Given the following partial differential equation with initial condition $$tu_x+u_t=2t\cdot u,\;IV: u(x,0)=x.$$ So I apply the method of characteristic to solve this problem as following

Forming three characteristic equation

$$\frac{dx}{ds}=t,\;\frac{dt}{ds}=1,\; and\;\frac{du}{ds}=2t\cdot u $$ $$\frac{dt}{ds}=1\rightarrow dt = ds\rightarrow t = s+t_0\rightarrow t=s\;since\; t_0=0\; from\; IV$$ then $$\frac{dx}{ds}=t\rightarrow \frac{dx}{ds}=s\rightarrow x=\frac{s^2}{2}+x_0$$ also for the last equation we have $$\frac{du}{ds}=2t\cdot u\rightarrow \frac{du}{u}=2s\;dt\rightarrow u=e^{s^2+u_0}$$ Then we substitute the $s=t$, and $x_0=x-\frac{s^2}{2}$ to initial condition $$ u(x_0,0)=e^{0+u_0}=x\rightarrow u_0=lnx$$ Thus the finial solution I have should be $u(x,t)=e^{t^2+lnx}=xe^{t^2}$

However this is not same as the answer given, could someone tell me where I make the mistake in this problem?


1 Answer 1


$$tu_x-u_t=tu$$ Charpit-Lagrange system of characteristic ODEs : $$\frac{dx}{t}=\frac{dt}{1}=\frac{du}{tu}=ds$$ This is what you correctly found, but presented on an equivalent form.

A first characteristic equation comes from solving $\frac{dx}{t}=\frac{dt}{1}$ : $$t^2-2x=c_1$$ A second characteristic equation comes from solving $\frac{dt}{1}=\frac{du}{tu}$ : $$ue^{-x}=c_2$$ The general solution of the PDE expressed on the form of implicit equation $c_2=F(c_1)$ is : $$ue^{-x}=F(t^2-2x)$$ $F$ is an arbitrary function (to be determined later in order to satisfy the initial condition). $$\boxed{u(x,t)=e^xF(t^2-2x)}$$


$$u(x,0)=x=e^xF(-2x)$$ $$F(-2x)=xe^{-x}$$ Let $x=-\frac12 X$ $$F(X)=-\frac12 X e^{\frac12 X}$$ Now the function $F(X)$ is known. We put it into the above general solution where $X=t^2-2x$

$F(t^2-2x)=-\frac12 (t^2-2x) e^{\frac12 (t^2-2x)}$ $$u(x,t)=e^x\left(-\frac12 (t^2-2x) e^{\frac12 (t^2-2x)}\right)$$ After simplification, the particular solution which satisfies both the PDE and the condition is : $$\boxed{u(x,t)=(x-\frac{t^2}{2})e^{\frac12 t^2}}$$

  • $\begingroup$ Sorry, Why we could write C2 = F(C1)? $\endgroup$
    – Snow Max
    Nov 9, 2020 at 4:46
  • $\begingroup$ This is a well known property for first order quasilinear PDEs without using boundary condition (Using boundary condition is another way if you prefer). Intuitive oversimplified explanation : Since $c_1$ and $c_2$ are arbitrary constants they are an infinity of equations such as $\Phi(c_1,c_2)=0$ with arbitrary function $\Phi$. Or equivalently on the form of any relationship between the two arbitrary constants : $c_2=F(c_1)$ or equivalently $c_1=G(c_2)$ with arbitrary functions $F$ and $G$ one inverse of the other. $\endgroup$
    – JJacquelin
    Nov 9, 2020 at 8:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.