Approximate solutions of the hyperbolic Kepler equation

Abstract: We provide an approximate zero

w i d e t i l d e S (g, L)

for the hyperbolic Kepler's equation

S - g, m a s i n h (S) - L = 0

for

g i n (0, 1)

and

L i n [0, i n f t y)

. We prove, by using Smale's

a l p h a

-theory, that Newton's method starting at our approximate zero produces a sequence that converges to the actual solution

S (g, L)

at quadratic speed, i.e. if

S_{n}

is the value obtained after

n

iterations, then

| S_{n} - S | l e q 0. 5^{2^{n} - 1} | w i d e t i l d e S - S |

. The approximate zero

w i d e t i l d e S (g, L)

is a piecewise-defined function involving several linear expressions and one with cubic and square roots. In bounded regions of

(0, 1) i m e s [0, i n f t y)

that exclude a small neighborhood of

g = 1, L = 0

, we also provide a method to construct simpler starters involving only constants.

Recommendations

Cites work

Cited in

(8)

This page was built for publication: Approximate solutions of the hyperbolic Kepler equation