3
$\begingroup$

I have a 1-dimensional homogeneous heat equation: $$ u''(x, t) = \dot u(x, t)$$

The initial value is $u(x, 0) = \exp\left(-x^2\right)$.

I plugged this into the solution formula: $$ u(x, t) = \frac{1}{\sqrt{4 \pi t}} \int_{-\infty}^\infty \mathrm dy \, \exp\left(-\frac{(x-y)^2}{4t} - y^2\right)$$

Since I am not sure how to calculate this integral, I typed it into Mathematica and got:

1/Sqrt[4 Pi t] Integrate[
  Exp[((y - x)^2)/(4 t) - y^2], {y, -Infinity, Infinity}]

$$\text{ConditionalExpression}\left[\frac{\exp\left(\frac{x^2}{4 t-1}\right)}{\sqrt{4-\frac{1}{t}} \sqrt{t}},\left(4 \Re(t)\neq 1\lor \Re\left(\frac{x}{t}\right)>0\right)\land \Re\left(\frac{1}{t}\right)<4\right]$$

Plotting this looks strange at best:

http://wstaw.org/m/2012/11/10/1_1.png

For $t=0$, it is not even close to the initial value.

  • How do I solve this integral?

  • Is what Mathematica gave me correct in some sort?

$\endgroup$
4
$\begingroup$

Setting $$ \xi=\sqrt{\frac{1+4t}{4t}}\left(y-\frac{x}{1+4t}\right), $$ we have: \begin{eqnarray} \frac{(x-y)^2}{4t}+y^2&=&\frac{(1+4t)y^2-2xy+x^2}{4t}\\ &=&\frac{1+4t}{4t}\left[y^2-\frac{2x}{1+4t}y+\frac{x^2}{1+4t}\right]\\ &=&\frac{1+4t}{4t}\left[\left(y-\frac{x}{1+4t}\right)^2+\frac{x^2}{1+4t}-\frac{x^2}{(1+4t)^2}\right]\\ &=&\frac{1+4t}{4t}\left[\left(y-\frac{x}{1+4t}\right)^2+\frac{4x^2t}{(1+4t)^2}\right]\\ &=&\xi^2+\frac{x^2}{1+4t}. \end{eqnarray} It follows that \begin{eqnarray} u(x,t)&=&\frac{1}{\sqrt{4\pi t}}\int_{-\infty}^\infty\exp\left(-\frac{(x-y)^2}{4t}-y^2\right)\, dy\\ &=&\frac{1}{\sqrt{4\pi t}}\sqrt{\frac{4t}{1+4t}}\exp\left(-\frac{x^2}{1+4t}\right)\int_{-\infty}^\infty\exp(-\xi^2)\, d\xi\\ &=&\frac{1}{\sqrt{\pi(1+4t)}}\exp\left(-\frac{x^2}{1+4t}\right)\int_{-\infty}^\infty\exp(-\xi^2)\, d\xi. \end{eqnarray} Using the fact that $$ \int_{-\infty}^\infty\exp(-\xi^2)\, d\xi=\sqrt{\pi}, $$ we get $$ u(x,t)=\frac{1}{\sqrt{1+4t}}\exp\left(-\frac{x^2}{1+4t}\right). $$

$\endgroup$
  • $\begingroup$ Thank you! Is there any way I could have come up with this by myself? Is that $\xi$ something like parabolic coordinates? $\endgroup$ – Martin Ueding Nov 10 '12 at 16:45
  • $\begingroup$ Following Mercy's example, I answered my own question. link $\endgroup$ – Patrick Windance Dec 20 '15 at 6:45

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.