A computer algebra package for bivariate asymptotics with explicit error terms
Benjamin Hackl, Stephan Wagner
TL;DR
The paper presents a SageMath-based toolkit, dependent_bterms, for obtaining rigorous multivariate asymptotics with explicit error terms for binomial-sum expressions by extending SageMath's $B$-term framework to a dependent variable. It demonstrates the approach on two concrete problems: Ramanujan's Q-function and a Bóna–DeJonge inequality for 132-avoiding permutations, delivering fully quantified expansions and computable error bounds. The method combines a Mellin-transform framework, region-splitting, residue calculus, and automated error tracking to produce precise asymptotics and certify results for large $n$ while providing verified checks for small $n$. This yields both theoretical guarantees and practical computational tools for rigorous asymptotic analysis in combinatorics and related fields, streamlining what used to be highly manual, case-by-case labor.
Abstract
Making use of a newly developed package in the computer mathematics system SageMath, we show how to perform a full asymptotic analysis of certain types of sums that occur frequently in combinatorics, including explicit error bounds. We present two applications of the general approach to illustrate its use: the first concerns a classical problem due to Ramanujan, while the second one concerns a question of Bóna and DeJonge on 132-avoiding permutations with a unique longest increasing subsequence that can be translated into an inequality for a certain binomial sum.
