# heat equation problem

• Oct 11th 2011, 04:29 AM
mgarson
heat equation problem
Hi!

I've got the following problem:

$f:\mathbb{R}\rightarrow\mathbb{R}$ a $C^2$-function such that $f$ is convex and $f(0)=f'(0)=0$. $g\in C^{\infty}([0,\infty),\mathscr{S}(\mathbb{R}^d))$ a real-valued solution to the heat equation $g_t=\Delta g$. Using this show that $F\in C^1$ and $F$ decrasing, where:

$F(t)=\int_{\mathbb{R}^d}f(g(x,t))dx$.

----------------------------------------------------------------
'I started by just diff. $F$ and I got

$F'(t) = \frac{d}{dt}\int_{\mathbb{R}^d}f(g(x,t))dx = \int_{\mathbb{R}^d} \frac{d}{dt}f(g(x,t))dx = \int_{\mathbb{R}^d}\frac{df(g)}{dg}g_tdx = \int_{\mathbb{R}^d} \frac{df}{dg}\Delta gdx$.

What next?
• Oct 14th 2011, 03:05 PM
Jose27
Re: heat equation problem
Notice

$F'(t)= \int_{ \mathbb{R}^d } f'(g)\Delta g dx = - \int_{ \mathbb{R}^d } \nabla \left( f'(g) \right) \cdot \nabla (g)dx = -\int_{ \mathbb{R}^d } f''(g)|\nabla(g)|^2 dx\leq0$

where we first integrated by parts and then used that $f''\geq 0$.