# Thread: PDEs - Heat equation

1. ## PDEs - Heat equation

Hi, I have a heat equation question which I'm unsure of. The first two parts are ok, but I'll post them anyway so you have a bit of background info:

============================
a) Show that the solution of the heat equation $u_{t} = u_{xx}$ with initial condition $u(x,0) = f(x)$ can be written as

$u(x,t) = \int_{-\infty}^{\infty} \frac{1}{2\sqrt{{\pi}t}}\exp\left(-\frac{(x-x')^{2}}{4t}\right)f(x') dx'$

b) Show that the solutions of the heat equation can be mapped into non-trivial solutions of Burger's equation $v_{t} + vv_{x} = v_{xx}$ by the Cole-Hopf transformation

$v = \frac{{\alpha}u_{x}}{u}$
and find $\alpha$

These are just bookwork. We get $\alpha = -2$ for part b).
============================

The next bit is the problem:

c) By taking the singular data for the heat equation $f(x) = \delta^{\left(n\right)}(x)$ construct a solution $u(x,t)$ of the heat equation.

Now, I've gone through the solutions and the line has jumped from:

$u(x,t) = \frac{1}{2\sqrt{{\pi}t}}\int_{-\infty}^{\infty} \exp\left(-\frac{(x-x')^{2}}{4t}\right)f(x') dx'$

to

$u(x,t) = \frac{1}{2\sqrt{{\pi}t}}\left(-\frac{\partial}{{\partial}x}\right)^{n}\exp\left(-\frac{x^{2}}{4t}\right)$

This is where I'm stuck. I'm guessing it's some kind of delta function property but I'm not sure. The $\delta^{\left(n\right)}$ threw me a little. If someone can explain what's going on, that'd be great!

Thanks.

2. Originally Posted by sathys
[snip]
the solution of the heat equation $u_{t} = u_{xx}$ with initial condition $u(x,0) = f(x)$ can be written as

$u(x,t) = \int_{-\infty}^{\infty} \frac{1}{2\sqrt{{\pi}t}}\exp\left(-\frac{(x-x')^{2}}{4t}\right)f(x') dx'$

[snip]

The next bit is the problem:

c) By taking the singular data for the heat equation $f(x) = \delta^{\left(n\right)}(x)$ construct a solution $u(x,t)$ of the heat equation.

Now, I've gone through the solutions and the line has jumped from:

$u(x,t) = \frac{1}{2\sqrt{{\pi}t}}\int_{-\infty}^{\infty} \exp\left(-\frac{(x-x')^{2}}{4t}\right)f(x') dx'$

to

$u(x,t) = \frac{1}{2\sqrt{{\pi}t}}\left(-\frac{\partial}{{\partial}x}\right)^{n}\exp\left(-\frac{x^{2}}{4t}\right)$

This is where I'm stuck. I'm guessing it's some kind of delta function property but I'm not sure. The $\delta^{\left(n\right)}$ threw me a little. If someone can explain what's going on, that'd be great!

Thanks.
$\delta^{\left(n\right)}(x)$ denotes the nth derivative of the Dirac delta function. The 'jump' uses a property of the nth derivative of the Dirac delta function.

For example, $\int_{-\infty}^{+ \infty} \delta^{\left(1\right)}(x') \, f(x', x) \, dx' = -\frac{df(x)}{dx}$.

More generally:

$\int_{-\infty}^{+ \infty} \delta^{\left(n\right)}(x') \, f(x', x) \, dx' = (-1)^n \frac{d^n f(x)}{dx^n}$.