1

Find the generalized entropic solution of the Cauchy problem $$u_t+\left ( \frac{u^2}{2} \right )_x=0,u|_{t=0}=u_0$$ When $x<-2\Rightarrow u_0(x)=1$, when $-2<x<1\Rightarrow u_0(x)=-1$, when $x>1\Rightarrow u_0(x)=0$

I know it is a nonlinear Burgers equation $$u_t+\left ( \frac{u^2}{2} \right )_x=0, \; u(x, 0) = \begin{cases} 1, & \text{if} x < -2, \\ -1, & \text{if } -2 < x < 1, \\ 0, & \text{if } x > 1. \end{cases}$$

But there's a big problem with jumps. My attempt:

I will write the Burgers equation as a system of ordinary differential equations

$$ \frac{dx}{dt} = u, \quad \frac{du}{dt} = 0 $$

The characteristic equations become - $x(t) = ut + x_0$

  • At $x<-2$, $u=1$, so $x(t)=t+x_0$. Hence, $x_0=x-t$ and the solution $u=1$ for $x<t-2$.
  • At $-2<x<1$, $u=-1$, so $x(t)=-t+x_0$. Hence, $x_0=x+t$ and the solution $u=-1$ for $x>-t-1$ and $x<-t+1$.
  • At $x>1$, $u=0$, so $x(t)=x_0$. Hence, $x_0=x$ and the solution $u=0$ for $x>1$.

We have regions where jumps in the value of $u$ are possible. Consequently, we must treat them according to the entropy condition, which states that jumps must propagate at a rate that is the average of the velocities on the two sides of the jump. In our case, at the point $x=-2$ the jump goes from $1$ to $-1$, so that the speed of the jump is $(1 - 1) / 2 = 0$, which means that the jump is not moving. Similarly, at $x=1$ the jump goes from $-1$ to $0$, so the speed of the jump is $(-1 + 0) / 2 = -1.2$. And I got this solution

$$ u(x, t) = \begin{cases} 1, & \text{if } x < t - 2, \\ -1, & \text{if } -t - 1 < x < \min(-t + 1, 1 - 0.5t), \\ 0, & \text{if } x > \max(1 - 0.5t, -t + 1) \end{cases} $$

Did I get the cases and the jumps right? If there is an easier way, could you show me a simpler solution?

Partim
  • 891
  • You have one shock and one rarefaction front. The solution changes again when these two start to interact at time $t=3$. There could be a second event at $t=6$? A picture is always helpful to get the idea, see for instance https://math.stackexchange.com/questions/1534271/rarefaction-and-shock-waves-colliding-in-burgers-equation – Lutz Lehmann Jun 16 '23 at 04:04
  • Can you tell me if you've made the right decision or not? – Partim Jun 16 '23 at 20:42
  • No, it is not correct, you are missing the rarefaction sector. – Lutz Lehmann Jun 17 '23 at 02:37
  • Could you show me the right solution? – Partim Jun 17 '23 at 10:26

1 Answers1

1

No, you did not get the solution right. Let's draw the characteristics as they are close to the initial condition

enter image description here

Where the lines cross, there is a shock wave, here a standing wave, and where the lines are missing there will be a rarefaction wave, essentially filling the sector with rays from the point at the bottom.

enter image description here

Now we get intersecting lines from the left-most segment and the rarefaction rays. This is a dynamical situation, the shock wave gets computed via a differential equation $$ x_S'(t)=\frac{1+(x_S(t)-1)/t}{2}, ~~~x_S(3)=-2,~~ t\ge 3. $$ With its solution the shock wave continues on a parabolic path

enter image description here

Again there are unresolved intersections, now of rays from the left-most and right-most segment. The shock wave thus continues with constant slope $\frac12$. Remember that the slope of the ray is equal to the value of $u$ at each specific point.

Lutz Lehmann
  • 126,666