0
$\begingroup$

While trying to evaluate the Laplace transform below $$I = \int_{0}^{\infty}e^{-st}B(\frac{1}{2}-it,\frac{3}{2}+it)\mathrm{d}t,$$

invoking

LaplaceTransform[Beta[0.5-it, 1.5+it], t, s],

yields

$$\frac{B(\frac{1}{2}-it,\frac{3}{2}+it)}{s},$$

which is quite weird since the output of a Laplace transform has to be an exclusive function of $s$, but this output still includes $t$. What am I missing here?

$\endgroup$
8
  • $\begingroup$ Looks like a bug, you shouldn't have s in output..I came across similar bugs, the solution was to rewrite integral using Feynam trick $\endgroup$ Feb 6 at 1:39
  • $\begingroup$ @YaroslavBulatov: $s$ is OK. One should not have $t$ in the output. $\endgroup$
    – User
    Feb 6 at 1:42
  • $\begingroup$ There were couple of bugs I ran across in laplace.transform, my workaround is using ilaplaceFeynmann and ilaplaceMellin helpers from mathematica.stackexchange.com/q/286895/217 $\endgroup$ Feb 6 at 1:43
  • $\begingroup$ @YaroslavBulatov: The definitions of these transforms are apparently not the same as that ofLaplaceTransform. So, how do their results help here? $\endgroup$
    – User
    Feb 6 at 1:46
  • 3
    $\begingroup$ The Beta function does not include t, only the variable it. What did you intend? i*t or I*t? $\endgroup$
    – Bob Hanlon
    Feb 6 at 2:27

1 Answer 1

2
$\begingroup$
$Version

(* "14.0.0 for Mac OS X ARM (64-bit) (December 13, 2023)" *)

Clear["Global`*"]

Mathematica appears to be able to handle imaginary values for i

LaplaceTransform[Beta[1/2 - I*t, 3/2 + I*t] //
   FunctionExpand, t, s] // FullSimplify

(* (1/(8 π))(-2 π PolyGamma[0, (π + s)/(4 π)] + 
  2 π PolyGamma[0, 1/4 (3 + s/π)] + 
  I (PolyGamma[1, (π + s)/(4 π)] - PolyGamma[1, 1/4 (3 + s/π)])) *)

Letting i == a*I,

f[i_, s_] = ((LaplaceTransform[(Beta[1/2 - i*t, 3/2 + i*t] /. i -> a*I) //
        FunctionExpand, t, s] // FullSimplify) /. a -> i/I) // FullSimplify

(* (1/(8 i π))(-2 I π (PolyGamma[0, 1/4 + (I s)/(4 i π)] - 
     PolyGamma[0, 3/4 + (I s)/(4 i π)]) - 
  PolyGamma[1, 1/4 (1 + (I s)/(i π))] + 
  PolyGamma[1, 1/4 (3 + (I s)/(i π))]) *)
$\endgroup$

Your Answer

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

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