3
$\begingroup$

I'm after the numerical inverse Laplace transform of a function, so I type

f[t_?NumericQ] :=   InverseLaplaceTransform[1/(1 + s + ArcTanh[1/(s - 1)]), s, t];

This function may possibly suck, but it looks like its inverse Laplace transform should exist, albeit not in closed form (for instance, by the initial value theorem we know that f[t->0+]=1). However, when confronted with this task Mathematica simply replicates my command without giving a numerical solution.

Of course ArcTanh is a multivalued function in the complex plane, but shouldn't the algorithm used for Laplace inversion know how to handle this?

$\endgroup$
2
  • 1
    $\begingroup$ Support for numeric InverseLaplaceTransform was introduced in version 12.2. $\endgroup$ Jan 19, 2022 at 16:03
  • 2
    $\begingroup$ Your function works fine in Mathematica 13.0. $\endgroup$
    – John Doty
    Jan 19, 2022 at 17:39

1 Answer 1

3
$\begingroup$

We can expand function 1/(1 + s + ArcTanh[1/(s - 1)]) by series:

f[s_] := 1/(1 + s + ArcTanh[1/(s - 1)]);

g[t_, M_] := InverseLaplaceTransform[Series[f[s], {s, Infinity, M}] // Normal, s, t]; 
Plot[{g[t, 50]}, {t, 0, 3}]

N[g[2, 50], 15](*for t=2*)
(*-1.39795847070074*)

InverseLaplaceTransform[f[s], s, 2.0, WorkingPrecision -> 14]
(*for t=2. Works only in version 12.2 or above.*)
(*-1.39795847070073 + 0.*10^-17 I*)

Probably does not have finite closed-form expression in terms of very large class of special functions.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.