Rational Approximation

Numerical Methods

Source inspiration: (Mathew 2000-2019).

Animations

The animation below compares the diagonal Padé approximant \(R_{[n/n]}(x)\) with the Taylor polynomial \(T_{2n}(x)\) of equal total degree for \(f(x) = e^x\) on \([-6,6]\), as the order \(n\) increases from 1 to 4.

Julia source scripts that generated these animations are linked under each case.

Case 1 — Padé \([n/n]\) vs Taylor \(T_{2n}\) for \(e^x\), \(n=1,\ldots,4\)

Behavior: Both approximations use \(2n+1\) parameters. The Taylor polynomial is accurate near \(x=0\) but diverges away from the origin. The Padé rational approximant distributes its accuracy more evenly, remaining close to \(e^x\) over a much wider interval. For \(n=4\) the Padé \([4/4]\) tracks \(e^x\) well past \(x=\pm5\), where \(T_8\) has already diverged.

Julia source

Pade [n/n] approximant (green) vs Taylor T_{2n} (red dashed) for e^x on [-6,6] for n=1,2,3,4; Pade stays close to the true function far from the origin while Taylor diverges

Derivation Notes (Planned)

Short derivations will be added to explain the core equations and assumptions.

References

Mathew, John H. 2000-2019. Numerical Analysis - Numerical Methods Modules. https://web.archive.org/web/20190808102217/http://mathfaculty.fullerton.edu/mathews/n2003/NumericalUndergradMod.html.