Two questions about PDE (sum of solutions)

439 Views Asked by At

In an old exercise, I tried to solve a the following problem

$$ \begin{cases} u_t + xu_x = u\\ u(0,x)=x^3 \end{cases} $$

I solved the equation with the equalities

$$\frac{1}{dt}=\frac{x}{dx}=\frac{u}{du}$$

Treating 1st and 2nd side of the equalities

$$\frac{1}{dt}=\frac{x}{dx} \quad\rightarrow\quad dt=\frac{dx}{x} \quad\rightarrow\quad t+k_1 = \ln|x| + k_2 \quad\rightarrow\quad k_1 - k_2 =\ln|x|-t$$

$$\rightarrow\quad e^{k_1 - k_2} =e^{\ln|x|-t} \quad\rightarrow\quad e^{k_1 - k_2} =e^{\ln|x|}e^{-t} \quad\rightarrow\quad \pm e^{k_1 - k_2} =xe^{-t} $$

Fixing $\zeta$

$$\zeta:=\pm e^{k_1 - k_2} =xe^{-t} $$

A part of general solution yields

$$\Phi(\zeta)=\Phi(xe^{-t}) $$

Treating now 1st and 3rd

$$\frac{1}{dt}=\frac{u}{du} \quad\rightarrow\quad \frac{du}{u}=dt \quad\rightarrow\quad \ln|u|+k_3=t+k_4 \quad\rightarrow\quad \ln|u|=t+k_4-k_3$$

$$\rightarrow\quad u=\pm e^{k_4-k_3}e^t $$

Fixed the constant $K:=\pm e^{k_4-k_3}$, we can say

$$ u=Ke^t $$

Detected the constant in the general solution, we replace it with $\Phi(\zeta)$.

$$ u_1(t,x)=e^t\Phi(xe^{-t}) $$

Then I tried to solve by verifying the equality between 2nd and 3rd side.

$$\frac{x}{dx}=\frac{u}{du} \quad\rightarrow\quad \frac{du}{dx}=\frac{u}{x} \quad\rightarrow\quad \frac{u'(x)}{u(x)}=\frac{1}{x} \quad\rightarrow$$

$$\rightarrow\quad \ln|u(x)|+k_5=\ln|x|+k_6 \quad\rightarrow\quad \ldots \quad\rightarrow\quad u(x)=\pm e^{k_6-k_5}x $$

Fixed as constant $H:=\pm e^{k_6-k_5}$, we can say

$$ u(x)=Hx $$

Detected the constant in the general solution, we replace it with $\Phi(\zeta)$ found in the first phase.

$$ u_2(t,x)=x\Phi(xe^{-t}) $$

So, both $u_1(t,x)$ and $u_2(t,x)$ are suitable for the PDE.

I tried to sum them up and I noticed that $u(t,x)=u_1(t,x)+u_2(t,x)$ also verifies the PDE!

Question 1: Is there a general form (even more general than the sum) to indicate the most general solution for the PDE?

Then:

I also tried to evaluate the initial condition in order to find the particular solution but I fulfilled the request using $u_1(t,x)$ ONLY. I will show you what I have done

$$x^3 = u(0,x)=\Phi(x)$$

$$\zeta:=x \quad\rightarrow\quad x^3=\zeta^3$$

Yielding:

$$ u_1(t,x)=e^t(xe^{-t})^3=x^3 e^{-2t} $$

But I failed in evaluate the initial condition with $u_2(t,x)$.

Question 2: Can anyone help me to find the particular solution using $u(t,x)=u_1(t,x)+u_2(t,x)$? Can anyone show me the correct passages?

Thanks to all in advance.

1

There are 1 best solutions below

1
On BEST ANSWER

The sum of these solutions is a solution because both are different forms for the most general solution possible.

Consider $u_1(t,x)=e^t\Phi(xe^{-t})$ and $u_2(t,x)=x\Psi(xe^{-t})$

Now, we can always set $\Phi(xe^{-t})=xe^{-t}\Phi_1(xe^{-t})$ for some $\Phi_1$, continuous and differentiable too. Then,

$u_1(t,x)=e^t\Phi(xe^{-t})=e^txe^{-t}\Phi_1(xe^{-t})=x\Phi_1(xe^{-t})$

So is, $u_1=u_2$ because obviously $\Psi=\Phi_1$

You had the general solution inasmuch you had an expression depending on an arbitrary function (to determine by the initial conditions)

Solving using $u_2(t,x)=x\Psi(xe^{-t})$

$u_2(0,x)=x^3=x\Psi(x)$

$\Psi(x)=x^2$ and

$u_2(t,x)=x(xe^{-t})^2=x^3e^{-2t}$

The same function, as expected, so, no need to sum them.