数学杂志  2017, Vol. 37 Issue (6): 1317-1327   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
李增明
黄莉茸
两粒子的Vlasov-Maxwell-Boltzmann方程组中一个算子的紧性证明
李增明, 黄莉茸    
暨南大学信息科学技术学院数学系, 广东 广州 510632
摘要:本文主要研究紧算子在Vlasov-Maxwell-Boltzmann方程组中的应用的问题.利用紧算子的定义,获得了描述不同质量两粒子模型的线性Boltzmann算子的一个紧性结果.
关键词赋范空间    紧算子    Vlasov-Maxwell-Boltzmann方程组    Boltzmann碰撞算子    
A PROOF OF COMPACT OPERATOR OF THE TWO SPECIES VLASOV-MAXWELL-BOLTZMANN EQUATIONS
LI Zeng-ming, HUANG Li-rong    
Department of Mathematics, School of Information Science and Technology, Jinan University, Guangzhou 510632, China
Abstract: In this paper, we mainly investigate an application of compact operators to the Vlasov-Maxwell-Boltzmann equations. By using the definition of compact operators, we prove a compactness result of the linear Boltzmann operator of a two species model with different mass.
Key words: normed space     compact operator     Vlasov-Maxwell-Boltzmann equations     Boltzmann collision operator    
1 引言

本文研究带不同质量两粒子线性Boltzmann算子的性质.这类线性算子在Vlasov-Maxwell-Boltzmann模型中有着重要的应用. Vlasov-Maxwell-Boltzmann方程组是分子动理学理论中一类常见的方程组, 可以用来描述弱电离化的等离子体中各种粒子比如说离子和电子的动力学行为[19].对于两族的不考虑物理参数如粒子质量、电量等的Vlasov-Maxwell-Boltzmann方程组的经典解的存在性以及大时间行为的研究已经取得了很多重要的成果, 见Guo[16]、Strain[24]、Duan-Strain[13]、Duan-Lei-Yang-Zhao[12]、Lei-Zhao[23]、Ha-Xiao-Xiong-Zhao[17]等在碰撞核截断情形的工作, 以及Duan-Liu-Yang-Zhao[11]、Fang-Lei[15]等在碰撞核非截断情形的工作.我们知道带正电的离子和带负电的电子其质量一般相差较大, 因此从物理角度来讲, Vlasov-Maxwell-Boltzmann方程组应该考虑两种粒子的质量的影响, 注意到Duan-Liu[10]以及Huang-Liu[18]近期关于带不同质量的Vlasov-Poisson-Boltzmann方程组的非平凡解稳定性的研究工作, 证实了两种粒子质量不同, 所遇到的困难会更大.因此对带不同质量两粒子线性Boltzmann算子的研究具有深刻的意义.此外, 还注意到Boudin-Grec-Pravic-Salvarani[8]最近研究了多族的在整体Maxwellian附近的线性Boltzmann算子的紧性性质, 本文讨论来源于Vlasov-Maxwell-Boltzmann方程组由两流体分解得到的在局部Maxwellian附近的线性Boltzmann算子的性质.

2 紧算子

在这里先对紧算子的相关定义、性质及其一些判别准则做一个整理和总结.紧算子的相关等价定义

定义2.1 [1]$X$$Y$是赋范线性空间, 若算子$T: X \rightarrow Y$$X$中任何有界集映成$Y$中的列紧集, 则称$T$是紧算子.如果紧算子$T$还是线性算子, 则称$T$为紧线性算子.

定义2.2 [2]$X$, $Y$是Banach空间, 设$T: X \rightarrow Y$线性; 称$T$是紧算子, 如果$\overline{T(B_{1})}$$Y$中是紧集, 其中$B_{1}$$X$中的单位球.

定义2.3 [3]$T$是赋范空间$X$上到赋范空间$Y$中的线性算子, 如果对$X$中任意有界集$M$, $\overline{TM}$$Y$中紧集, 称$T$是紧算子.

定义2.4 [4]$T$: $D \subset X \rightarrow Y$是一连续映射, 若对任意有界序列$\{x_{n}\}\subset D$, $Tx_{n}$恒有收敛子列, 则称$T$为紧算子或紧映射.

由定义不难证明, 紧算子的下列性质

引理2.5 [1]$T: X \rightarrow Y$是线性算子, 则下述条件相互等价

(1) $T$是紧算子;

(2) $T$$X$中的单位球映成$Y$中的列紧集;

(3) 对$X$中任何有界点列$\{x_{k}\}$, $\{Tx_{k}\}$存在收敛子列.

引理2.6 [2]$X$, $Y$是赋范空间, $T\in B(X, Y)$, $B(X, Y)$$X$$Y$的有界线性算子空间, 如果$T$是紧算子, 则$T$$X$中的弱收敛点列映为$Y$中的强收敛点列.

下面给出三个判别紧算子的准则和方法.

引理2.7 [3]$\{T_{n}\}$是赋范空间$X$上到Banach空间$Y$中的紧算子列且按范数收敛于算子$T$, 则$T$也是紧算子.

引理2.8 [4]$\Omega\in\mathbb{R}^n$是一个可测集, 又设$K\in L^2(\Omega\times\Omega)$, 则

$ T: f(u)\mapsto\int_{\Omega}K(u, v)f(u)du, \ (\forall f\in L^2(\Omega)) $

$L^2(\Omega)$上的紧算子.

引理2.9 [5] 假设$K(u, v) \geq 0$$u, v$的函数, 设$Tf(\upsilon) = \displaystyle{\int} K(u, \upsilon)f(u)du, $ $K_{n} = K(u, v)\chi_{\Omega_n}$, 这里$\Omega_{n} = A_{n} \bigcap B_{n}$, 其中$A_{n} = \{(u, \upsilon): |u - \upsilon| \geq \frac{1}{n}\}, \ \ B_{n} = \{(u, \upsilon): |\upsilon|\leq n\}$, $n$为正整数.若

(1) $\displaystyle{\int} K(u, v)du$关于$v$是有界的;

(2) 对所有的$n$, $K \in L^{2}(\Omega_{n})$;

(3) $\sup \limits_{v} \displaystyle{\int} (K - K_{n})du \rightarrow 0, \ n \rightarrow \infty$,

$T$$L^{2}$上是紧的.引理的证明在参考文献[9]中已给出, 这里证明略去.

3 Vlasov-Maxwell-Boltzmann方程组K算子的紧性证明

在弱电离化的等离子体中, 带正电的离子和带负电的电子的动力学行为可以用如下Vlasov-Maxwell-Boltzmann方程组来刻画(见文献[16, 19])

$ \begin{equation}\label{equ-1.1} \left\{ \begin{array}{lll} \begin{split} &\partial_{t} F_{i} + \xi\cdot \nabla_{x} F_{i} + \frac{Ze}{m_i}(E + \xi\times B) \cdot \nabla_{\xi} F_{i} = Q_{ii}(F_{i}, F_{i}) + Q_{ie}(F_{i}, F_{e}), \\ &\partial_{t} F_{e} + \xi\cdot \nabla_{x} F_{e} - \frac{e}{m_e}(E + \xi \times B) \cdot \nabla_{\xi} F_{e} = Q_{ee}(F_{e}, F_{e}) + Q_{ei}(F_{e}, F_{i}), \end{split} \end{array} \right. \end{equation} $ (3.1)

其中电场$E$和磁场$B$满足Maxwell方程组

$ \begin{equation}\label{M} \left\{ \begin{split} &\partial_t E-\nabla_x\times B=-\int_{\mathbb{R}^3} \xi (ZeF_i-eF_e)\, d\xi, \\ &\partial_t B+\nabla_x\times E =0, \\ &\nabla_x\cdot E =\int_{\mathbb{R}^3} (ZeF_i-eF_e)\, d\xi, \ \ \nabla_x \cdot B=0, \end{split}\right. \end{equation} $ (3.2)

这里$F_{i}=F_{i}(t, x, \xi)\geq 0, F_{e}=F_{e}(t, x, \xi)\geq 0$可分别看作离子和电子的密度函数, $x$$\xi$分别表示离子和电子的空间位置和速度, $(t, x, \xi)\in (0, \infty) \times \mathbb{R}^{3} \times \mathbb{R}^{3}$. $m_i$, $Ze$分别表示离子的质量和电量, 而$m_e$, $-e$分别表示电子的质量和电量.

(3.1)式中$Q_{\alpha\beta}(\cdot, \cdot)$ $(\alpha, \beta\in\{i, e\})$是Boltzmann碰撞算子, 这里取如下硬球模型的非对称形式

$ \begin{equation}\label{g.cop} Q_{\alpha\beta}(F_\alpha, F_\beta)=\frac{(\sigma_\alpha+\sigma_\beta)^2}{4}\int_{\mathbb{R}^3\times \mathbb{S}_+^2} [F_\alpha(\xi')F_\beta(\xi'_\ast)-F_\alpha(\xi)F_\beta(\xi_\ast)]|(\xi-\xi_\ast)\cdot\omega|\, d\xi_\ast d\omega, \end{equation} $ (3.3)

这里$\mathbb{S}_+^2=\{\omega\in\mathbb{S}^2: (\xi-\xi_\ast)\cdot\omega\geq0\}$, $\sigma_\alpha$$\alpha$粒子的直径, 本文取$\sigma_\alpha=\sigma_\beta=1$. $(\xi, \xi_{\ast})$$(\xi', \xi_{\ast}')$分别表示粒子碰撞前以及碰撞后的速度.由于粒子碰撞前后动量和能量守恒, 故

$ \begin{equation}\label{con.law} m_\alpha\xi+m_\beta\xi_{\ast}=m_\alpha\xi'+m_\beta\xi'_{\ast}, \quad m_\alpha|\xi|^2+m_\beta|\xi_{\ast}|^2=m_\alpha|\xi'|^2+m_\beta|\xi'_{\ast}|^2, \end{equation} $ (3.4)

由此易知

$ \begin{equation*} \xi'=\xi-\frac{2m_\alpha}{m_\alpha+m_\beta}[(\xi-\xi_\ast)\cdot\omega]\omega, \quad \xi_\ast'=\xi_\ast+\frac{2m_\alpha}{m_\alpha+m_\beta}[(\xi-\xi_\ast)\cdot\omega]\omega. \end{equation*} $

为方便起见, 将方程组(3.1)写成如下向量形式

$ \begin{equation}\label{ve.F} \partial_{t} F + \xi \cdot \nabla_{x}F + m^{-1}q \left(E + \xi \times B\right) \cdot \nabla_{\xi}F = C(F, F), \end{equation} $ (3.5)

这里记

$ F = \left( \begin{array}{ccc} F_{i}\\ F_{e}\\ \end{array} \right), \ \ m^{-1} = \left( \begin{array}{ccc} \frac{1}{m_{i}}&0 \\ 0&\frac{1}{m_{e}}\\ \end{array} \right), \ \ q = \left( \begin{array}{ccc} Ze&0 \\ 0&-e\\ \end{array} \right), \ \ C(F, F)= \left( \begin{array}{ccc} C_i(F, F) \\ C_e(F, F) \end{array} \right), $

其中$C_\alpha(F, F)=\sum\limits_{\beta}Q_{\alpha\beta}(F_\alpha, F_\beta)$.

现在令

$ \begin{eqnarray}\label{two.de} F = \left( \begin{array}{ccc} F_{i}\\ F_{e}\\ \end{array} \right) = \left( \begin{array}{ccc} {\bf{M}}_{i}\\ {\bf{M}}_{e}\\ \end{array} \right) + \left( \begin{array}{ccc} {\bf{G}}_{i}\\ {\bf{G}}_{e}\\ \end{array} \right), \end{eqnarray} $ (3.6)

并称此形式为一个两流体分解, 其中

$ \begin{equation} {\bf{M}}_{\alpha}(\xi) = \frac{n_{\alpha}}{(2\pi k_{\alpha}\theta)^{3/2}}\exp \left(- \frac{m_{\alpha}|\xi - u|^{2}}{2 k_{\alpha}\theta}\right), \ \ \alpha = i, e \end{equation} $ (3.7)

为标准的麦克斯韦分布, 这里$k_\alpha=k_B/m_\alpha$, $k_B$为Boltzmann常数, $n_\alpha, u, \theta$分别表示$\alpha$ -粒子的宏观密度、速度和温度.具体来讲, (3.7)式中的$[n_i, n_e, u, \theta](t, x)$由下式确定

$ \begin{eqnarray}\label{macroscopic-component} \left\{ \begin{array}{rll} \begin{split} &m_in_i\equiv \int_{{\mathbb{R}}^3}\psi_{0i}\cdot F(t, x, \xi)d\xi, \ \ m_en_e\equiv \int_{{\mathbb{R}}^3}\psi_{0e}\cdot F(t, x, \xi)d\xi, \\ &(m_{i}n_i+m_en_e)u_j\equiv\int_{{\mathbb{R}}^3} \psi_{j}\cdot F(t, x, \xi)d\xi, \ \ j=1, 2, 3, \\ &n_i\left(\theta+\frac{1}{2}m_i|u(t, x)|^2\right)+n_e\left(\theta+\frac{1}{2}m_e|u(t, x)|^2\right)\equiv \int_{{\mathbb{R}}^3}\psi_{4}\cdot F(t, x, \xi)d\xi, \end{split} \end{array}\right. \end{eqnarray} $ (3.8)

其中$\psi_{0i}$, $\psi_{0e}$$\psi_{j}$, $j=1, 2, 3, 4, $是六个碰撞不变量, 有如下形式

$ \begin{eqnarray} \psi_{0i}=\left(\begin{array}{cc} m_i \\ 0 \end{array}\right), \ \ \psi_{0e}=\left(\begin{array}{cc} 0 \\ m_e \end{array}\right), \ \ \psi_{j}=\left(\begin{array}{cc} m_i \xi_j \\ m_e\xi_j \end{array}\right), \ j=1, 2, 3, \ \ \psi_{4} =\left(\begin{array}{cc} \frac{1}{2}m_i|\xi|^2 \\ \frac{1}{2}m_e|\xi|^2 \end{array}\right). \end{eqnarray} $ (3.9)

进一步可验证

$ \begin{eqnarray*} &&\int_{{\mathbb{R}}^3} Q_{\alpha\beta}(F_\alpha, F_{\beta})d\xi=0, \ \ \int_{{\mathbb{R}}^3}m_\alpha\xi_j Q_{\alpha\alpha}(F_\alpha, F_{\alpha}) d\xi=0, \ j=1, 2, 3, \\ &&\int_{{\mathbb{R}}^3}\frac{1}{2}m_\alpha|\xi|^2 Q_{\alpha\alpha}(F_\alpha, F_{\alpha}) d\xi=0 \end{eqnarray*} $

$ \begin{eqnarray*} \left\{ \begin{array}{rll} \begin{split} &\int_{{\mathbb{R}}^3}m_\alpha\xi_j Q_{\alpha\beta}(F_\alpha, F_{\beta}) d\xi+\int_{{\mathbb{R}}^3}m_\beta\xi_j Q_{\beta\alpha}(F_\beta, F_{\alpha}) d\xi=0, \ j=1, 2, 3, \\ &\int_{{\mathbb{R}}^3}\frac{1}{2}m_\alpha|\xi|^2 Q_{\alpha\beta}(F_\alpha, F_{\beta}) d\xi+\int_{{\mathbb{R}}^3}\frac{1}{2}m_\mathcal {B}|\xi|^2 Q_{\beta\alpha}(F_\beta, F_{\alpha}) d\xi=0 \end{split} \end{array}\right. \end{eqnarray*} $

$\alpha\neq\beta$$(\beta\in\{i, e\})$成立.对于(3.7)式中定义的${\bf{M}}_\alpha$, 有

引理3.1${\bf{M}}_\alpha$定义如(3.7)式, 则

$ \begin{equation}\label{M.pro} Q_{\alpha\beta}({\bf{M}}_\alpha, {\bf{M}}_\beta)=0. \end{equation} $ (3.10)

将(3.7)式代入(3.3)式, 再利用(3.4)式即可得(3.10)式.现将$F$的两流体分解(3.6)代入(3.5)式可得

$ \partial_{t} ({\bf{M}} + {\bf{G}}) + \xi \cdot \nabla_{x}({\bf{M}} + {\bf{G}}) + m^{-1}q (E + \xi \times B) \cdot \nabla_{\xi}({\bf{M}} + {\bf{G}}) = L_{\bf{M}} {\bf{G}}+\Gamma({\bf{G}}, {\bf{G}}), $

这里

$ \begin{eqnarray} &L_{{\bf{M}}}{\bf{G}} = \left( \begin{array}{ccc} Q_{ii}({\bf{M}}_{i}, {\bf{G}}_{i}) + Q_{ii}({\bf{G}}_{i}, {\bf{M}}_{i}) + Q_{ie}({\bf{M}}_{i}, {\bf{G}}_{e}) + Q_{ie}({\bf{G}}_{i}, {\bf{M}}_{e})\\ Q_{ee}({\bf{M}}_{e}, {\bf{G}}_{e}) + Q_{ee}({\bf{G}}_{e}, {\bf{M}}_{e}) + Q_{ei}({\bf{M}}_{e}, {\bf{G}}_{i}) + Q_{ei}({\bf{G}}_{e}, {\bf{M}}_{i}) \end{array} \right), \\ &\Gamma ({\bf{G}}, {\bf{G}}) = \left( \begin{array} Q_{ii}({\bf{G}}_{i}, {\bf{G}}_{i}) + Q_{ie}({\bf{G}}_{i}, {\bf{G}}_{e}) \\ Q_{ee}({\bf{G}}_{e}, {\bf{G}}_{e}) + Q_{ei}({\bf{G}}_{e}, {\bf{G}}_{i}) \end{array} \right).\end{eqnarray} $ (3.11)

注意到定义式(3.3), 记

$ \begin{equation} \nonumber \begin{split} Q_{\alpha\beta}(F_{\alpha}, F_{\beta}) =& \int_{\mathbb{R}^3\times\mathbb{S}^2_+} {\bf{B}} \left[F_{\alpha}(\xi^{'})F_{\beta}(\xi_{*}^{'})-F_{\alpha}(\xi)F_{\beta}(\xi_{*})\right]d\xi_{*}d\omega\\ =& \int_{\mathbb{R}^3\times\mathbb{S}^2_+} {\bf{B}} F_{\alpha}(\xi^{'})F_{\beta}(\xi_{*}^{'})d\xi_{*}d\omega - \int_{\mathbb{R}^3\times\mathbb{S}^2_+} {\bf{B}} F_{\alpha}(\xi)F_{\beta}(\xi_{*})d\xi_{*}d\omega\\ =& Q_{\alpha\beta}^{{\rm gain}}(F_{\alpha}, F_{\beta}) + Q_{\alpha\beta}^{{\rm loss}}(F_{\alpha}, F_{\beta}), \end{split} \end{equation} $

其中$Q_{\alpha\beta}^{{\rm gain}}$$Q_{\alpha\beta}^{{\rm loss}}$分别称为“增益项”和“亏损项”, ${\bf{B}}=|(\xi-\xi_\ast)\cdot\omega|$.由(3.11)式, 可作如下分解

$ \begin{eqnarray} \nonumber L_{{\bf{M}}}{\bf{G}} =-\left( \begin{array}{ccc} \nu_{i}{\bf{G}}_i\\ \\ \nu_{e}{\bf{G}}_e \end{array}\right) +\left( \begin{array}{ccc} \sqrt{{\bf{M}}_{i}}K_{i}{\bf{G}}\\ \sqrt{{\bf{M}}_{e}}K_{e}{\bf{G}} \end{array}\right), \end{eqnarray} $

其中

$ -\nu_\alpha=\sum\limits_{\alpha\in\left\{i, e\right\}}Q_{\alpha\beta}^{{\rm loss}}(1, {\bf{M}}_\alpha), ~~ K_\alpha=K_\alpha^1+K_\alpha^2+K_\alpha^3+K_\alpha^4, $

这里

$ \begin{eqnarray} \nonumber K_{\alpha}^{1} {\bf{G}} &=& {\bf{M}}^{-\frac{1}{2}}_{\alpha}\sum\limits_{\beta\in\left\{i,e\right\}}Q_{\alpha\beta}^{{\rm loss}}({\bf{M}}_{\alpha}, {\bf{G}}_{\beta}) \\%+ Q_{ie}^{{\rm loss}}({\bf{M}}_{i}, {\bf{G}}_{e}) \nonumber\\ &=& - \int_{\mathbb{R}^3\times\mathbb{S}^2_+} B \sqrt{{\bf{M}}_{\alpha}(\xi)} \sqrt{{\bf{M}}_{\alpha}(\xi_{\ast})} \left(\frac{{\bf{G}}_{\alpha}}{\sqrt{{\bf{M}}_{\alpha}}}\right)(\xi_{\ast}) d\xi_{\ast} d\omega \nonumber\\&&- \int_{\mathbb{R}^3\times\mathbb{S}^2_+} B \sqrt{{\bf{M}}_{\alpha}(\xi)} \sqrt{{\bf{M}}_{\beta}(\xi_{\ast})} \left(\frac{{\bf{G}}_{\beta}}{\sqrt{{\bf{M}}_{\beta}}}\right)(\xi_{\ast}) d\xi_{\ast} d\omega,\ \ \alpha\neq\beta,\nonumber\\ K_{\alpha}^{2} {\bf{G}} &=& {\bf{M}}^{-\frac{1}{2}}_{\alpha}\left\{Q_{\alpha\alpha}^{{\rm gain}}({\bf{M}}_{\alpha}, {\bf{G}}_{\alpha}) + Q_{\alpha\alpha}^{{\rm gain}}({\bf{G}}_{\alpha}, {\bf{M}}_{\alpha})\right\} \\ &=& \int_{\mathbb{R}^3\times\mathbb{S}^2_+} B \sqrt{{\bf{M}}_{\alpha}(\xi_{\ast})} \left[\sqrt{{\bf{M}}_{\alpha}(\xi')} \left(\frac{{\bf{G}}_{\alpha}}{\sqrt{{\bf{M}}_{\alpha}}}\right)(\xi'_{\ast}) + \\\sqrt{{\bf{M}}_{\alpha}(\xi'_{\ast})} \left(\frac{{\bf{G}}_{\alpha}}{\sqrt{{\bf{M}}_{\alpha}}}\right)(\xi')\right] d\xi_{\ast} d\omega,\nonumber\\ %=& \int\int \widetilde{B_{ii}} \sqrt{M_{i}(\xi)} \sqrt{M_{i}(\xi_{\ast})} \sqrt{M_{i}(\xi'_{\ast})} % \frac{G_{i}}{\sqrt{M_{i}}}(\xi') d\xi_{\ast} d\omega\\ \nonumber K_{\alpha}^{3} {\bf{G}} &=& {\bf{M}}^{-\frac{1}{2}}_{\alpha}Q_{\alpha\beta}^{{\rm gain}}({\bf{M}}_{\alpha}, {\bf{G}}_{\beta}) \nonumber\\ &=& \int_{\mathbb{R}^3\times\mathbb{S}^2_+} B \sqrt{{\bf{M}}_{\beta}(\xi_{\ast})} \sqrt{{\bf{M}}_{\alpha}(\xi')} \left(\frac{{\bf{G}}_{\beta}}{\sqrt{{\bf{M}}_{\beta}}}\right)(\xi'_{\ast}) d\xi_{\ast} d\omega,\ \ \alpha\neq\beta,\nonumber\\ K_{\alpha}^{4} {\bf{G}} &=& {\bf{M}}^{-\frac{1}{2}}_{\alpha}Q_{\alpha\beta}^{{\rm gain}}({\bf{G}}_{\alpha}, {\bf{M}}_{\beta}) \nonumber\\ &=& \int_{\mathbb{R}^3\times\mathbb{S}^2_+} B \sqrt{{\bf{M}}_{\beta}(\xi_{\ast})} \sqrt{{\bf{M}}_{\beta}(\xi'_{\ast})} \left(\frac{{\bf{G}}_{\alpha}}{\sqrt{{\bf{M}}_{\alpha}}}\right)(\xi') d\xi_{\ast} d\omega,\ \ \alpha\neq\beta.\nonumber \end{eqnarray} $ (3.12)

现在记

$ \begin{eqnarray} \label{def.K} K{\bf{G}} =\left( \begin{array}{ccc} \sqrt{{\bf{M}}_{i}}K_i{\bf{G}} \\ \sqrt{{\bf{M}}_{e}}K_e{\bf{G}} \end{array}\right). \end{eqnarray} $ (3.13)

下面将证明$K$是某个Hilbert空间上的紧算子, 为此先引入加权$L^2$空间, 规定$f\in L^2(\mathbb{R}^3, \frac{1}{\sqrt{{\bf{M}}_\alpha}})$当且仅当$f{\bf{M}}_\alpha^{-1/2}\in L^2(\mathbb{R}^3)$.下面是本文的主要结论.

定理3.2$K$的定义如(3.13)式, 则$K$$L^2(\mathbb{R}^3, \frac{1}{\sqrt{{\bf{M}}_i}})\times L^2(\mathbb{R}^3, \frac{1}{\sqrt{{\bf{M}}_e}})$$L^2(\mathbb{R}^3, \frac{1}{\sqrt{{\bf{M}}_i}})\times L^2(\mathbb{R}^3, \frac{1}{\sqrt{{\bf{M}}_e}})$的紧算子.

现在证明$K$是紧算子, 即证明$K_{i}^{1}$, $K_{i}^{2}$, $K_{i}^{3}$, $K_{i}^{4}$, $K_{e}^{1}$, $K_{e}^{2}$, $K_{e}^{3}$, $K_{e}^{4}$是紧算子, 事实上只需要证明$K_{i}^{1}$, $K_{i}^{2}$, $K_{i}^{3}$, $K_{i}^{4}$是紧算子, $K_{e}^{1}$, $K_{e}^{2}$, $K_{e}^{3}$, $K_{e}^{4}$为紧算子同理可证.证明分以下四部分:

$1^{\circ}$易知

$ \sqrt{{\bf{M}}_{i}(\xi)} \sqrt{{\bf{M}}_{e}(\xi_{\ast})}\in L^2(\mathbb{R}^3\times\mathbb{R}^3), $

再结合引理2.2 [4]可知$K_{i}^{1}$显然是$L^2(\mathbb{R}^3, \frac{1}{\sqrt{{\bf{M}}_i}})\times L^2(\mathbb{R}^3, \frac{1}{\sqrt{{\bf{M}}_e}})$上的紧算子.

$2^{\circ}$现证明$K_{i}^{2}$为紧算子.对$K_{i}^{2}$算子, 注意到

$ \begin{equation} \nonumber \xi^{'} = \xi - [(\xi-\xi_{*})\cdot\omega]\omega, ~~ \xi_{*}^{'} = \xi_{*} + [(\xi-\xi_{*})\cdot\omega]\omega. \end{equation} $

$V = \xi_{*}-\xi$, 设$\omega^{\bot} \bot \omega$, 有$V = \omega(\omega \cdot V ) + \omega^{\bot}(\omega^{\bot} \cdot V )$, 且

$ \begin{eqnarray*} &&\xi^{'} = \xi + (\omega \cdot V )\omega = \xi_{*} -\omega^{\bot}(\omega^{\bot} \cdot V ), \\ && \xi_{*}^{'} = \xi_{*} - (\omega \cdot V )\omega = \xi+ \omega^{\bot}(\omega^{\bot} \cdot V ). \end{eqnarray*} $

从而变量替换$\omega\rightarrow \omega^{\bot}$意味着$ \xi'\rightarrow\xi_\ast', \ \ \xi_\ast'\rightarrow\xi'. $基于此令$\widetilde{B} = B(|\xi - \xi_{*}|, \omega) + B(|\xi - \xi_{*}|, \omega^{\bot})$, 从而$K_{\alpha}^{2}$可化简为

$ \begin{equation} \begin{split} K_{i}^{2} {\bf{G}}= \int_{\mathbb{R}^3\times\mathbb{S}^2_+} \widetilde{B} \sqrt{{\bf{M}}_{i}(\xi_{\ast})} \sqrt{{\bf{M}}_{i}(\xi'_{\ast})} \left(\frac{{\bf{G}}_{i}}{\sqrt{{\bf{M}}_{i}}}\right)(\xi') d\xi_{\ast} d\omega. \end{split} \end{equation} $ (3.14)

接着作变量替换$V = \xi_{*} - \xi $, 可得

$ \begin{equation} \begin{split} K_{i}^{2} {\bf{G}}= \int_{\mathbb{R}^3\times\mathbb{S}^2_+} \widetilde{B} \sqrt{{\bf{M}}_{i}(\xi_{\ast})} \sqrt{{\bf{M}}_{i}(\xi'_{\ast})} \left(\frac{{\bf{G}}_{i}}{\sqrt{{\bf{M}}_{i}}}\right)(\xi') dV d\omega. \end{split} \end{equation} $ (3.15)

进一步设$ V=\upsilon+W, $这里$\upsilon= (V \cdot \omega )\omega, \ \ W= (V -(V \cdot \omega )\omega ), $则有$dV = 2dW d|\upsilon|$, 并且

$ d\omega dV = d\omega \cdot 2dW d|\upsilon| = 2dW \cdot \frac{|\upsilon|^{2}d|\upsilon| d\omega}{|\upsilon|^{2}} = \frac{2d\upsilon dW}{|\upsilon|^{2}}. $

再由$\xi_{\ast} = \upsilon+W + \xi, \ \ \xi^{'} = \xi + \upsilon, \ \xi_\ast^{'} = \xi + W$, 可将$K_{i}^{2}$改写为

$ \begin{equation}\label{k2.2} \begin{split} K_{i}^{2}{\bf{G}} = 4\int_{\mathbb{R}^3\times \Pi} {\bf{M}}^{\frac{1}{2}}_{i}(\xi+\upsilon +W) {\bf{M}}_{i}^{\frac{1}{2}}(\xi + W) \left(\frac{{\bf{G}}_{i}}{\sqrt{{\bf{M}}_{i}}}\right)(\xi + \upsilon) \frac{1}{|\upsilon|} d\upsilon dW, \end{split} \end{equation} $ (3.16)

这里$\Pi=\{\upsilon\}^\perp$, 在推导上式中还用到了$\widetilde{B}d\omega=2|V|\cos \varphi\sin\phi d\varphi d\phi$.

为计算(3.16)式, 令$\eta = \xi + \upsilon$, 此时有

$ \begin{equation}\label{k2.3} \begin{split} K_{i}^{2}{\bf{G}}= 4\int_{\mathbb{R}^3\times \Pi} {\bf{M}}^{\frac{1}{2}}_{i}(\eta + W) {\bf{M}}_{i}^{\frac{1}{2}}(\xi + w) \left(\frac{{\bf{G}}_{i}}{\sqrt{{\bf{M}}_{i}}}\right)(\eta) \frac{1}{|\eta - \xi|} d\eta dW. \end{split} \end{equation} $ (3.17)

接下来, 对${\bf{M}}^{\frac{1}{2}}_{i}(\eta + W) {\bf{M}}_{i}^{\frac{1}{2}}(\xi+W)$作如下计算

$ \begin{equation}\label{k2.4} \begin{split} &{\bf{M}}^{\frac{1}{2}}_{i}(\eta + W) {\bf{M}}_{i}^{\frac{1}{2}}(\xi+W) \\=&\frac{n_{i}}{(2\pi k_{i}\theta)^{3/2}}\exp \left(- \frac{m_{i}\left(|\eta+W - u|^{2}+|\xi+W-u|^2\right)}{4 k_{B}\theta}\right)\\ =&\frac{n_{i}}{(2\pi k_{i}\theta)^{3/2}}\exp \left(- \frac{m_{i}\left(2|W +\frac{\xi+\eta}{2}- u|^{2}+\frac{1}{2}|\xi-\eta|^2\right)}{4 k_{B}\theta}\right) \\ =&\frac{n_{i}}{(2\pi k_{i}\theta)^{3/2}}\exp \left(- \frac{m_{i}|W +\frac{\xi+\eta}{2}- u|^{2}}{2 k_{B}\theta}\right)\exp \left(- \frac{m_{i}|\xi-\eta|^2}{8 k_{B}\theta}\right) \\ =&\frac{n_{i}}{(2\pi k_{i}\theta)^{3/2}}\exp \left(- \frac{m_{i}|W +\varsigma|^2}{2 k_{B}\theta}\right) \exp \left(- \frac{m_{i}|(\xi-u)^2-(\eta-u)^2|^{2}}{8 k_{B}\theta|\xi-\eta|^2}\right)\\ & \exp \left(- \frac{m_{i}|\xi-\eta|^2}{8 k_{B}\theta}\right), \end{split} \end{equation} $ (3.18)

这里

$ \varsigma=\frac{\xi+\eta-2u}{2}-\frac{\xi+\eta-2u}{2}\cdot\frac{\xi-\eta}{|\xi-\eta|} \frac{\xi-\eta}{|\xi-\eta|}. $

将(3.18)式代入(3.17)式可得

$ \begin{align} & [K_{i}^{2}{\bf{G}}](\xi )=[K_{i}^{2}{{\bf{G}}_{i}}](\xi )= \\ & \int_{{{\mathbb{R}}^{3}}}{\exp }\left( -\frac{{{m}_{i}}|{{(\xi -u)}^{2}}-{{(\eta -u)}^{2}}{{|}^{2}}}{8{{k}_{B}}\theta |\xi -\eta {{|}^{2}}} \right)\exp \left( -\frac{{{m}_{i}}|\xi -\eta {{|}^{2}}}{8{{k}_{B}}\theta } \right)\left( \frac{{{\bf{G}}_{i}}}{\sqrt{{{\bf{M}}_{i}}}} \right)(\eta )\frac{1}{|\eta -\xi |} \\ & \times 4\int_{\Pi }{\frac{{{n}_{i}}}{{{(2\pi {{k}_{i}}\theta )}^{3/2}}}}\exp \left( -\frac{{{m}_{i}}|W+\varsigma {{|}^{2}}}{2{{k}_{B}}\theta } \right)dWd\eta = \\ & 4{{n}_{i}}\int_{{{\mathbb{R}}^{3}}}{\frac{1}{|\xi -\eta |}}\exp \left( -\frac{{{m}_{i}}|{{(\xi -u)}^{2}}-{{(\eta -u)}^{2}}{{|}^{2}}}{8{{k}_{B}}\theta |\xi -\eta {{|}^{2}}} \right) \\ & \times \exp \left( -\frac{{{m}_{i}}|\xi -\eta {{|}^{2}}}{8{{k}_{B}}\theta } \right)\left( \frac{{{\bf{G}}_{i}}}{\sqrt{{{\bf{M}}_{i}}}} \right)(\eta )d\eta , \\ \end{align} $ (3.19)

从而$K_{i}^{2}$的核可表示为

$ \kappa^2_{\alpha}(\xi, \eta) = 4n_{i}\frac{1}{|\xi-\eta|}\exp \left(- \frac{m_{i}|(\xi-u)^2-(\eta-u)^2|^{2}}{8 k_{B}\theta|\xi-\eta|^2}\right) \exp \left(- \frac{m_{i}|\xi-\eta|^2}{8 k_{B}\theta}\right) . $

利用引理2.2 [5]可证$K_{i}^{2}$为紧算子, 具体证明可见文献[8, 9].

$3^{\circ}$ $K_{i}^{3}$的证明要更复杂, 这里先证明$K_{i}^{4}$为紧算子.

$K_{i}^{4}$, 与$K_{i}^{2}$不同的地方在于此时

$ \begin{equation} \nonumber \begin{split} \xi^{'} = \xi - \frac{2m_{e}}{m_{i} + m_{e}} [(\xi- \xi_{*})\cdot\omega]\omega, \ \ \xi_{*}^{'} = \xi_{*} + \frac{2m_{i}}{m_{i} + m_{e}} [(\xi-\xi_{*})\cdot\omega]\omega. \end{split} \end{equation} $

现在令$\xi_{*} - \xi = V = \upsilon + W$, 其中$ \upsilon = (V \cdot \omega) \omega, ~~W = V - (V \cdot \omega) \omega, $则有

$ \begin{equation} \nonumber \begin{split} \xi^{'} =& \xi + \frac{2m_{e}}{m_{i} + m_{e}} \upsilon, \\ \xi_{*}^{'} =& \upsilon + W + \xi - \frac{2m_{i}}{m_{i} + m_{e}}\upsilon = \frac{m_{e} - m_{i}}{m_{i} + m_{e}} \upsilon + W + \xi. \end{split} \end{equation} $

利用上面的关系式, 与(3.16)式的计算类似, 这时$K_{i}^{4}$可转化为

$ \begin{equation} \nonumber \begin{split} K_{i}^{4}{\bf{G}} = &4\int_{\mathbb{R}^3\times\Pi} \sqrt{{\bf{M}}_{e}(\upsilon + W + \xi)} \sqrt{{\bf{M}}_{e}\left(\frac{m_{e} - m_{i}}{m_{i} + m_{e}}\upsilon + W + \xi\right) }\\ &\times \left(\frac{{\bf{G}}_{i}}{\sqrt{{\bf{M}}_{i}}}\right)\left(\xi + \frac{2m_{e}}{m_{i} + m_{e}}\upsilon\right) \frac{d\upsilon dW}{|\upsilon|}. \end{split} \end{equation} $

进一步设$\eta=\xi + \frac{2m_{e}}{m_{i} + m_{e}} \upsilon $$\upsilon = \frac{m_{i} + m_{e}}{2m_{e}}(\eta - \xi).$从而

$ \begin{equation}\label{k4.1} \begin{split} K_{i}^{4}{\bf{G}} =& 4\left(\frac{2m_{e}}{m_{i}+m_{e}}\right)^{-3}\int_{\mathbb{R}^3\times \Pi} {\bf{M}}_{e}^{\frac{1}{2}}\left(\frac{m_{i}+m_{e}}{2m_{e}}(\eta-\xi)+W+\xi \right) \\ &\times{\bf{M}}_{e}^{\frac{1}{2}} \left(\frac{m_{e}-m_{i}}{2m_{e}}(\eta-\xi)+W+\xi \right) \left(\frac{{\bf{G}}_{i}}{\sqrt{{\bf{M}}_{i}}}\right)(\eta) \frac{1}{|\eta-\xi|} d\eta dW. \end{split} \end{equation} $ (3.20)

现在计算

$ \begin{equation}\label{k4.2} \begin{split} &{\bf{M}}_{e}^{\frac{1}{2}} \left(\frac{m_{i}+m_{e}}{2m_{e}}(\eta-\xi)+W+\xi \right) {\bf{M}}_{e}^{\frac{1}{2}} \left(\frac{m_{e}-m_{i}}{2m_{e}}(\eta-\xi)+W+\xi \right) \\ =&{\bf{M}}_{e}^{\frac{1}{2}}\left(\frac{m_{i}+m_{e}}{2m_{e}}\eta + \frac{m_{e}-m_{i}}{2m_{e}}\xi +W \right)\cdot {\bf{M}}_{e}^{\frac{1}{2}}\left(\frac{m_{i}+m_{e}}{2m_{e}}\xi + \frac{m_{e}-m_{i}}{2m_{e}}\eta +W \right) \\ =&{\bf{M}}_{e}^{\frac{1}{2}}\left(\frac{m_{i}}{2m_{e}}(\eta-\xi) + \frac{1}{2}(\eta+\xi) +W \right)\cdot {\bf{M}}_{e}^{\frac{1}{2}}\left(\frac{m_{i}}{2m_{e}}(\xi-\eta) + \frac{1}{2}(\eta+\xi) +W \right) \\ =&n_{e} (2\pi k_e \theta)^{-3/2} \exp \left(\frac{-m_{e}\left(\frac{m_{i}}{2m_{e}}(\eta-\xi)+\frac{1}{2}(\eta+\xi)+ W - u \right) ^{2}}{4k_B \theta} \right) \\ &\times\exp\left(\frac{-m_{e}\left(\frac{m_{i}}{2m_{e}}(\xi-\eta)+\frac{1}{2}(\eta+\xi)+ W - u \right)^{2}}{4k_B \theta} \right)\\ =&n_{e} (2k_e \pi \theta)^{-3/2} \exp \left(\frac{-m_{e}\cdot \frac{m_{i}^{2}}{2m_{e}^{2}}(\xi-\eta)^{2}} {4k_B \theta} - \frac{m_{e}\cdot 2 (W + \frac{1}{2}(\eta+\xi) -u)^{2}} {4k_B\theta} \right) \\ =&n_{e} (2k_e \pi \theta)^{-3/2} \exp \left(- \frac{m_{i}^{2}}{m_{e}} \frac{|\xi-\eta|^{2}}{8k_B \theta} - \frac{m_{e}}{2} \frac{((\eta-u)^{2}-(\xi-u)^{2})^{2}}{4k_B \theta |\eta-\xi|^{2}} \right) \\ &\times\exp \left(- \frac{m_{e} |W+\zeta_{2}|^{2}}{2k_B\theta} \right). \end{split} \end{equation} $ (3.21)

将(3.21)式代入(3.20)式可知$K_{i}^{4}$的核可以写为下列形式

$ \kappa^4_{i}(\xi, \eta) =4n_{e}\frac{1}{|\xi-\eta|} \exp\left(-\frac{m_{i}^{2}}{m_{e}}\frac{(\xi-\eta)^{2}}{8k_B\theta} -\frac{m_e((\eta-u)^{2}-(\xi-u)^{2})^{2}}{8k_B\theta|\xi-\eta|^{2}}\right). $

再次利用引理2.2 [5]可知$K_{i}^{4}$为紧算子.

$4^{\circ}$ 现在证明$K_{i}^{3}$是紧算子.与$K_{i}^{2}$, $K_{i}^{4}$的证明一样, 为了证明$K_{i}^{3}$是紧算子, 首先要将$K_{i}^{3}$化为一个更加简洁的形式, 为此先引入下面引理.

引理3.3 存在$b>0$对任意的$i, e$满足$m_{i}\neq m_{e}$及任意的$\xi, \xi_{*}\in \mathbb{R}^{3}$$\xi^{'}, \xi_{*}^{'}$

$ \begin{equation} \label{M4.0} \begin{split} \sqrt{{\bf{M}}_{\beta}(\xi_{*})}\sqrt{{\bf{M}}_{\alpha}(\xi^{'})} \leq b \left(\sqrt{{\bf{M}}_{\beta}(\xi_{*}^{'})} \sqrt{{\bf{M}}_{\alpha}(\xi)} \right). \end{split} \end{equation} $ (3.22)

现在对此引理给出证明, 选择$m_i\neq m_e$, 进行变量变换$(\xi-u, \xi_\ast-u, \xi'-u, \xi_\ast'-u)\rightarrow (\xi, \xi_\ast, \xi', \xi_\ast')$:

$ \begin{equation} \nonumber \begin{split} \xi^{'} = \xi - \frac{2m_{e}}{m_{i} + m_{e}} [(\xi- \xi_{*})\cdot\omega]\omega, \ \ \xi_{*}^{'} = \xi_{*} + \frac{2m_{i}}{m_{i} + m_{e}} [(\xi-\xi_{*})\cdot\omega]\omega \end{split} \end{equation} $

可化为

$ \begin{eqnarray} \label{v4.3} && \xi^{'} = \left( I_{3} - \frac{2m_{e}}{m_{i} + m_{e}}\omega \omega^{T} \right)\xi + \frac{2m_{e}}{m_{i} + m_{e}}\omega \omega^{T}\xi_{*}, \\ \label{v4.4} && \xi_{*}^{'} = \left( I_{3} - \frac{2m_{i}}{m_{i} + m_{e}}\omega \omega^{T} \right)\xi_{*} + \frac{2m_{e}}{m_{i} + m_{e}}\omega \omega^{T}\xi, \end{eqnarray} $ (3.23)

这里$I_{3}$$3\times 3$单位矩阵.由(3.24)式可以得到

$ \left( I_{3} - \frac{2m_{i}}{m_{i} + m_{e}}\omega \omega^{T} \right)\xi_{*} = \xi_{*}^{'} - \frac{2m_{e}}{m_{i} + m_{e}}\omega \omega^{T}\xi. $

为表达简便, 记

$ {\bf{A}} = \left( I_{3} - \frac{2m_{i}}{m_{i} + m_{e}}\omega \omega^{T} \right), $

因为$\det {\bf{A}} = \frac{m_{e} - m_{i}}{m_{e} + m_{i}}$$m_i\neq m_e$, 所以${\bf{A}}$可逆.因此有

$ \begin{equation} \label{v4.5} \begin{split} \xi_{*} =& \left( I_{3} - \frac{2m_{i}}{m_{i} + m_{e}}\omega \omega^{T} \right)^{-1}\xi_{*}^{'} - \left( I_{3} - \frac{2m_{i}}{m_{i} + m_{e}}\omega \omega^{T} \right)^{-1}\frac{2m_{i}}{m_{i} + m_{e}}\omega \omega^{T}\xi\\ =& {\bf{A}}^{-1}\xi_{*}^{'} + (I_{3} - {\bf{A}}^{-1})\xi. \end{split} \end{equation} $ (3.25)

把(3.25)式代入(3.23)式又可得到

$ \xi^{'} = - \frac{m_{e}}{m_{i}}(I_{3} - {\bf{A}}^{-1})\xi_{*}^{'} + \left( \frac{m_{i} + m_{e}}{m_{i}}I_{3} - \frac{m_{e}}{m_{i}} {\bf{A}}^{-1} \right)\xi. $

现在考虑如下的分块矩阵

$ \begin{eqnarray} \nonumber {\bf{B}} =\left( \begin{array}{ccc} {\bf{A}}^{-1}&\sqrt{\frac{m_{e}}{m_{i}}}(I_{3} - {\bf{A}}^{-1})\\ \frac{m_{i} + m_{e}}{m_{i}}I_{3} - \frac{m_{e}}{m_{i}}A^{-1} & -\sqrt{\frac{m_{e}}{m_{i}}}(I_{3} - {\bf{A}}^{-1}) \end{array}\right). \end{eqnarray} $

通过计算有$\det {\bf{A}} = 1$${\bf{A}}^{-1} ={\bf{A}} $, 则有

$ \begin{eqnarray} \nonumber \left( \begin{array}{ccc} \sqrt{m_{e}}\xi_{*}\\ \sqrt{m_{i}}\xi^{'} \end{array}\right) = {\bf{B}} \left( \begin{array}{ccc} \sqrt{m_{e}}\xi_{*}^{'}\\ \sqrt{m_{i}}\xi \end{array}\right). \end{eqnarray} $

事实上(3.22)式是通过找下面这个式子的下界得到的

$ \frac{\|{\bf{B}}[\sqrt{m_{e}}\xi_{*}^{'} \ \ \sqrt{m_{i}}\xi]^{T}\|^{2}}{\|[\sqrt{m_{e}}\xi_{*}^{'} \ \ \sqrt{m_{i}}\xi]^{T}\|^{2}}, $

且其是关于$\omega$的正函数并有$ b=\min \limits_{\omega \in S^{2}} \| {\bf{B}} \|^{2} > 0, $这就证明了引理3.2.

这时为了证明$K_{i}^{3}$是紧算子, 利用关系

$ \begin{eqnarray} \nonumber \left( \begin{array}{ccc} \xi_{*}-u\\ \xi^{'}-u \end{array}\right) = {\bf{B}} \left( \begin{array}{ccc} \xi_{*}^{'}-u\\ \xi-u \end{array}\right), \end{eqnarray} $

$K_{i}^{3}$可化为

$\begin{equation} \label{K1.1} K_{i}^{3} = \int_{\mathbb{R}^3\times\mathbb{S}^2_+} {\bf{B}} \sqrt{{\bf{M}}_e(\xi_{\ast}(\xi_{*}^{'}, \xi))} \sqrt{{\bf{M}}_i(\xi'(\xi_{*}^{'}, \xi))} \left(\frac{{\bf{G}}_e}{\sqrt{{\bf{M}}_e}}\right)(\xi_{*}^{'}) d\xi_{\ast} d\omega. \end{equation} $ (3.26)

再通过变量变换$\xi_{*} \rightarrow \xi_{*}^{'}$, 可以得到

$ \begin{equation} \nonumber \begin{split} (3.26) = \int_{\mathbb{R}^3\times\mathbb{S}^2_+} {\bf{B}} \sqrt{{\bf{M}}_e(\xi_{\ast}(\xi_{*}^{'}, \xi))} \sqrt{{\bf{M}}_i(\xi'(\xi_{*}^{'}, \xi))} \left(\frac{{\bf{G}}_e}{\sqrt{{\bf{M}}_e}}\right)(\xi_{*}^{'})|{\bf{A}}^{-1}| d\xi'_{\ast} d\omega. \end{split} \end{equation} $

上式也有确定的核形式, 其证明过程与$K_{i}^{2}$$K_{i}^{4}$类似, 可知$K_{i}^{3}$是紧的.同理可证$K_{e}^{1}$, $K_{e}^{2}$, $K_{e}^{3}$, $K_{e}^{4}$也是紧算子.

综合上述可证明积分算子$K$为紧算子.

参考文献
[1] 吴翊, 屈田心. 应用泛函分析[M]. 长沙: 国防科技大学出版社, 2002.
[2] 张恭庆, 林源渠. 泛函分析[M]. 北京: 北京大学出版社, 1987.
[3] 刘炳初. 泛函分析[M]. 北京: 科学出版社, 1998.
[4] 胡适耕. 泛函分析[M]. 北京: 科学出版社, 2003.
[5] 孙炯, 王万义, 赫建文. 泛函分析讲义[M]. 北京: 高等教育出版社, 2010.
[6] Peter, LaxD. 泛函分析[M]. 北京: 人民邮电出版社出版发行, 2010.
[7] 邸孝君. Banach空间的(L, P)性质[J]. 数学杂志, 1990, 1(1): 1–4.
[8] Boudin L, Grec B, Pavic M, Salvarani F. Diffusion asymptotics of a kinetic model for gaseous mixtures[J]. Kinet. Relat. Models, 2013, 6(1): 137–157.
[9] Glassey R T. The Cauchy problem in kinetic theory[M]. Philadelphia: Soc. Indus. Appl. Math., 1996.
[10] Duan R J, Liu S Q. Stability of the rarefaction wave of the Vlasov-Poisson-Boltzmann system[J]. SIAM J. Math. Anal., 2015, 47(5): 3585–3647. DOI:10.1137/140995179
[11] Duan R J, Liu S Q, Yang T, Zhao H J. Stability of the nonrelativistic Vlasov-Maxwell-Boltzmann system for angular non-cutoff potentials[J]. Kin. Rel. Mod., 2013, 6(1): 159–204.
[12] Duan R J, Lei Y J, Yang T, Zhao H J. The Vlasov-Maxwell-Boltzmann system near Maxwellians in the whole space with very soft potentials[J]. arXiv:1411.5150. http://link.springer.com/article/10.1007/s00220-017-2844-7
[13] Duan R J, Strain R M. Optimal large-time behavior of the Vlasov-Maxwell-Boltzmann system in the whole space[J]. Comm. Pure Appl. Math., 2011, 64(11): 1497–1546.
[14] Cercignani C. The Boltzmann equation and its apllications[M]. New York: Springer, 1988.
[15] Fan Y Z, L ei, Y J. Global solutions and time decay of the non-cutoff Vlasov-Maxwell-Boltzmann system in the whole space[J]. J. Stat. Phys., 2015, 161(5): 1059–1097. DOI:10.1007/s10955-015-1380-0
[16] Guo Y. The Vlasov-Maxwell-Boltzmann system near Maxwellians[J]. Invent. Math., 2003, 153: 593–630. DOI:10.1007/s00222-003-0301-z
[17] Ha S Y, Xiao Q H, Xiong L J, Zhao H J. L2-stability of the Vlasov-Maxwell-Boltzmann system near global Maxwellians[J]. J. Math. Phys., 2013, 54(12): 121509. DOI:10.1063/1.4850775
[18] Huang L R, Liu S Q. Stability of contact discontinuity for the two-species Vlasov-Poisson-Boltzmann system[J]. Math. Mod. Meth. Appl. Sci., 2016, 26(7): 1243–1318. DOI:10.1142/S0218202516500305
[19] Krall N A, Trivelpiece A W. Principles of plasma physics[M]. New York: McGraw-Hill Book Company, 1973.
[20] Lions P L. Compactness in Boltzmann's equation via Fourier integral operators and applications Ⅰ[J]. J. Math. Kyoto Univ., 1994, 34: 391–427. DOI:10.1215/kjm/1250519017
[21] Lions P L. Compactness in Boltzmann's equation via Fourier integral operators and applications Ⅱ[J]. J. Math. Kyoto Univ., 1994, 34: 429–461. DOI:10.1215/kjm/1250519018
[22] Lions P L. Compactness in Boltzmann's equation via Fourier integral operators and applications Ⅲ[J]. J. Math. Kyoto Univ., 1994, 34: 539–584. DOI:10.1215/kjm/1250518932
[23] Lei Y J, Zhao H J. The Vlasov-Maxwell-Boltzmann system with a uniform ionic background near Maxwellians[J]. J. Diff. Equ., 2016, 260(3): 2830–2897. DOI:10.1016/j.jde.2015.10.021
[24] Strain R M. The Vlasov-Maxwell-Boltzmann system in the whole space[J]. Comm. Math. Phys., 2006, 268(2): 543–567. DOI:10.1007/s00220-006-0109-y