Weak Solution to Dx u + 3yu = 0 with Heaviside Step Function Boundary Condition
Jump to navigation
Jump to search
Theorem
Consider the boundary value problem:
- $\begin{cases}
\dfrac {\partial u} {\partial x} + 3 y u = 0 & : x \in \R_{>0},~ y \in \R \\ & \\ \map u {0, y} = \map H y & : y \in \R \\ \end{cases}$
Then it has a weak solution of the form:
- $u = e^{-3 y x} \map H y$
Proof
Let $u = e^{-3 y x} \map H y$
We have that:
- Heaviside Step Function is Locally Integrable
- Locally Integrable Function defines Distribution
- Multiplication of Distribution induced by Locally Integrable Function by Smooth Function
Hence, we can define a distribution $T_u \in \map {\DD'} {\R^2}$ associated with $u$.
Then in the distributional sense we have that:
\(\ds \dfrac {\partial u}{\partial x}\) | \(=\) | \(\ds \dfrac {\partial}{\partial x} \paren {e^{-3 y x} \map H y}\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds e^{-3 y x} \paren {-3y} \map H y + e^{-3 y x} \dfrac {\partial \map H y}{\partial x}\) |
That is:
\(\ds \dfrac {\partial T_u}{\partial x}\) | \(=\) | \(\ds \dfrac {\partial}{\partial x} T_{e^{-3 y x} \map H y}\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds \dfrac {\partial}{\partial x} \paren {e^{-3 y x} T_{\map H y} }\) | Multiplication of Distribution induced by Locally Integrable Function by Smooth Function | |||||||||||
\(\ds \) | \(=\) | \(\ds e^{-3 y x} \paren {-3y} T_{\map H y} + e^{-3 y x} \dfrac {\partial}{\partial x} T_{\map H y}\) | Product Rule for Distributional Derivatives of Distributions multiplied by Smooth Functions |
Let $\phi \in \map \DD {\R^2}$ be a test function.
Then:
\(\ds \dfrac {\partial}{\partial x} \map {T_{\map H y} } \phi\) | \(=\) | \(\ds - \map {T_{\map H y} } {\dfrac {\partial \phi}{\partial x} }\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds - \int_{-\infty}^\infty \int_{-\infty}^\infty \map H y \dfrac {\partial \phi}{\partial x} \rd x \rd y\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds - \int_0^\infty \int_{-\infty}^\infty \dfrac {\partial \phi}{\partial x} \rd x \rd y\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds - \int_0^\infty \paren {\bigintlimits {\map \phi {x, y} } {x \mathop = - \infty} {x \mathop = \infty} } \rd y\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds - \int_0^\infty 0 \rd y\) | Definition of Test Function | |||||||||||
\(\ds \) | \(=\) | \(\ds 0\) |
Hence, in the distributional sense we have that:
- $\dfrac \partial {\partial x} \map H y = \mathbf 0$
where $\mathbf 0$ is the zero distribution.
Therefore:
\(\ds \dfrac {\partial u} {\partial x} + 3 y u\) | \(=\) | \(\ds e^{-3xy} \paren {-3y} \map H y + 3y e^{-3yx} \map H y\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds \mathbf 0\) |
Moreover:
\(\ds \map u {0, y}\) | \(=\) | \(\ds e^{-0} \map H y\) | ||||||||||||
\(\ds \) | \(=\) | \(\ds \map H y\) |
$\blacksquare$
Sources
- 2017: Amol Sasane: A Friendly Approach to Functional Analysis ... (previous) ... (next): Chapter $\S 6.4$: A glimpse of distribution theory. Multiplication by $C^\infty$ functions