This is from a homework question 13.22 part (c) from "Mathematical Methods for Physic and Engineering" by Riley et. al on p. 464

I don't understand why the heaviside function is in the solution to this problem. I would be very grateful if someone could point me in the right direction.

The question states:

Find the function y(t) whose Laplace transforms is the following: $$e^{-(\gamma+s)}/[(s+\gamma)^2+b^2]$$

The correct solution uses the convolution formula for a Laplace transform to get:

$$ \begin{align} f(t) &= \frac{e^{-\gamma t_0}}{b} \int_0^{t} e^{-\gamma u}\sin(bu)H(u)\delta(t-u-t_0) \, du \\&= \frac{e^{-\gamma t_0}}{b} e^{-\gamma (t-t_0)} \sin[b(t-t_0)]H(t-t_0) \\ &= \frac{1}{b} e^{-\gamma t} sin[b(t-t_0)] H(t-t_0)\end{align}$$

where $\delta$ is the Dirac delta function and $H$ is the heavyside function.

To approach this problem I took $\bar g(s) = e^{-s t_0}$ and $\bar f(s) = \frac{b}{(s^2 + \gamma)^2 + b^2}$ as the Laplace transforms. From there I looked up the Laplace transforms in a table provided by the book and got $g(t) = \delta (t-t_0)$ and $f(t) = e^{-\gamma t} \sin(bt)$. I then applied the convolution theorem in the following way: $$\begin{align} \mathcal{L}^{-1} \left(\tfrac{1}{b} e^{-\gamma t_0 } \bar f(s) \bar g(s) \right) &= \frac{1}{b} e^{-\gamma t_0} \int_0^t e^{-\gamma u} \sin(bu) \delta(t-u-t_0) \, du \\ &= \frac{1}{b} e^{-\gamma t} \sin(b(t-t_0))\end{align}$$

where $\mathcal{L}$ is the Laplace transform. As you can see, my solution was similar to the correct solution, but has no heaviside function. I still can't figure out where the heaviside function comes from! Thank you in advance for your help.

  • $\begingroup$ the Heaviside function means that $f(t)$ is zero for $t \lt t_0$ $\endgroup$ – David Holden Nov 23 '14 at 3:58
  • $\begingroup$ @DavidHolden Oh I see, thank you. Is that something that should be evident from the context of the problem, or does it require some background in physics? $\endgroup$ – student Nov 23 '14 at 22:08
  • 1
    $\begingroup$ could you explain the use of the symbol $t_0$ in the solutions you give? $\endgroup$ – David Holden Nov 23 '14 at 23:54

By using the inverse Laplace transform, we should be able to clear up the confusion of where the Heaviside function comes into play. The inverse Laplace transform is $$ \mathcal{L}^{-1}\{F(s)\} = \frac{1}{2\pi i}\int_{\sigma-i\infty}^{\sigma + i\infty}F(s)e^{st}ds = \sum\text{Res} $$ If we expand the denominator, we have that $$ (s+\gamma)^2 +b^2 = s^2 + 2\gamma s+\gamma^2 + b^2. $$ Then the poles occur when the denominator is zero. Using the quadratic equation, we obtain that the poles occur when $$ s = -\gamma\pm ib. $$ Now, let's set up the inverse Laplace transform. $$ \mathcal{L}^{-1}\{F(s)\} = \frac{e^{-\gamma}}{2\pi i}\int_{\sigma-i\infty}^{\sigma + i\infty}\frac{\exp[s(t-1)]}{(s+\gamma - ib)(s+\gamma+ib)}ds $$ The numerator is $\exp[s(t-1)]$. In order for convergence, $s(t-1)<0$ or $t<1$. The definition of the Heaviside function (Unit step function) is $$ \mathcal{H}(t) = \mathcal{U}(t) = \begin{cases} 1, & t>0\\ 0, & t<0 \end{cases} $$ In order for convergence of the exponential term, we need $t<1$ or $\mathcal{H}(t-1)$ which will zero out all the terms with $t<1$. \begin{align} \frac{e^{-\gamma}}{2\pi i}\int_{\sigma-i\infty}^{\sigma + i\infty}\frac{\exp[s(t-1)]}{(s+\gamma - ib)(s+\gamma+ib)}ds &= e^{-\gamma}\biggl[\lim_{s\to -\gamma+ib}(s+\gamma-ib)\frac{e^{s(t-1)}}{(s+\gamma - ib)(s+\gamma+ib)}+\lim_{s\to -\gamma-ib}(s+\gamma+ib)\frac{e^{s(t-1)}}{(s+\gamma - ib)(s+\gamma+ib)}\biggr]\\ &=\frac{e^{-\gamma-\gamma(t-1)}\mathcal{H}(t-1)}{b}\biggl[\frac{e^{ib(t-1)}-e^{-ib(t-1)}}{2i}\biggr]\\ &= \frac{e^{-\gamma-\gamma(t-1)}\mathcal{H}(t-1)}{b}\sin[b(t-1)] \end{align} enter image description here

where I took $\gamma = b = 1$.


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.