gauss hypergeometric function
0 references
Integral representation
0 references
Runckel's theorem
0 references
0 references