Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

According to the basic rules of symbolic caculus,$[a(x,D),x_{j}]=-ia^{j}[x,D]$.So we have $[(1-\triangle)^{\frac{1}{2}},x_i]=\partial_i(1-\triangle)^{-\frac{1}{2}}$ which is $L^2$ bounded. It's also true that the $[(1-\triangle)^{\frac{1}{2}},\langle x \rangle]$ is $L^2$ bounded.Can we write the explicit expression of this commutator? More generally,how to show that $[(1-\triangle)^{\frac{\alpha}{2}},\langle x \rangle^{\alpha}]$ ($\alpha \leq1$)is $L^2$ bounded? it is obviously true when $\alpha<0$ ?

what if we use $(-\triangle)^{\alpha}$ instead which the symbol of it is not smooth at 0?

share|improve this question
I assume that $\langle x\rangle =\sqrt{1+x^2}$. If so, here are two simple observations: 1) For $\alpha \le 0$ the operators $(1-\Delta)^{\frac{\alpha}{2}}$ and $\langle x\rangle^\alpha$ are each individually bounded on $L^2$ and so their commutator is bonded, with norm no larger than $2$. 2) For even positive $\alpha$, the commutator is an unbounded partial differential operator. For example $$ [(1-\Delta),\langle x \rangle^2]= -4 [x\cdot \nabla +\nabla\cdot x].$$ Thus seems likely that the commutator is unbounded for $\alpha \ge 2$. Not sure about $0<\alpha <2$. –  Jeff Schenker Apr 24 '12 at 1:12
note that only $\alpha< -\frac{n}{2}$ can assure boundedness.what about $0>\alpha \geq -\frac{n}{2}$ ? –  user23078 Apr 24 '12 at 7:13
No. The multiplication operator $f \mapsto \langle x \rangle^\alpha f$ is a bounded operator on $L^2$ with norm one once $\alpha \le 0$. Indeed, $$ \int |\langle x \rangle^\alpha f(x)|^2 dx \le \int |f(x)|^2 dx$$ since the function $\langle x \rangle^\alpha \le 1$. (The criterion $\alpha < -\frac{n}{2}$ indicates when the functions $\langle x \rangle^\alpha$ is itself in $L^2$, but that is not relevant to deciding if the multiplication operator is bounded.) –  Jeff Schenker Apr 24 '12 at 14:40
oops...you are right. –  user23078 Apr 24 '12 at 15:03
add comment

2 Answers

up vote 2 down vote accepted

Here is a technique that shows the commutator is bounded for $\alpha <1$. First note, as I observed in my comment, it is a bounded operator for $\alpha \le 0$. Now note that if $0<\eta<1$ and $A$ is a strictly positive operator then we have the identity $$ A^\eta = c_\eta \int_0^\infty t^\eta \left (\frac{1}{t}- \frac{1}{t+A} \right ) dt $$ where $\frac{1}{c_\eta} =\int_0^\infty t^\eta \left [\frac{1}{t}- \frac{1}{t+1} \right ]dt$ and the integral is to be understood in the strong sense (apply both sides to a vector in a dense core for the domain of $A^\eta$.) It follows that if $A$ and $B$ are two such operators then $$ [A^\eta,B^\eta]=c_\eta^2 \int_0^\infty\int_0^\infty t^\eta s^\eta \left [ \frac{1}{t+A},\frac{1}{s+B} \right]ds dt. $$ The commutator of resolvents can be computed, for $A=1-\Delta$ and $B=1+x^2$ to give $$\left [\frac{1}{t+1-\Delta},\frac{1}{s+1+x^2} \right ] = -\frac{1}{t+1 -\Delta} \left [ 1-\Delta,\frac{1}{s+1 +x^2} \right ] \frac{1}{t+1 -\Delta} $$ $$= - 2 \frac{1}{t+1-\Delta} \left ( \frac{x}{(s+1+x^2)^2}\cdot \nabla+\nabla \cdot\frac{x}{(s+1+x^2)^2} \right ) \frac{1}{t+1-\Delta}.$$ The operator norm of the result is bounded by $$ 4 \frac{1}{(t+1)^{\frac{3}{2}}} \frac{1}{(s+1)^{\frac{3}{2}}}.$$ (To see this note that $\sup_x |x|/\sqrt{s+1 +x^2}= 1$. A similar computation on the Fourier side gives $\|\nabla/\sqrt{t+1-\Delta}\|=1$.) Plugging this into the integral representation gives $$\left \| \left [ (1-\Delta)^\frac{\alpha}{2}, (1+x^2)^\frac{\alpha}{2} \right ] \right \| \le \left (2 c_\eta \int_0^\infty \frac{t^\frac{\alpha}{2}}{(t+1)^{3/2}} dt \right )^2$$ which is finite if $0 <\alpha <1$.

Clearly this argument misses something since it doesn't give the boundary case $\alpha=1$, however I feel that a modification of this argument will show the commutator to be unbounded once $1<\alpha <2$ but I don't see the details. For $\alpha \ge 2$ I believe the commutator is unbounded and one can certainly show this, as I mentioned in my comment above, if $\alpha$ is an even natural number since the result is a partial differential operator.

share|improve this answer
add comment

Take a pseudodifferential operator with a symbol $p\in S^1_{1,0}$ and $f$ a Lipschitz-continuous function. Then the commutator $$ [p(x,D),f] $$ is bounded on $L^2$. When $f$ is $C^\infty$ with bounded derivatives, the principal symbol of this commutator is the Poisson bracket -i{$p,f$}, i.e. in that case $$ c_0(x,\xi)=\frac{1}{i}\sum_{1\le j\le n}\frac{\partial p}{\partial \xi_j}\cdot \frac{\partial f}{\partial x_j},\quad c_0\in S^0_{1,0}. $$ In that case, $ [p(x,D),f]=c_0(x,D)+ r(x,D),\quad r\in S^{-1}_{1,0}. $


share|improve this answer
add comment

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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