Fast computation of the Gauss hypergeometric function with all its parameters complex with application to the Pöschl-Teller-Ginocchio potential wave functions

Abstract The fast computation of the Gauss hypergeometric function F 1 2 with all its parameters complex is a difficult task. Although the F 1 2 function verifies numerous analytical properties involving power series expansions whose implementation is apparently immediate, their use is thwarted by instabilities induced by cancellations between very large terms. Furthermore, small areas of the complex plane, in the vicinity of z = e ± i π 3 , are inaccessible using F 1 2 power series linear transformations. In order to solve these problems, a generalization of R.C. Forrey's transformation theory has been developed. The latter has been successful in treating the F 1 2 function with real parameters. As in real case transformation theory, the large canceling terms occurring in F 1 2 analytical formulas are rigorously dealt with, but by way of a new method, directly applicable to the complex plane. Taylor series expansions are employed to enter complex areas outside the domain of validity of power series analytical formulas. The proposed algorithm, however, becomes unstable in general when | a | , | b | , | c | are moderate or large. As a physical application, the calculation of the wave functions of the analytical Poschl–Teller–Ginocchio potential involving F 1 2 evaluations is considered. Program summary Program title: hyp_2F1, PTG_wf Catalogue identifier: AEAE_v1_0 Program summary URL: http://cpc.cs.qub.ac.uk/summaries/AEAE_v1_0.html Program obtainable from: CPC Program Library, Queen's University, Belfast, N. Ireland Licensing provisions: Standard CPC licence, http://cpc.cs.qub.ac.uk/licence/licence.html No. of lines in distributed program, including test data, etc.: 6839 No. of bytes in distributed program, including test data, etc.: 63 334 Distribution format: tar.gz Programming language: C++, Fortran 90 Computer: Intel i686 Operating system: Linux, Windows Word size: 64 bits Classification: 4.7 Nature of problem: The Gauss hypergeometric function F 1 2 , with all its parameters complex, is uniquely calculated in the frame of transformation theory with power series summations, thus providing a very fast algorithm. The evaluation of the wave functions of the analytical Poschl–Teller–Ginocchio potential is treated as a physical application. Solution method: The Gauss hypergeometric function F 1 2 verifies linear transformation formulas allowing consideration of arguments of a small modulus which then can be handled by a power series. They, however, give rise to indeterminate or numerically unstable cases, when b − a and c − a − b are equal or close to integers. They are properly dealt with through analytical manipulations of the Lanczos expression providing the Gamma function. The remaining zones of the complex plane uncovered by transformation formulas are dealt with Taylor expansions of the F 1 2 function around complex points where linear transformations can be employed. The Poschl–Teller–Ginocchio potential wave functions are calculated directly with F 1 2 evaluations. Restrictions: The algorithm provides full numerical precision in almost all cases for | a | , | b | , and | c | of the order of one or smaller, but starts to be less precise or unstable when they increase, especially through a, b, and c imaginary parts. While it is possible to run the code for moderate or large | a | , | b | , and | c | and obtain satisfactory results for some specified values, the code is very likely to be unstable in this regime. Unusual features: Two different codes, one for the hypergeometric function and one for the Poschl–Teller–Ginocchio potential wave functions, are provided in C++ and Fortran 90 versions. Running time: 20,000 F 1 2 function evaluations take an average of one second.

[1]  V. Pierro,et al.  Computation of hyperngeometric functions for gravitationally radiating binary stars , 2002 .

[2]  S. Dong,et al.  Exact solutions of the s-wave Schrödinger equation with Manning–Rosen potential , 2007 .

[3]  N. Michel Precise Coulomb wave functions for a wide range of complex l, eta and z , 2007, Comput. Phys. Commun..

[4]  Adrian A. Husain Ark , 2010 .

[5]  J. Ginocchio A class of exactly solvable potentials. I. One-dimensional Schrödinger equation☆ , 1984 .

[6]  V. Villalba,et al.  Scattering of a Klein-Gordon particle by a Woods-Saxon potential , 2005, hep-th/0503108.

[7]  Cooper,et al.  Relationship between supersymmetry and solvable potentials. , 1987, Physical review. D, Particles and fields.

[8]  R. L. Robinson,et al.  COULOMB EXCITATION OF ⁷⁵As. , 1967 .

[9]  B. M. Fulk MATH , 1992 .

[10]  Y. Sucu,et al.  Exact solution of Dirac equation in 2+1 dimensional gravity , 2007 .

[11]  Philip M. Morse,et al.  On the Vibrations of Polyatomic Molecules , 1932 .

[12]  A. Bohr,et al.  Study of Nuclear Structure by Electromagnetic Excitation with Accelerated Ions , 1956 .

[13]  G. G. Stokes "J." , 1890, The New Yale Book of Quotations.

[14]  M. Stoitsov,et al.  Analytically solvable mean-field potential for stable and exotic nuclei , 1997, nucl-th/9703045.

[15]  Nico M. Temme,et al.  Numerically satisfactory solutions of hypergeometric recursions , 2007, Math. Comput..

[16]  Robert C. Forrey,et al.  Computing the Hypergeometric Function , 1997 .

[17]  Path integral treatment for the one‐dimensional Natanzon potentials , 1993 .

[18]  William H. Press,et al.  Numerical recipes in C , 2002 .

[19]  L. Chetouani,et al.  Path integral treatment for the generalized Ginocchio potentials , 1995 .

[20]  P ? ? ? ? ? ? ? % ? ? ? ? , 1991 .

[21]  H. V. Haeringen,et al.  Analytic T matrices for Coulomb plus rational separable potentials , 1975 .

[22]  M. Stoitsov,et al.  Use of the Ginocchio potential in mean-field studies and beyond , 1998 .

[23]  C. Eckart The Penetration of a Potential Barrier by Electrons , 1930 .

[24]  Nicholas J. Higham,et al.  INVERSE PROBLEMS NEWSLETTER , 1991 .

[25]  P. Schmelcher,et al.  The analytic continuation of the Gaussian hypergeometric function 2 F 1 ( a,b;c;z ) for arbitrary parameters , 2000 .

[26]  Guo Jian-you,et al.  Solution of the Dirac Equation with Special Hulthén Potentials , 2003 .

[27]  J. Adam A nonlinear eigenvalue problem in astrophysical magnetohydrodynamics: Some properties of the spectrum , 1989 .

[28]  R. M. EL-ASHWAH,et al.  HYPERGEOMETRIC FUNCTIONS , 2004 .

[29]  Levere C. Hostler Relativistic Coulomb Sturmian matrix elements and the Coulomb Green’s function of the second‐order Dirac equation , 1987 .

[30]  K. S. Kölbig,et al.  Errata: Milton Abramowitz and Irene A. Stegun, editors, Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables, National Bureau of Standards, Applied Mathematics Series, No. 55, U.S. Government Printing Office, Washington, D.C., 1994, and all known reprints , 1972 .

[31]  Guo Jian-you,et al.  Solution of the relativistic Dirac-Woods-Saxon problem , 2002 .

[32]  Contiguous relations of hypergeometric series , 2001, math/0109222.

[33]  J. Dobaczewski,et al.  Continuum effects for the mean-field and pairing properties of weakly bound nuclei , 1999, nucl-th/9905031.

[34]  Ericka Stricklin-Parker,et al.  Ann , 2005 .

[35]  Annie Cuyt,et al.  Gamma function and related functions , 2008 .

[36]  N. Temme Numerical aspects of special functions , 2007, Acta Numerica.

[37]  P. Cordero,et al.  Algebraic solution for the Natanzon hypergeometric potentials , 1994 .

[38]  Wolfgang Büring An analytic continuation of the hypergeometric series , 1987 .

[39]  Andrew G. Glen,et al.  APPL , 2001 .

[40]  E. Guth,et al.  Electric Excitation and Disintegration of Nuclei. I. Excitation and Disintegration of Nuclei by the Coulomb Field of Positive Particles , 1951 .