4
$\begingroup$

I want to calculate Inverse Laplace Transform of $s^n$. I have an idea, but I do not know if it works?

We have a formula to compute inverse laplace transforms of functions as below,

$$\mathcal{L}^{-1} [ F(s) ] = -\frac{\mathcal{L}^{-1} [ F^{\prime}(s) ]}{t}.$$

So from the given formula, we can obtain

$$\mathcal{L}^{-1} [ s ]= -\frac{\mathcal{L}^{-1} [ 1 ]}{t}= -\frac{\delta (t)}{t}.$$ and as a result, $$\mathcal{L}^{-1} [ s^n ] = (-1)^n\frac{n!\delta (t)}{t^n}$$ Is it right? In fact, I want to know the necessary conditions to use the given formula.

$\endgroup$
0
6
$\begingroup$

Intuitively, the derivative of the Dirac delta function $\delta'$ has Laplace transform $s$. The derivative of the Dirac delta is a generalized function that pulls out the derivative of the function with a change of sign: for any interval $[a,b]$ where $a < 0 < b$,

$$\int_a^b \delta'(t)f(t) \ dt = \left[ \delta(t) f(t) \right]_a^b - \int_a^b \delta(t) f'(t) \ dt = -f'(0)$$

Applying that procedure inductively,

$$L^{-1}\{s^n\}(t) = \delta^{(n)}(t)$$

$\endgroup$
1
  • 1
    $\begingroup$ This answer is not correct. The inverse Laplace transform of $s^n$, for non-negative integer values of $n$ is given by $$\mathscr{L}^{-1}\{s^n\}=\delta^{(n)}(t)$$There is no factor of $(-1)^n$. Note that for $f(t)=e^{-st}$, $(-1)^n \frac{d^n e^{-st}}{dt^n}=s^ne^{-st}$. I'll edit the sign error accordingly. $\endgroup$ – Mark Viola Sep 30 '20 at 16:40
2
$\begingroup$

$\newcommand{\bbx}[1]{\,\bbox[8px,border:1px groove navy]{\displaystyle{#1}}\,} \newcommand{\braces}[1]{\left\lbrace\,{#1}\,\right\rbrace} \newcommand{\bracks}[1]{\left\lbrack\,{#1}\,\right\rbrack} \newcommand{\dd}{\mathrm{d}} \newcommand{\ds}[1]{\displaystyle{#1}} \newcommand{\expo}[1]{\,\mathrm{e}^{#1}\,} \newcommand{\ic}{\mathrm{i}} \newcommand{\mc}[1]{\mathcal{#1}} \newcommand{\mrm}[1]{\mathrm{#1}} \newcommand{\pars}[1]{\left(\,{#1}\,\right)} \newcommand{\partiald}[3][]{\frac{\partial^{#1} #2}{\partial #3^{#1}}} \newcommand{\root}[2][]{\,\sqrt[#1]{\,{#2}\,}\,} \newcommand{\totald}[3][]{\frac{\mathrm{d}^{#1} #2}{\mathrm{d} #3^{#1}}} \newcommand{\verts}[1]{\left\vert\,{#1}\,\right\vert}$ \begin{align} &\left.\int_{0^{+}\ -\ \infty\ic}^{0^{+}\ -\ \infty\ic}s^{n}\expo{ts} {\dd s \over 2\pi\ic}\,\right\vert_{\ t\ >\ 0} = \int_{0^{+}\ -\ \infty\ic}^{0^{+}\ -\ \infty\ic}s^{n}\expo{ts}{\dd s \over 2\pi\ic} \\[5mm] \stackrel{\mrm{as}\ \epsilon\ \to\ 0^{+}}{\sim} &\ -\int_{-\infty}^{-\epsilon}\pars{-s}^{n}\expo{n\pi\ic}\expo{ts} \,{\dd s \over 2\pi\ic} - \int_{\pi}^{-\pi}\epsilon^{n}\expo{\ic n\theta}\epsilon\expo{\ic\theta}\ic\, \,{\dd s \over 2\pi\ic} - \int_{-\epsilon}^{-\infty}\pars{-s}^{n}\expo{-n\pi\ic}\expo{ts} \,{\dd s \over 2\pi\ic} \\[5mm] = &\ -\expo{n\pi\ic}\int_{\epsilon}^{\infty}s^{n}\expo{-ts}\,{\dd s \over 2\pi\ic} - {\epsilon^{n + 1}\sin\pars{n\pi} \over \pars{n + 1}\pi} + \expo{-n\pi\ic}\int_{\epsilon}^{\infty}s^{n}\expo{-ts}\,{\dd s \over 2\pi\ic} \\[5mm] = &\ -\,{\epsilon^{n + 1}\sin\pars{n\pi} \over \pars{n + 1}\pi} - {\sin\pars{n\pi} \over \pi} \int_{\epsilon}^{\infty}s^{n}\expo{-ts}\,\dd s \\[5mm] = &\ -\,{\epsilon^{n + 1}\sin\pars{n\pi} \over \pars{n + 1}\pi} - {\sin\pars{n\pi} \over \pi}\bracks{% -\,{\epsilon^{n + 1} \over n + 1}\,\expo{-t\epsilon} - \int_{\epsilon}^{\infty}{s^{n + 1} \over n + 1}\expo{-ts}\pars{-t}\,\dd s} \\[5mm] \stackrel{\mrm{as}\ \epsilon\ \to\ 0^{+}}{\sim} &\ -\,{\sin\pars{n\pi} \over \pars{n + 1}\pi}\,t \int_{\epsilon}^{\infty}s^{n + 1}\expo{-ts}\,\dd s \stackrel{\mrm{as}\ \epsilon\ \to\ 0^{+}}{\to} -\,{\sin\pars{n\pi} \over \pars{n + 1}\pi}\,t^{-n - 1}\,\Gamma\pars{n + 2} \\[5mm] = &\ -\,{\sin\pars{n\pi} \over \pars{n + 1}\pi}\,t^{-n - 1}\, {\pi \over \Gamma\pars{-1 - n}\sin\pars{\pi\bracks{n + 2}}} = -\,{1 \over n + 1}\,t^{-n - 1}\, {1 \over \Gamma\pars{-n}/\pars{-1 - n}} \\[5mm] = & \bbx{\ds{1 \over t^{n + 1}\Gamma\pars{-n}}} \\ & \end{align}

$\endgroup$
1
  • 1
    $\begingroup$ I feel that you should give some more insights (especially regarding the difference to the answer of Simon S). I guess your result is only valid for $n<0$? $\endgroup$ – Fabian Mar 6 '18 at 11:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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