If I had a function $\widehat f(s)$ how would I know if there exists a function $f(t)$ so that the laplace transform of $f$ is $\widehat f$?

From looking at the formula for finding the laplace transform $$\widehat f(s) = \int_o^\infty e^{-st}f(t)dt$$ it seems that $\widehat f$ must be a non-increasing function since as $s$ increases the integrand approaches $0$ would this be correct?

However I have found that $$\mathcal{L}^{-1}\{s\}=\delta'(t)$$ even though $s$ is increasing. Although I am not quite sure what the derivative of the dirac delta function means since the dirac delta function not exactly a function and is called a 'generalised function'?

Are there any conditions on $\widehat f$ would prevent it from having an inverse laplace transform?

  • $\begingroup$ Do you want the inverse transform to exist in the "usual sense" as a function? Or do you want to the inverse to exist in the sense of distribution theory? How much distribution theory do you know? $\endgroup$ – mrf Sep 10 '15 at 11:27
  • $\begingroup$ Probably the usual sense.. No idea what distribution theory is. $\endgroup$ – user3128875 Sep 10 '15 at 11:41

Maybe it helps to consider the complex Laplace transform. Thus $$ \hat{f}(z)=\int_{0}^{\infty }dt\exp [izt]f(t),\;{Im}z>0 $$ This object will exist for $f(t)$ absolutely integrable or square integrable. Now put $z=\omega +i\delta $. Then ($\theta (t)$ is the Heaviside step function) $$ \hat{f}(\omega +i\delta )=\int_{0}^{\infty }dt\exp [i(\omega +i\delta )t]f(t)=\int_{-\infty }^{+\infty }dt\theta (t)\exp [-\delta t]f(t)\exp [i\omega t] $$ so it is the Fourier transform of $\theta (t)\exp [-\delta t]f(t)$ and% \begin{eqnarray*} \theta (t)\exp [-\delta t]f(t) &=&\frac{1}{2\pi }\int_{-\infty }^{+\infty }d\omega \exp [-i\omega t]\hat{f}(\omega +i\delta ) \\ f(t) &=&\frac{1}{2\pi }\int_{-\infty }^{+\infty }d\omega \exp [-i(\omega +i\delta )t]\hat{f}(\omega +i\delta ),\;t>0 \end{eqnarray*} In the literature the integral is usually given in terms of a straight integration path parallel to the real axis. You see that in your case $% s=\delta $ but $\omega $ vanishes. What you need is a requirement about the analytic continuation in a strip including the real axis.

Your special case requires an extension of the theory of Fourier transforms to distributions. But that is a different story.


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.