数学杂志  2018, Vol. 38 Issue (5): 933-942   PDF    
扩展功能
加入收藏夹
复制引文信息
加入引用管理器
Email Alert
RSS
本文作者相关文章
赵健
杨丛丽
潘维烨
Fock型空间上的加权复合算子
赵健, 杨丛丽, 潘维烨    
贵州师范大学数学科学学院, 贵州 贵阳 550001
摘要:本文研究了从Fα2Fα2,m上的加权复合算子的一些性质.利用一个重要的估计,获得了有关有界性和紧性的几个等价刻画,并且给出了该算子是Fα2,m上的Hilbert-Schmidt类算子的一个充分必要条件.
关键词Fock型空间    加权复合算子    有界性    紧性    Hilbert-Schmidt类算子    
WEIGHTED COMPOSITION OPERATORS OF THE FOCK-TYPE SPACES
ZHAO Jian, YANG Cong-li, PAN Wei-ye    
School of Mathematical Science, Guizhou Normal University, Guiyang 550001, China
Abstract: In this paper, we research some characters of the weighted composition operator of the Fα2 to Fα2, m. By using a significant estimate, several equivalent conditions on the bounded and compactness of the weighted composition operator are given, and we give a sufficient and necessary condition that it is Hilbert-Schmidt operator.
Key words: Fock-type spaces     weighted composition operator     boundedness     compactness     Hilbert-Schmidt operators    
1 引言

$\mathbb{C}$表示复平面, $H(\mathbb{C})$表示在$\mathbb{C}$上的解析函数全体, $dA$表示$\mathbb{C}$上的Lebesgue面积测度, $F^{2}_{\alpha}$是由$L^{2}(\mathbb{C}, e^{-\alpha|z|^{2}}dA(z))$中的整函数所构成的空间, $0<\alpha<\infty$, 即

$ F^{2}_{\alpha}=\{f\in{H(\mathbb{C})}:\|f\|^{2}=\frac{\alpha}{\pi}\int_{\mathbb{C}}{|f(z)|^{2}e^{-\alpha|z|^{2}}dA(z)}<\infty\}. $

$\alpha=1$时, $F^{2}_{\alpha}$就是经典的Fock空间$F^{2}$, $F^{2}_{\alpha}$是一个再生核Hilbert空间, 其上的内积定义为

$ \langle{f, g}\rangle=\int_{\mathbb{C}}{f(z)\overline{g(z)}e^{-\alpha|z|^{2}}dA(z)},\;\;f, g\in{F^{2}_{\alpha}}, $

其规范正交系为

$ e_{n}(z)=\frac{(\sqrt{\alpha}z)^{n}}{\sqrt{n!}},\;\;n=0, 1, 2\cdots. $

因此, 多项式函数在$F^{2}_{\alpha}$中稠密, 其再生核函数为$K_{\omega}(z)=e^{\alpha z\overline{\omega}}, \omega, z\in{\mathbb{C}}.$并且对任意的$f\in{F^{2}_{\alpha}}$, $f(\omega)=\langle{f, K_{\omega}}\rangle, \omega, z\in{\mathbb{C}}.$直接计算可知, $K_{\omega}$的范数为$\|K_{\omega}\|=e^{\frac{\alpha|\omega|^{2}}{2}}, \omega\in{\mathbb{C}}.$因此对每个固定的$\omega, K_{\omega}\in{F^{2}_{\alpha}}$, 设$k_{\omega}$$K_{\omega}$的正规化再生核, 即$k_{\omega}=\frac{K_{\omega}}{\|K_{\omega}\|}$, 则

$ k_{\omega}(z)=e^{\alpha z\overline{\omega}-\frac{\alpha|\omega|^{2}}{2}}, z, \omega\in{\mathbb{C}}. $

在过去的几十年中, 学者们对Fock空间的研究主要是针对$F^{p}_{\alpha}$展开的, 可参见文献[1-7]等.$2012$年Cho和Zhu[8]引进了Fock-Sobolev空间$F^{p, m}$的概念, 当$p=2$时, 对任意给定的正整数$m, $Fock-Sobolev空间$F^{2, m}$$m$阶导函数属于$F^{2}$的整函数构成, $F^{2, m}$的范数为

$ \|f\|_{2, m}=|f(0)|+\cdots+|f^{m-1}(0)|+\|f^{(m)}\|, \;\;\forall f\in{F^{2, m}}, $

这里$\|\cdot\|$表示$F^{2}$的范数.

Cho和Zhu[8]证明了$f\in{F^{2, m}}$当且仅当$z^{m}f\in{F^{2}}$.因此, $F^{2, m}$的范数也可以表示为

$ \|f\|^{2}_{2, m}=\frac{1}{\pi m!}\int_{\mathbb{C}}{|z^{m}f(z)|^{2}e^{-|z|^{2}}dA(z)},\;\;\forall f\in{F^{2, m}}, $

这里的常数$\frac{1}{\pi m!}$刚好使得积分$\frac{1}{\pi m!}\int_{\mathbb{C}}{|z|^{2m}e^{-|z|^{2}}dA(z)}=1$.稍作一般化, 对任意$0<\alpha<\infty$, 仍然有$f\in F^{2, m}_\alpha$当且仅当$z^mf(z)\in F^2_\alpha$.在结果中可能会相差一个常数因子, 而在具体的证明过程中, 常数$\alpha$并没有本质作用.显然, $ F^{2, m}_\alpha$是一个Hilbert空间, 并且$ F^{2, m}_\alpha \subset F^2_\alpha$.

为了使用方便, 本文将使用记号$\| \cdot \|$$ F^2_\alpha$的范数, 使用记号$\| \cdot \| _{2, m}$$F^{2, m}_\alpha$的范数, 那么可定义Fock空间$F^{2, m}_\alpha$,

$ F^{2, m}_\alpha=\{f\in{H(\mathbb{C})}:\|f\|^{2}_{2, m}=\frac{\alpha^{m+1}}{\pi m!}\int_{\mathbb{C}}{|f(z)|^{2}e^{-\alpha|z|^{2}}|z|^{2m}dA(z)}<\infty\}. $

同样的, 这里的常数$\frac{\alpha^{m+1}}{\pi m!}$刚好使得积分$\frac{\alpha^{m+1}}{\pi m!}\displaystyle\int_{\mathbb{C}}{|z|^{2m}e^{-\alpha|z|^{2}}dA(z)}=1$.类似于文献[8], 通过计算可知, $F^{2, m}_\alpha$的规范正交系为

$ e_{n}(z)=\sqrt{\frac{m!}{(m+n)!}}(\sqrt{\alpha}z)^{n},\;\; n=0, 1, 2, \cdots. $

有关Fock-Sobolev空间的更多研究, 可参见文献[9-14].

$\varphi$$\mathbb{C}$上解析自映射, $\mu\in{H(\mathbb{C})}$, 则线性算子$C^{\mu}_{\varphi}=\mu f\circ\varphi, f\in{H(\mathbb{C})}$称为加权复合算子.有关Fock空间上加权复合算子的更多信息可参见文献[15-20, 23], 其中, 文献[19]给出了$F^{2}$空间上加权复合算子的有界性和紧性的完全刻画.文献[20]则给出了加权复合算子是$F^{2}$上的Hilbert-Schmidt算子的一个等价刻画.

$T\in{B(X)}$, $X$是Hilbert空间, 称$T$$X$上的Hilbert-Schmidt算子, 如果

$ \sum\limits_{n=0}^{\infty}\|Te_{n}\|^{2}_{X}<\infty. $

本文研究从$F^{2}_{\alpha}$$F^{2, m}_{\alpha}$上加权复合算子的有界性和紧性, 并且给出了加权复合算子是$F^{2, m}_{\alpha}$上的Hilbert-Schmidt算子的一个充分必要条件.

在本文中$M$表示一个正常数, 每次出现不一定相同.

2 主要结果及证明

为了证明本文的主要结果, 需要用到下面的一些辅助结论.

引理2.1 设$\varphi$$\mathbb{C}$上的解析自映射, $\mu\in{H(\mathbb{C})}, 0<\alpha<\infty$, 则$C^{\mu}_{\varphi}:F^{2}_{\alpha}\longrightarrow F^{2, m}_{\alpha}$为紧的充分必要条件为$C^{\mu}_{\varphi}:F^{2}_{\alpha}\longrightarrow F^{2, m}_{\alpha}$有界且对$F^{2}_{\alpha}$中任意有界并在$\mathbb{C}$的任意紧子集上一致收敛于$0$的序列$\{f_{n}\}$, $n\in{\mathbb{N}^{+}}$, 有

$ \lim\limits_{n\rightarrow\infty}\|C^{\mu}_{\varphi}f_{n}\|_{2, m}=0. $

通过使用类似于文献[21, 命题3.11]的方法, 可以证明引理2.1.

引理2.2 若序列$\{f_{n}\}$$F^{2}_{\alpha}$中弱收敛于$0$, 则$\{f_{n}\}$一致有界且在$\mathbb{C}$的任意紧子集上一致收敛于$0$.

 由于序列${\{f_{n}\}}$$F^{2}_{\alpha}$中弱收敛于$0$, 则对任意$f\in{F^{2}_{\alpha}}$, $\lim\limits_{n\rightarrow\infty}\langle{f_{n}, f}\rangle=0.$于是

$ f_{n}(z)=\langle{f_{n}, K_{z}}\rangle\rightarrow 0,\;\;n\rightarrow \infty. $

因此存在与$z$有关的常数$M_{z}>0$, 使得$|f_{n}(z)|\leq M_{z}$, 而$\mathbb{C}$是完备的, 由一致有界原理可得序列${\{f_{n}\}}$一致有界, 即存在$M>0$, 使得$M=\sup\limits_{n\geq 1}{\parallel{f_{n}}\parallel}.$$\mathbb{D}$$\mathbb{C}$中的任一紧子集, $K_{z}$$F^{2}_{\alpha}$的再生核, 可将$K_{z}$看成是$\mathbb{C}$$F^{2}_{\alpha}$上的一个映射, 且对任意$z_{0}\in{\mathbb{D}}$, 有

$ \begin{align*} \|K_{z}-K_{z_{0}}\|^{2}&=\langle{K_{z}-K_{z_{0}}, K_{z}-K_{z_{0}}}\rangle \\&=\|K_{z}\|^{2}+\|K_{z_{0}}\|^{2}-\langle{K_{z}, K_{z_{0}}}\rangle-\langle{K_{z_{0}}, K_{z}}\rangle \\&=e^{\alpha|z|^{2}}+e^{\alpha|z_{0}|^{2}}-e^{\alpha z_{0}\overline{z}}-e^{\alpha z\overline{z_{0}}}. \end{align*} $

显然, 当$z\rightarrow z_{0}$时, $\|K_{z}-K_{z_{0}}\|\rightarrow 0$, 这说明$K_{z}$为连续映射, 由于$\mathbb{D}$是紧的, 故$K_{z}$$\mathbb{D}$上一致连续.因此对任意的$\varepsilon>0$, 存在$\delta>0$, 对任意的$z', z''\in{\mathbb{D}}$, 当$|z-z_{0}|<\delta$时,

$ \|K_{z'}-K_{z''}\|<\frac{\varepsilon}{2M}. $

由于$\mathbb{D}$紧, 则$\mathbb{D}$是完全有界的, 于是, 对上述的$\delta$, 存在$z_{1}, z_{2}, \cdots, z_{q}\in{\mathbb{D}}$, 使得$\mathbb{D}\subset{\bigcup\limits_{i=1}^{q}B(z_{i}, \delta)}$, 由于

$ \lim\limits_{n\rightarrow\infty}f_{n}(z_i)=\lim\limits_{n\rightarrow\infty}\langle{f_{n}, K_{z_{i}}}\rangle=0, \;\;i=1, 2, \cdots, q. $

因此对上述的$\varepsilon$, 存在$N_{i}>0, i=1, 2, \cdots, q$, 当任意$n>N_{i}$时, $|f_{n}(z_{i})|<\frac{\varepsilon}{2}.$从而取$N=\max\{N_{1}, N_{2}, \cdots, N_{q}\}$, 对任意的$z\in{\mathbb{D}}$, 一定存在$z_{j}\in{\mathbb{D}}, 1\leq j\leq q$, 使得$|z-z_{j}|<\delta$, 当任意的$n>N$时, 有

$ \begin{align*} |f_{n}(z)|&\leq|f_{n}(z)-f_{n}(z_{j})|+|f_{n}(z_{j})| \leq\|f_{n}\|\|K_{z}-K_{z_{j}}\|+|f_{n}(z_{j})| \\&<M\cdot\frac{\varepsilon}{2M}+\frac{\varepsilon}{2} =\varepsilon. \end{align*} $

$\{f_{n}\}$$\mathbb{C}$的任一紧子集上一致收敛于$0$.

引理2.3 设$\varphi$$\mathbb{C}$上的解析自映射, $\mu\in{H(\mathbb{C})}, 0<\alpha<\infty$, 则$C^{\mu}_{\varphi}:F^{2}_{\alpha}\longrightarrow F^{2, m}_{\alpha}$为紧的充分必要条件为$C^{\mu}_{\varphi}:F^{2}_{\alpha}\longrightarrow F^{2, m}_{\alpha}$有界且对$F^{2}_{\alpha}$中任意弱收敛于$0$的序列$\{f_{n}\}$, $n\in{N^{+}}$, 有

$ \lim\limits_{n\rightarrow\infty}\|C^{\mu}_{\varphi}f_{n}\|_{2, m}=0. $

 由引理2.1和引理2.2即可得证.

引理2.4 设$k_{\omega}$$F^{2}_{\alpha}$中的正规化再生核, 则当$|\omega|\rightarrow\infty$时, $k_{\omega}$弱收敛于$0$.即对任意的$f\in{F^{2}_{\alpha}}$,

$ \lim\limits_{|\omega|\rightarrow\infty}\langle{f, k_{\omega}}\rangle=0. $

 设$p$$F^{2}_{\alpha}$中的任一多项式, 则

$ p(\omega)=\langle{p, K_{\omega}}\rangle=\|K_{\omega}\|\langle{p, k_{\omega}}\rangle, \;\;\omega\in{\mathbb{C}}. $

于是

$ \langle{p, k_{\omega}}\rangle=e^{-\frac{\alpha|\omega|^{2}}{2}}p(\omega)\rightarrow0, \;\;|\omega|\rightarrow\infty. $

由多项式在$F^{2}_{\alpha}$中稠密知, 对任意$f\in{F^{2}_{\alpha}}$, $\lim\limits_{|\omega|\rightarrow\infty}\langle{f, k_{\omega}}\rangle=0.$

下面的引理是文献[22, 推论2.8]稍作推广后的结果, 它在本文主要结果的证明中有着重要作用.当$\alpha=1$时, 就是文献[22]中的结论.

引理2.5 设$f\in{H(\mathbb{C})\cap L^{p}(\mathbb{C}, \frac{\alpha^{m+1}}{\pi m!}e^{-\alpha|z|^{2}}|z|^{2m}dA(z))}, 0<\alpha<\infty$, 则对任意$p>0, z\in{\mathbb{C}}$, 存在常数$M>0$, 使得

$ |f(z)|^{p}e^{-\alpha|z|^{2}}|z|^{2m}\leq\frac{M\alpha^{m+1}}{\pi m!} \int_{\mathbb{C}}{|f(z)e^{-\frac{\alpha|z|^{2}}{p}}z^{\frac{2m}{p}}|^{p}dA(z)}. $

使用类似于文献[22]的方法可以证明引理2.5.

下面的结论来自于文献[23].

引理2.6 设$\varphi$$\mathbb{C}$上的解析自映射, $\mu\in{H(\mathbb{C})}, 0<\alpha<\infty$, 若

$ \sup\limits_{z\in{\mathbb{C}}}|\mu(z)|^{2}e^{\alpha|\varphi(z)|^{2}-\alpha|z|^{2}}<\infty, $

则存在常数$a, b\in{\mathbb{C}}$$|a|\leq1$, 使得$\varphi(z)=az+b, z\in{\mathbb{C}}.$进一步, 若

$ \lim\limits_{|z|\rightarrow\infty}|\mu(z)|^{2}e^{\alpha|\varphi(z)|^{2}-\alpha|z|^{2}}=0, $

$\varphi(z)=az+b, z\in{\mathbb{C}}$$|a|<1$.

严格地说, 引理2.6是文献[23]中结论稍作推广后的结果, 当$\alpha=1$时, 引理2.6就是文献[23]中的结论.实际上, 上述各个引理中的常数$\alpha$在具体的证明中并没有本质作用.

接下来证明本文的主要结论.

定理2.1 设$\varphi$$\mathbb{C}$上的解析自映射, $\mu\in{H(\mathbb{C})}$$\mu\neq0, 0<\alpha<\infty$, 则下列陈述等价:

(ⅰ)$C^{\mu}_{\varphi}:F^{2}_{\alpha}\longrightarrow F^{2, m}_{\alpha}$有界;

(ⅱ)

$ \sup\limits_{\omega\in{\mathbb{C}}}\frac{\alpha^{m+1}}{\pi m!}\int_{\mathbb{C}}{|\mu(z)|^{2}|e^{\alpha\varphi(z)\overline{\omega}}|^{2}e^{-\alpha(|z|^{2}+|\omega|^{2})}|z|^{2m}dA(z)}<\infty; $ (2.1)

(ⅲ)$\mu\in{F^{2, m}_{\alpha}}$, 存在$a, b\in{\mathbb{C}}, |a|\leq1$, 使得$\varphi(z)=az+b, z\in{\mathbb{C}}$

$ \sup\limits_{z\in\mathbb{C}}|\mu(z)|^{2}e^{\alpha|\varphi(z)|^{2}-\alpha|z|^{2}}|z|^{2m}<\infty. $ (2.2)

 (ⅰ)$\Rightarrow$(ⅱ).设$C^{\mu}_{\varphi}:F^{2}_{\alpha}\longrightarrow F^{2, m}_{\alpha}$有界, $k_{\omega}$$F^{2}_{\alpha}$中的正规化再生核, 显然, $k_{\omega}\in{F^{2}_{\alpha}}, \omega\in{\mathbb{C}}$, 因此

$ \begin{align*} \|C^{\mu}_{\varphi}k_{\omega}\|^{2}_{2, m}&=\frac{\alpha^{m+1}}{\pi m!}\int_{\mathbb{C}}{|C^{\mu}_{\varphi}k_{\omega}(z)|^{2}|e^{-\alpha|z|^{2}}|z|^{2m}dA(z)} \\&=\frac{\alpha^{m+1}}{\pi m!}\int_{\mathbb{C}}{|\mu(z)|^{2}|e^{\alpha\varphi(z)\overline{\omega}}|^{2}e^{-\alpha(|z|^{2}+|\omega|^{2})}|z|^{2m}dA(z)} \\&\leq\|C^{\mu}_{\varphi}\|^{2}\|k_{\omega}\|^{2}<\infty. \end{align*} $

从而

$ \sup\limits_{\omega\in{\mathbb{C}}}\frac{\alpha^{m+1}}{\pi m!}\int_{\mathbb{C}}{|\mu(z)|^{2}|e^{\alpha\varphi(z)\overline{\omega}}|^{2}e^{-\alpha(|z|^{2}+|\omega|^{2})}|z|^{2m}dA(z)}<\infty. $

即(2.1)式成立.

(ⅱ)$\Rightarrow$(ⅲ).若$(2.1)$式成立, 则对任意$\omega\in{\mathbb{C}}$,

$ \frac{\alpha^{m+1}}{\pi m!}\int_{\mathbb{C}}{|\mu(z)|^{2}|e^{\alpha\varphi(z)\overline{\omega}}|^{2}e^{-\alpha(|z|^{2}+|\omega|^{2})}|z|^{2m}dA(z)}<\infty, $

$ \frac{\alpha^{m+1}}{\pi m!}\int_{\mathbb{C}}{|C^{\mu}_{\varphi}k_{\omega}(z)|^{2}e^{-\alpha|z|^{2}}|z|^{2m}dA(z)}<\infty. $

这说明$C^{\mu}_{\varphi}k_{\omega}\in{F^{2, m}_{\alpha}}$, 由引理$2.5$, 存在$M>0$, 使得

$ \begin{align*} |C^{\mu}_{\varphi}k_{\omega}(z)|^{2}e^{-\alpha|z|^{2}}|z|^{2m}&<\frac{M\alpha^{m+1}}{\pi m!}\int_{\mathbb{C}}{|C^{\mu}_{\varphi}k_{\omega}(z)|^{2}e^{-\alpha|z|^{2}}|z|^{2m}dA(z)}<\infty, \end{align*} $

$ |\mu(z)|^{2}|e^{\alpha\varphi(z)\overline{\omega}-\frac{\alpha|\omega|^{2}}{2}}|^{2}e^{-\alpha|z|^{2}}|z|^{2m}<\infty,\;\; \omega\in{\mathbb{C}}. $

特别地, 取$\omega=\varphi(z)$, 则

$ |\mu(z)|^{2}e^{\alpha|\varphi(z)|^{2}-\alpha|z|^{2}}|z|^{2m}=|\mu(z)|^{2}|e^{\alpha\varphi(z)\overline{\omega}-\frac{\alpha|\omega|^{2}}{2}}|^{2}e^{-\alpha|z|^{2}}|z|^{2m}<\infty, \;\;z\in{\mathbb{C}}. $

从而(2.2)式成立, 由(2.2)式容易得到$\sup\limits_{z\in\mathbb{C}}|\mu(z)|^{2}e^{\alpha|\varphi(z)|^{2}-\alpha|z|^{2}}<\infty.$由引理2.6, 存在常数$a, b\in{\mathbb{C}}$, 且$|a|\leq1$, 使得$\varphi(z)=az+b, z\in{\mathbb{C}}.$在(2.1)式中, 令$\omega=0$, 即可得

$ \frac{\alpha^{m+1}}{\pi m!}\int_{\mathbb{C}}{|\mu(z)|^{2}e^{-\alpha|z|^{2}}|z|^{2m}dA(z)}<\infty. $

这表明$\mu\in{F^{2, m}_{\alpha}}$.

(ⅲ)$\Rightarrow$(ⅰ).若存在常数$a, b\in{\mathbb{C}}, |a|\leq1$, 使得$\varphi(z)=az+b, z\in{\mathbb{C}}.$$\omega=\varphi(z)$, 则当$0<|a|\leq1$时, $z=\frac{\omega}{a}-\frac{b}{a}, \omega\in{\mathbb{C}}.$由(2.2)式, 存在$M>0$, 使得

$ M=\sup\limits_{z\in{\mathbb{C}}}|\mu(z)|^{2}e^{\alpha|\varphi(z)|^{2}-\alpha|z|^{2}}|z|^{2m}. $

于是对任意的$f\in{F^{2}_{\alpha}}$,

$ \begin{align*} \|C^{\mu}_{\varphi}f\|^{2}_{2, m}&=\frac{\alpha^{m+1}}{\pi m!}\int_{\mathbb{C}}{|\mu(z)|^{2}|f(\varphi(z))|^{2}e^{-\alpha|z|^{2}}|z|^{2m}dA(z)} \\&=\frac{\alpha^{m+1}}{\pi m!}\int_{\mathbb{C}}{|\mu(z)|^{2}e^{\alpha|\varphi(z)|^{2}-\alpha|z|^{2}}|z|^{2m}|f(\varphi(z))|^{2}e^{-\alpha|\varphi(z)|^{2}}dA(z)} \\&\leq\frac{M\alpha^{m+1}}{\pi m!}\int_{\mathbb{C}}{|f(\varphi(z))|^{2}e^{-\alpha|\varphi(z)|^{2}}dA(z)} \\&=\frac{M\alpha^{m+1}}{\pi m!}\int_{\mathbb{C}}{|f(\omega)|^{2}e^{-\alpha|\omega|^{2}}dA(\frac{\omega}{a}-\frac{b}{a})} \\&=\frac{M\alpha^{m}}{|a|^{2}m!}\|f\|^{2}<\infty. \end{align*} $

这说明$C^{\mu}_{\varphi}:F^{2}_{\alpha}\longrightarrow F^{2, m}_{\alpha}$有界, 并且这时有

$ \|C^{\mu}_{\varphi}\|^{2}\leq\frac{M\alpha^{m}}{|a|^{2}m!}. $

$|a|=0$时, $\varphi(z)=b$, 于是对任意的$f\in{F^{2}_{\alpha}}, C^{\mu}_{\varphi}f=\mu f(b)$.因此

$ \|C^{\mu}_{\varphi}f\|^{2}_{2, m}=|f(b)|^{2}\|\mu\|^{2}_{2, m}, $

$\mu\in{F^{2, m}_{\alpha}}$易知$C^{\mu}_{\varphi}:F^{2}_{\alpha}\longrightarrow F^{2, m}_{\alpha}$有界.

定理2.2 设$\varphi$$\mathbb{C}$上的解析自映射, $\mu\in{H(\mathbb{C})}$$\mu\neq0, 0<\alpha<\infty$, 则下列陈述等价:

(ⅰ)$C^{\mu}_{\varphi}:F^{2}_{\alpha}\longrightarrow F^{2, m}_{\alpha}$是紧的;

(ⅱ)$C^{\mu}_{\varphi}:F^{2}_{\alpha}\longrightarrow F^{2, m}_{\alpha}$有界, 且

$ \lim\limits_{|\omega|\rightarrow\infty}\frac{\alpha^{m+1}}{\pi m!}\int_{\mathbb{C}}{|\mu(z)|^{2}|e^{\alpha\varphi(z)\overline{\omega}}|^{2}e^{-\alpha(|z|^{2}+|\omega|^{2})}|z|^{2m}dA(z)}=0; $ (2.3)

(ⅲ)$\mu\in{F^{2, m}_{\alpha}}$, 存在常数$a, b\in{\mathbb{C}}, |a|<1$, 使得$\varphi(z)=az+b, z\in{\mathbb{C}}$

$ \lim\limits_{|z|\rightarrow\infty}|\mu(z)|^{2}e^{\alpha|\varphi(z)|^{2}-\alpha|z|^{2}}|z|^{2m}=0. $ (2.4)

 (ⅰ)$\Rightarrow$(ⅱ).设$C^{\mu}_{\varphi}:F^{2}_{\alpha}\longrightarrow F^{2, m}_{\alpha}$是紧的, 由引理2.3, 对于$F^{2}_{\alpha}$中任意弱收敛于$0$的点列$\{f_{n}\}$, 有

$ \lim\limits_{n\rightarrow\infty}\|C^{\mu}_{\varphi}f_{n}\|_{2, m}=0. $

$k_{\omega}$$F^{2}_{\alpha}$中的正规化再生核, $\{\omega_{n}\}$$\mathbb{C}$中任意趋于无穷的点列, 由引理2.4, 当$n\rightarrow\infty$时, $\{k_{\omega_{n}}\}$弱收敛于$0$.于是

$ \lim\limits_{n\rightarrow\infty}\|C^{\mu}_{\varphi}k_{\omega_{n}}\|^{2}_{2, m}=0, $

从而

$ \lim\limits_{|\omega|\rightarrow\infty}\frac{\alpha^{m+1}}{\pi m!}\int_{\mathbb{C}}{|\mu(z)|^{2}|e^{\alpha\varphi(z)\overline{\omega}}|^{2}e^{-\alpha(|z|^{2}+|\omega|^{2})}|z|^{2m}dA(z)}=0. $

即(2.3)式成立.$C^{\mu}_{\varphi}:F^{2}_{\alpha}\longrightarrow F^{2, m}_{\alpha}$有界是显然的.

(ⅱ)$\Rightarrow$(ⅲ).设$C^{\mu}_{\varphi}:F^{2}_{\alpha}\longrightarrow F^{2, m}_{\alpha}$有界, 由定理2.1知$\mu\in{F^{2, m}_{\alpha}}$, 且存在常数$a, b\in{\mathbb{C}}, |a|\leq1$, 使得$\varphi(z)=az+b, z\in{\mathbb{C}}.$由于$C^{\mu}_{\varphi}k_{\omega}\in{F^{2, m}_{\alpha}}$, 由引理2.5, 存在$M>0$, 使得

$ \begin{align*} |C^{\mu}_{\varphi}k_{\omega}(z)|^{2}e^{-\alpha|z|^{2}}|z|^{2m}&<\frac{M\alpha^{m+1}}{\pi m!}\int_{\mathbb{C}}{|C^{\mu}_{\varphi}k_{\omega}(z)|^{2}e^{-\alpha|z|^{2}}|z|^{2m}dA(z)} \\&=\frac{M\alpha^{m+1}}{\pi m!}\int_{\mathbb{C}}{|\mu(z)|^{2}|e^{\alpha\varphi(z)\overline{\omega}}|^{2}e^{-\alpha(|z|^{2}+|\omega|^{2})}|z|^{2m}dA(z)}, \omega\in{\mathbb{C}}. \end{align*} $

由(2.3)式可得

$ \lim\limits_{|\omega|\rightarrow\infty}|\mu(z)|^{2}|e^{\alpha\varphi(z)\overline{\omega}-\frac{\alpha|\omega|^{2}}{2}}|^{2}e^{-\alpha|z|^{2}}|z|^{2m}=\lim\limits_{|\omega|\rightarrow\infty}|C^{\mu}_{\varphi}k_{\omega}(z)|^{2}e^{-\alpha|z|^{2}}|z|^{2m}=0. $

由上式, 任意的$\varepsilon>0$, 存在$R_{\varepsilon}>0$, 当$|\omega|>R_{\varepsilon}$时,

$ |\mu(z)|^{2}|e^{\alpha\varphi(z)\overline{\omega}-\frac{\alpha|\omega|^{2}}{2}}|^{2}e^{-\alpha|z|^{2}}|z|^{2m}<\varepsilon. $

$0<|a|\leq1$时, 令$\omega=\varphi(z)$, 则$|\omega|=|\varphi(z)|\rightarrow\infty, |z|\rightarrow\infty$, 因此, 对上述的$R_{\varepsilon}$, 存在$R'>0$, 当$|z|>R'$时, 有$|\omega|=|\varphi(z)|>R_{\varepsilon}.$于是对上述的$\varepsilon$, 取$R=R'$, 当$|z|>R$时, $|\omega|=|\varphi(z)|>R_{\varepsilon}$

$ |\mu(z)|^{2}e^{\alpha|\varphi(z)|^{2}-\alpha|z|^{2}}|z|^{2m}=|\mu(z)|^{2}|e^{\alpha\varphi(z)\overline{\omega}-\frac{\alpha|\omega|^{2}}{2}}|^{2}e^{-\alpha|z|^{2}}|z|^{2m}<\varepsilon. $

$\lim\limits_{|z|\rightarrow\infty}|\mu(z)|^{2}e^{\alpha|\varphi(z)|^{2}-\alpha|z|^{2}}|z|^{2m}=0.$因此这时(2.4)式成立.

$|a|=0$时, $\varphi(z)=b$, 由上述情形易知, 这时(2.4)式仍然成立.

由(2.4)式容易得到

$ \lim\limits_{|z|\rightarrow\infty}|\mu(z)|^{2}e^{\alpha|\varphi(z)|^{2}-\alpha|z|^{2}}=0. $

再由引理2.6知$|a|<1$.

(ⅲ)$\Rightarrow$(ⅰ).设$\{f_{n}\}$$F^{2}_{\alpha}$中任意有界且在$\mathbb{C}$的任一紧子集上一致收敛于$0$的序列, 则存在$C_{1}>0$, 使得$\|f_{n}\|\leq C_{1}$.由于$\mu\in{F^{2, m}_{\alpha}}$, 则存在$C_{2}>0$, 使得$\|\mu\|_{2, m}\leq C_{2}$.若存在常数$a, b\in{\mathbb{C}}, |a|<1$, 使得$\varphi(z)=az+b, z\in{\mathbb{C}}.$$\omega=\varphi(z)$, 当$0<|a|<1$时,

$ z=\frac{1}{a}\omega-\frac{b}{a}, \;\;\omega\in{\mathbb{C}}. $

由(2.4)式, 对任意的$\varepsilon>0$, 存在$R>0$, 当$|z|>R$时,

$ |\mu(z)|^{2}e^{\alpha|\varphi(z)|^{2}-\alpha|z|^{2}}|z|^{2m}<\frac{L}{2C_{1}}\varepsilon, $

这里$L=\frac{|a|^{2}m!}{\alpha^{m}}$.记$\mathbb{D}=\{z\in{\mathbb{C}}:|z|\leq R\}$, 则$\mathbb{D}$$\mathbb{C}$中的紧子集, 由于$\varphi$$\mathbb{C}$上的解析自映射, 则$\varphi$连续, 因此$\varphi(\mathbb{D})$亦为$\mathbb{C}$中的紧子集.从而, 对上述的$\varepsilon$, 存在$N_{\varepsilon}>0$, 当$n>N_{\varepsilon}$时, 对任意的$z\in{\mathbb{D}}$, 有$\varphi(z)\in{\varphi(\mathbb{D})}$, 进而

$ |f_{n}(\varphi(z))|^{2}<\frac{\varepsilon}{2C^{2}_{2}}, $

那么取$N=N_{\varepsilon}$, 当$n>N$时, 有

$ \begin{align*} \|C^{\mu}_{\varphi}f_{n}\|^{2}_{2, m}=&\frac{\alpha^{m+1}}{\pi m!}\int_{\mathbb{C}}{|\mu(z)|^{2}|f_{n}(\varphi(z))|^{2}e^{-\alpha|z|^{2}}|z|^{2m}dA(z)} \\=&\frac{\alpha^{m+1}}{\pi m!}\int_{\mathbb{D}}{|\mu(z)|^{2}|f_{n}(\varphi(z))|^{2}e^{-\alpha|z|^{2}}|z|^{2m}dA(z)} \\&+\frac{\alpha^{m+1}}{\pi m!}\int_{\mathbb{C}\setminus\mathbb{D}}{|\mu(z)|^{2}e^{\alpha|\varphi(z)|^{2}-\alpha|z|^{2}}|z|^{2m}|f_{n}(\varphi(z))|^{2}e^{-\alpha|\varphi(z)|^{2}}dA(z)} \\<&\frac{\varepsilon}{2C^{2}_{2}}\|\mu\|^{2}_{2, m}+\frac{L}{2C^{2}_{1}}\varepsilon\cdot\frac{\alpha^{m+1}}{|a|^{2}\pi m!}\int_{\mathbb{C}\setminus \mathbb{D}}{|f_{n}(\omega)|^{2}e^{-\alpha|\omega|^{2}}dA(\omega)} \\=&\frac{\varepsilon}{2C^{2}_{2}}\|\mu\|^{2}_{2, m}+\frac{L}{2C^{2}_{1}}\varepsilon\cdot\frac{\alpha^{m}}{|a|^{2}m!}\|f_{n}\|^{2} \\\leq&\frac{\varepsilon}{2C^{2}_{2}}C^{2}_{2}+\frac{L}{2C^{2}_{1}}\varepsilon\cdot\frac{1}{L}C^{2}_{1}=\varepsilon. \end{align*} $

$\lim\limits_{n\rightarrow\infty}\|C^{\mu}_{\varphi}f_{n}\|_{2, m}=0.$

$|a|=0$时, $\varphi(z)=b$, 此时有

$ \lim\limits_{n\rightarrow\infty}\|C^{\mu}_{\varphi}f_{n}\|_{2, m}=\|\mu\|_{2, m}\lim\limits_{n\rightarrow\infty}|f_{n}(b)|=0. $

综上, 当$|a|<1$时, 由引理2.1可得, $C^{\mu}_{\varphi}:F^{2}_{\alpha}\longrightarrow F^{2, m}_{\alpha}$是紧的.

接下来讨论Hilbert空间上的一类重要算子-Hilbert-Schmidt算子.文献[8]给出了Fock空间$F^{2}$上的加权复合算子是Hilbert-Schmidt算子的完全刻画, 其证明过程可以平行地推广到$\mathbb{C}^{n}$上的Fock空间.下面给出$F^{2, m}_{\alpha}$空间上的加权复合算子$C^{\mu}_{\varphi}$是Hilbert-Schmidt算子的一个充分必要条件.

定理2.3 设$\varphi$$\mathbb{C}$上的解析自映射, $\mu\in{H(\mathbb{C})}, 0<\alpha<\infty$, 若$C^{\mu}_{\varphi}:F^{2, m}_{\alpha}\longrightarrow F^{2, m}_{\alpha}$有界, 则$C^{\mu}_{\varphi}$$F^{2, m}_{\alpha}$上的Hilbert-Schmidt算子的充分必要条件为

$ \int_{\mathbb{C}}{\frac{|\mu(z)|^{2}e^{\alpha|\varphi(z)|^{2}-\alpha|z|^{2}}|z|^{2m}}{|\varphi(z)|^{2m}}dA(z)}<\infty. $ (2.5)

 设(2.5)式成立, 记$e_{n}$$F^{2, m}_{\alpha}$的规范正交系, 则

$ e_{n}(z)=\sqrt{\frac{m!}{(m+n)!}}(\sqrt{\alpha}z)^{n}, z\in{\mathbb{C}}, n=0, 1, 2\cdots. $

从而

$ \begin{align*} \sum^{\infty}_{n=0}\|C^{\mu}_{\varphi}e_{n}\|^{2}_{2, m}=&\sum^{\infty}_{n=0}\frac{\alpha^{m+1}}{\pi m!}\int_{\mathbb{C}}{|\mu(z)|^{2}|e_{n}(\varphi(z))|^{2}e^{-\alpha|z|^{2}}|z|^{2m}dA(z)} \\=&\sum^{\infty}_{n=0}\frac{\alpha^{m+1}}{\pi m!}\int_{\mathbb{C}}{|\mu(z)|^{2}\frac{m!}{(m+n)!}\alpha^{n}|\varphi(z)|^{2n}e^{-\alpha|z|^{2}}|z|^{2m}dA(z)} \\=&\frac{\alpha}{\pi}\int_{\mathbb{C}}{\frac{|\mu(z)|^{2}e^{-\alpha|z|^{2}}|z|^{2m}}{|\varphi(z)|^{2m}}\sum^{\infty}_{n=0}\frac{(\alpha|\varphi(z)|^{2})^{m+n}}{(m+n)!}dA(z)} \\=&\frac{\alpha}{\pi}\int_{\mathbb{C}}{\frac{|\mu(z)|^{2}e^{\alpha|\varphi(z)|^{2}-\alpha|z|^{2}}|z|^{2m}}{|\varphi(z)|^{2m}}dA(z)} \\&-\sum^{m-1}_{k=0}\frac{\alpha^{m+1}}{\pi}\int_{\mathbb{C}}{|\mu(z)|^{2}\frac{(\alpha|\varphi(z)|^{2})^{k-m}}{k!}e^{-\alpha|z|^{2}}|z|^{2m}dA(z)} \\<&\frac{\alpha}{\pi}\int_{\mathbb{C}}{\frac{|\mu(z)|^{2}e^{\alpha|\varphi(z)|^{2}-\alpha|z|^{2}}|z|^{2m}}{|\varphi(z)|^{2m}}dA(z)}<\infty. \end{align*} $

因此$C^{\mu}_{\varphi}$$F^{2, m}_{\alpha}$上的Hilbert-Schmidt算子.

反过来, 设$C^{\mu}_{\varphi}$$F^{2, m}_{\alpha}$上的Hilbert-Schmidt算子, 则

$ \sum\limits_{n = 0}^\infty {\left\| {C_\varphi ^\mu {e_n}} \right\|_{2,m}^2} < \infty . $

由上述证明过程有

$ \begin{align*} &\frac{\alpha}{\pi}\int_{\mathbb{C}}{\frac{|\mu(z)|^{2}e^{\alpha|\varphi(z)|^{2}-\alpha|z|^{2}}|z|^{2m}}{|\varphi(z)|^{2m}}dA(z)} \\=&\sum^{\infty}_{n=0}\|C^{\mu}_{\varphi}e_{n}\|^{2}_{2, m}+\sum^{m-1}_{k=0}\frac{\alpha^{m+1}}{\pi}\int_{\mathbb{C}}{|\mu(z)|^{2}\frac{(\alpha|\varphi(z)|^{2})^{k-m}}{k!}e^{-\alpha|z|^{2}}|z|^{2m}dA(z)}. \end{align*} $

显然, 要得到(2.5)式, 只需上式等号右边第二项有界即可.为此, 令

$ g_{k}(z)=\frac{(\sqrt{\alpha}z)^{k-m}}{\sqrt{k!}}, \;\;z\in{\mathbb{C}},\;\; k=0, 1, \cdots, m-1. $

对每个$k$, 由于$z^{m}g_{k}(z)=\frac{(\sqrt{\alpha}z)^{k}}{\sqrt{k!}}$$F^{2}_{\alpha}$规范正交系的前$m-1$项, 则$z^{m}g_{k}\in{F^{2}_{\alpha}}$, 于是$g_{k}\in{F^{2, m}_{\alpha}}$, 而$C^{\mu}_{\varphi}:F^{2, m}_{\alpha}\longrightarrow F^{2, m}_{\alpha}$有界, 因此, 存在$M>0$, 使得

$ \begin{align*} &\frac{\alpha^{m+1}}{\pi}\int_{\mathbb{C}}{|\mu(z)|^{2}\frac{(\alpha|\varphi(z)|^{2})^{k-m}}{k!}e^{-\alpha|z|^{2}}|z|^{2m}dA(z)} \\=&m!\cdot\frac{\alpha^{m+1}}{\pi m!}\int_{\mathbb{C}}{|\mu(z)|^{2}|g_{k}(\varphi(z))|^{2}e^{-\alpha|z|^{2}}|z|^{2m}dA(z)} \\=&m!\cdot\|C^{\mu}_{\varphi}g_{k}\|^{2}_{2, m}\leq m!\cdot M<\infty. \end{align*} $

进而有

$ \sum\limits_{k = 0}^{m - 1} {\frac{{{\alpha ^{m + 1}}}}{\pi }} \int_\mathbb{C} {|\mu (z){|^2}\frac{{{{(\alpha |\varphi (z){|^2})}^{k - m}}}}{{k!}}{e^{ - \alpha |z{|^2}}}|z{|^{2m}}dA(z)} < \infty . $

故(2.5)式成立.

参考文献
[1] Bargmann V. On a Hilbert space of analytic functions and an associated integral tranform[J]. Comm. Pure Appl. Math., 1961, 14: 187–214. DOI:10.1002/(ISSN)1097-0312
[2] Furdui O. On a class of integral operators[J]. Integ. Equ. Oper. The., 2010, 66: 469–483.
[3] Isralowitz J, Zhu K H. Toeplitz operators on the Fock space[J]. Integ. Equ. Oper. The., 2010, 66: 593–611. DOI:10.1007/s00020-010-1768-9
[4] Perälä A, Schuster A, Virtanen J A. Hankel operators on Fock spaces[J]. Oper The. Adv. Appl., 2014, 236: 377–390.
[5] Bauer W. Mean oscillation and Hankel operators on the Segal-Bergmann space[J]. Integ. Equ. Oper. The., 2005, 52: 1–15. DOI:10.1007/s00020-003-1272-6
[6] Gryc W E, Kemp T. Duality in Segal-Bergmann spaces[J]. J. Funct. Anal., 2011, 261: 1591–1623. DOI:10.1016/j.jfa.2011.05.014
[7] Bauer W, Isralowitz J. Compactness characterization of operators in the Toeplitz algebra of the Fock space Fαp[J]. J. Funct. Anal., 2012, 263: 1323–1355. DOI:10.1016/j.jfa.2012.04.020
[8] Cho H R, Zhu K. Fock-Sobolev spaces and their Carleson measures[J]. ArXiv:1212. 0737, 2012.
[9] Cho H R, Choe B R, Koo H. Fock-Sobolev spaces of fractional order[J]. Poten. Anal., 2015, 43: 199–240. DOI:10.1007/s11118-015-9468-3
[10] Cho H R, Choe B R, Koo H. Combinations of composition operators on the Fock-Sobolev space[J]. Poten. Anal., 2014, 41: 1223–1246. DOI:10.1007/s11118-014-9417-6
[11] Choe B R, Yang J. Commutants of Toeplitz operators with radial symbols on the Fock-Sobolev space[J]. J. Math. Anal. Appl., 2014, 415: 779–790. DOI:10.1016/j.jmaa.2014.02.018
[12] Mengestie T. Carleson type measures for Fock-Sobolev spaces[J]. Compl. Anal. Oper. The., 2014, 8: 1225–1256. DOI:10.1007/s11785-013-0321-7
[13] Wang X F, Cao G F, Xia J. Toeplitz operators on Fock-Sobolev spaces with positive measure symbols[J]. Sci. China Math., 2014, 57: 1443–1462. DOI:10.1007/s11425-014-4813-3
[14] Wang X F, Xia J, Cao G F. Bounded, compact and Sp-class operators on Fock-Sobolev spaces (in Chinese)[J]. Sci. Sin. Math., 2014, 44: 263–274. DOI:10.1360/012014-15
[15] Zhao L. Invertible weighted composition operators on the Fock space of ${\mathbb{C}^n}$[J]. J. Funct. Spaces, 2015, ID250358.
[16] Stevi S. Weighted composition operators between Fock-type spaces in ${\mathbb{C}^n}$[J]. Appl. Math. Comp., 2009, 215: 2750–2760. DOI:10.1016/j.amc.2009.09.016
[17] Zhao L. Unitary Weighted composition operators on Fock space of ${\mathbb{C}^n}$[J]. Compl. Anal. Oper. The., 2014, 8: 581–590. DOI:10.1007/s11785-013-0313-7
[18] Zhao L, Pang C. A class of weighted composition operators on Fock space[J]. J. Math. Res. Appl., 2015, 35(3): 303–310.
[19] Ueki S. Weighted composition operator on the Fock space[J]. Proc. Amer. Math. Soc., 2007, 135(5): 1405–1410. DOI:10.1090/S0002-9939-06-08605-9
[20] Ueki S. Hilbert-Schmidt weighted composition operator on the Fock space[J]. Int. J. Math. Anal., 2007, 1: 769–774.
[21] Cowen C C, Maccluer B D. Composition operators on spaces of analytic functions[M]. Stud. Adv. Math., Florida: CRC Press, 1995.
[22] Zhu K H. Analysis on Fock space[M]. New York: Springer-Verlag, 2012.
[23] Le T. Normal and isometric weighted composition operators on the Fock space[J]. Bull. London Math. Soc., 2014, 46: 847–856. DOI:10.1112/blms/bdu046