Fast and stable rational approximation of generalized hypergeometric functions
Richard Mikael Slevinsky

TL;DR
This paper introduces efficient, stable recurrence relations for rational approximations of generalized hypergeometric functions, improving computational stability and confirming the superiority of factorial Levin-type transformations over Drummond transformations.
Contribution
Derives new recurrence relations for rational approximations of ${}_pF_q$ functions that are more stable and computationally efficient than previous methods.
Findings
Recurrence relations require $ ext{O}[ ext{max}\{p,q ight ext{(n+k)}]$ flops.
Numerical evidence shows increased stability of the new recurrence relations.
Theoretical analysis confirms the superiority of factorial Levin-type transformation over Drummond transformation.
Abstract
Rational approximations of generalized hypergeometric functions of type are constructed by the Drummond and factorial Levin-type sequence transformations. We derive recurrence relations for these rational approximations that require flops. These recurrence relations come in two forms: for the successive numerators and denominators; and, for an auxiliary rational sequence and the rational approximations themselves. Numerical evidence suggests that these recurrence relations are much more stable than the original formul\ae~for the Drummond and factorial Levin-type sequence transformations. Theoretical results on the placement of the poles of both transformations confirm the superiority of factorial Levin-type transformation over the Drummond transformation.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsIterative Methods for Nonlinear Equations · Mathematical functions and polynomials · Numerical methods for differential equations
