3
$\begingroup$

Performing certain manipulations on pseudo-differential equations I have come across the following first order equation: $$ D_{t}u-\lambda(t,x,D_{t},D_{x})u=0, \ \ (*) $$ where $\lambda$ is a scalar pseudo-differential operator with the principal symbol being real-valued and independent of $D_{t}.$ But, the lower-order terms of $\lambda$ depend on $D_{t}$ (or $\tau$ at the symbol level).

I was expecting a hyperbolic equation. But, I find that standard text books(like M.Taylor's 'Pseudo-differential operators') treat only equations in which $\lambda$ term is independent of $D_{t}$ (or $\tau$ at the symbol level).

For the equation $(*)$ to be hyperbolic, is it necessary that $\lambda$ to be independent of $D_{t}$? Are there any references which discuss these issues?

$\endgroup$

2 Answers 2

3
$\begingroup$

Your principal symbol is $\tau-\lambda_1 (t,x,\xi)$, with a real-valued $\lambda_1$, but you have lower order terms which may depend on $\tau$. You can get rid of that dependency as follows: take for instance $\lambda_0$ of order $0$. You can find an operator $M$ of order $0$ such that $$ Op\bigl(\tau-\lambda_1 (t,x,\xi)-\lambda_0(t,x,\tau,\xi)\bigr)= e^{-iM}Op\bigl(\tau-\lambda_1 (t,x,\xi)\bigr)e^{iM}+Op(S^{-1}). $$ In fact, since $e^{iM}$ is a pseudodifferential operator of order 0, the composition formula gives $$ e^{-iM}Op\bigl(\tau-\lambda_1 (t,x,\xi)\bigr)e^{iM}=D_t+\frac{\partial M}{\partial t} -e^{-iM}[Op(\lambda_1),e^{iM}]-Op(\lambda_1), $$ and since $ e^{-iM}[Op(\lambda_1),e^{iM}]=Op(${$\lambda_1,m$}$)+Op(S^{-1}) $ where $m$ is of order 0 and {} is the Poisson bracket. To get rid of $\lambda_0$, you have only to solve $$ \frac{\partial m}{\partial t}-\text{ {$\lambda_1,m$ } }=-\lambda_0, $$ which is a linear transport equation of real principal type in $m$. You can of course iterate this business to get a remainder of order as negative as you like. This is explained in the Chapter 23 on hyperbolic equations in Hörmander third volume of ALPDO.

$\endgroup$
3
$\begingroup$

If I understand correctly, you have a PDE of the form $$ (D_t - \lambda_1(t, x, D_x) + \lambda_0(t, x, D_t, D_x))u = f, $$ where $\lambda_1$ is a first order pseudodifferential operator and $\lambda_0$ is a zero-th order pseudodifferential operator. It seems to me that the proofs of many if not all estimates, including energy integral estimates, and theorems about regularity, uniqueness, and existence for the equation $$ (D_t - a(t,x)D_x + b(t,x))u = f, $$ as presented in books like Taylor can be extended to your PDE.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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