EDP (équation de la chaleur) — solved exercise (Master's)
Statement
Let u be the bounded solution of ∂ₜu = ∂ₓₓu on ℝ with u(x, 0) = e^(−x²). Compute u(0, 2).
Strategy
Use the heat kernel representation for the bounded solution on $\mathbb R$ and evaluate the Gaussian convolution at $(x,t)=(0,2)$.
Step-by-step solution
1. Heat kernel formula
For the heat equation on $\mathbb R$, the bounded solution with initial data $f(x)=e^{-x^2}$ is
$$
u(x,t)=\frac{1}{\sqrt{4\pi t}}\int_{-\infty}^{\infty} e^{-\frac{(x-y)^2}{4t}} f(y)\,dy.$$
📖 This is the standard heat kernel representation for $\partial_t u=\partial_{xx}u$ on the real line.
2. Evaluate at $x=0$, $t=2$
Substituting $x=0$ and $t=2$ gives
$$u(0,2)=\frac{1}{\sqrt{8\pi}}\int_{-\infty}^{\infty} e^{-\frac{y^2}{8}}e^{-y^2}\,dy.$$
📖 We use $f(y)=e^{-y^2}$ and $(0-y)^2=y^2$.
3. Combine the Gaussian exponentials
Since
$$e^{-\frac{y^2}{8}}e^{-y^2}=e^{-\frac{9y^2}{8}},$$
we get
$$u(0,2)=\frac{1}{\sqrt{8\pi}}\int_{-\infty}^{\infty}e^{-\frac{9y^2}{8}}\,dy.$$
📖 Add the exponents: $-y^2/8-y^2=-9y^2/8$.
4. Gaussian integral
Using
$$\int_{-\infty}^{\infty}e^{-ay^2}\,dy=\sqrt{\frac{\pi}{a}},\qquad a>0,$$
with $a=9/8$, we obtain
$$u(0,2)=\frac{1}{\sqrt{8\pi}}\sqrt{\frac{8\pi}{9}}=\frac{1}{3}.$$
📖 Standard Gaussian integral formula.
Result
$$u(0,2)=\frac{1}{3}.$$
Question-by-question check (two independent resolutions)
| 1 | 1/3 | ✘ sqrt(78)/26 |
Key concepts
Heat equationHeat kernelGaussian convolutionGaussian integral
Tip
Remember that the heat equation on $\mathbb R$ smooths initial data by convolution with the Gaussian heat kernel $G_t(x)=\frac{1}{\sqrt{4\pi t}}e^{-x^2/(4t)}$.
Have a similar exercise? Get it solved step by step, for free.
🧮 Solve my exercise