广义线性回归极大似然估计的强相合性

设有该文第1节所描述的广义线性回归模型,以$\underline{\lambda}_n$和$\overline{\lambda}_n$分别记$\sum\limits_{i=1}^{n}Z_iZ_i^{\prime}$的最小和最大特征根,$\hat{\beta}_n$记$\beta_0$的极大似然估计.在文献[1]中,当\{$Z_i,i\ge1$\}有界时得到$\hat{\beta}_n$强相合的充分条件,在自然联系和非自然联系下分别为$\underline{\lambda}_n\rightarrow\infty$, $(\overline{\lambda}_n)^{1/2+\delta}=O(\underline{\lambda}_n)$（对某$\delta>0$)以及$\underline{\lambda}_n\rightarrow\infty$, $\overline{\lambda}_n=O(\underline{\lambda}_n)$.作者将后一结果改进为只要求$(\overline{\lambda}_n)^{1/2+\delta}=O(\underline{\lambda}_n)$,从而与自然联系情况下的条件达到一致.     

多维广义线性模型拟极大似然估计的弱相合性

本文考虑多维广义线性模型的拟似然方程$\tsm^n_{i=1}X_i(y_i-\mu(X_i'\xb))=0$, 在一定条件下证明了此方程的解$\wh\xb_n$渐近存在, 并得到了其收敛速度, 即$\wh\xb_n-\xb_0=O_p({\underline{\xl}}_n^{-1/2})$, 其中$\xb_0$为参数$\xb$的真值, $\underline{\xl}_n$是方阵$S_n=\tsm^n_{i=1}X_iX_i'$的最小特征值。     

UNBOUNDED SELF-ADJOINT OPERATOR ΠK SPACE

如果A是Πsubsub空间上的自共轭算子，由文[1]可知存在空间昨一个标准分解 \[{\Pi _k} = N \oplus \{ Z + {Z^*}\} \oplus P\] 在此分解下，A有三角模型\[A = \{ S,{A_N},{A_p},F,G,Q\} \].利用三角模型，我们直接证明了 定理1设A是\[{\Pi _k}\]上的-共轭算子，n是任何自然数，那末\[{A^n}\]也是自共轭算子. 定理2设A是\[{A^n}\]上的自共轭算子，那末对所有的\[{A^n}(n = 1,2,...)\]，存在一个公共 的标准分解，在此分解下 \[\begin{gathered} {A^n} = \{ {S^n},A_N^n,A_P^n,\sum\limits_{i = 0}^{n - 1} {{S^i}} FA_N^{n - 1 - i},\sum\limits_{i = 0}^{n - 1} {{S^i}GA_P^{n - 1 - i}} , \hfill \ \sum\limits_{i = 0}^{n - 1} {{S^i}} Q{S^{*n - 1 - i}} - \sum\limits_{i + j + k = n - 2} {{S^i}(FA_N^j{F^*} + GA_P^j{G^*}){S^{*k}}} \} \hfill \\ \end{gathered} \] 定理3 设A是瓜空间上的自共轭算子，\[\sigma (A) \subset [0,\infty ),0 \notin {\sigma _P}(A),\]，那末存在唯 一的自共轭算子A1，满足\[A_1^n = A,\sigma ({A_1}) \subset [0,\infty )\] 其次，我们研究了谱系在临界点附近的性状.记临界点全体为\[C(A)\]).对 \[{\lambda _0} \in C(A)\]记S与入0相应的最高阶根向量的阶数为\[r({\lambda _0})\] 定理4设A是\[{\Pi _k}\]空间上的无界自共轭算子，\[C(A) \cap ({\mu _1},{\nu _1}) = \{ {\lambda _0}\} \],那末以下四 个命题等价： (i）\[\mathop {sup}\limits_{\mu ,\nu } \{ \left\| {{E_{\mu \nu }}} \right\||{\lambda _0} \in (\mu ,\nu ) \subset ({\mu _1},{\nu _1})\} < \infty \] (ii）\[{\mu ^{{\text{1}}}}...,{\mu ^{{{\text{k}}_{\text{0}}}}}\]是全有限的测度； （iii）\[s - \lim {\kern 1pt} {\kern 1pt} {\kern 1pt} {E_{\mu \nu }}\]存在； （iv)A与\[{\lambda _0}\]相应的根子空间\[{\Phi _{{\lambda _0}}}\]非退化；这里\[{\mu ^{{\text{1}}}}...,{\mu ^{{{\text{k}}_{\text{0}}}}}\]是由\[{A_P}\]与G导出的测度. 定通5 设A是\[{\Pi _k}\]上自共轭算子，\[{\lambda _0} \in C(A),r({\lambda _0}) = n\]，那么 （i）\[{E_{\mu \nu }}\]在\[{{\lambda _0}}\]处的奇性次数不超过2n， （ii）\[s - \mathop {\lim }\limits_{\varepsilon \to 0} \int_{[{M_1},{\lambda _0} - \varepsilon )} {(t - {\lambda _0}} {)^{2n}}d{E_t},s - \mathop {\lim }\limits_{\varepsilon \to 0} \int_{[{\lambda _0} + \varepsilon ,{M_2})} {(t - {\lambda _0}} {)^{2n}}d{E_t},\]存在。这里\[{M_1},{M_2}\]满足\[[{M_1},{M_2}] \cap C(A) = \{ {\lambda _0}\} \] 定理6 设A是\[{\Pi _k}\]上的自共轭算子，临界点集\[C(A) = \{ {\lambda _1},...,{\lambda _l},{\lambda _{l + 1}},{\overline \lambda _{l + 1}},...,{\lambda _{l + p}},{\overline \lambda _{l + p}},\]，这里\[\operatorname{Im} {\lambda _v} = 0(1 \leqslant \nu \leqslant l),r({\lambda _\nu }) = {n_\nu }\]那么有 \[{(\lambda - A)^{ - 1}} = \int_{ - \infty }^\infty {K(\lambda ,t)d{E_t}} + \sum\limits_{\nu = 1}^l {\sum\limits_{i = 1}^{2{n_\nu } + 1} {\frac{{{B_{\nu i}}}}{{{{(\lambda - {\lambda _\nu })}^i}}}} } + \sum\limits_{\nu = l + 1}^{l + p} {\sum\limits_{i = 1}^{{n_\nu }} {[\frac{{{B_{\nu i}}}}{{{{(\lambda - {\lambda _\nu })}^i}}}} } + \frac{{B_{\nu i}^ + }}{{{{(\lambda - {{\overline \lambda }_v})}^i}}}]\] 这里 \[K(\lambda ,t) = \frac{1}{{\lambda - t}} - \sum\limits_{v = 1}^l {\delta (t - {\lambda _v}} )\sum\limits_{i = 1}^{2{n_v}} {\frac{{{{(t - {\lambda _v})}^{i - 1}}}}{{{{(\lambda - {\lambda _v})}^i}}}} ,\delta \lambda {\text{ = }}\left\{ \begin{gathered} {\text{1}}{\text{|}}\lambda {\text{| < }}\delta \hfill \ {\text{0}}{\text{|}}\lambda {\text{|}} \geqslant \delta \hfill \\ \end{gathered} \right.\] \[0 < \delta < \mathop {\min }\limits_\begin{subarray}{l} 1 \leqslant \mu ,v \leqslant l \\ {\lambda _\mu } \ne {\lambda _v} \end{subarray} |{\lambda _\mu } - {\lambda _v}|\].对\[1 \leqslant v \leqslant l\]，\[{B_{vi}}\]是\[{\Pi _k}\]上的有界自共轭算子，而当\[l + 1 \leqslant v \leqslant l + p\]时，\[{B_{vi}} = {({\lambda _\mu } - S)^{i - 1}}{P_{\lambda v}}\]是以与\[{{\lambda _v}}\]相应的根子空间为值域的某些平行投影. 定理7 在定理6的条件下，有 \[\begin{gathered} {\text{f}}(A) = \int_{ - \infty }^\infty {[f(t) - \sum\limits_{v = 1}^l {\delta (t - {\lambda _v}} } )\sum\limits_{i = 0}^{2{n_v} - 1} {\frac{{{f^{(i)}}({\lambda _v})}}{{i!}}} (t - {\lambda _v})d{E_t} \hfill \ {\text{ + }}\sum\limits_{{\text{v = 1}}}^{\text{l}} {\sum\limits_{i = 0}^{2{n_v}} {\frac{{{f^{(i)}}({\lambda _0})}}{{i!}}} } {B_v} + \sum\limits_{v = l + 1}^{l + p} {\sum\limits_{i = 0}^{{n_v} - 1} {[\frac{{{f^{(i)}}({\lambda _v})}}{{i!}}} } {B_{vi}} + \frac{{{f^{(i)}}({{\overline \lambda }_v})}}{{i!}}B_{vi}^ + ] \hfill \\ \end{gathered} \] 这里\[f(\lambda )\]在\[\sigma (A)\]的一个邻域内解析. 为了建立更一般的算子演算，我们引入两个特殊的代数： \[{\Omega _n} = \{ (f,\{ {a_i}\} _{i = 0}^{2n})|f\]为Borel可测函数，\[\{ {a_i}\} \]为一常数}。对\[F = (f,\{ {a_i}\} ) \in {\Omega _n},G = (g,\{ {b_i}\} ) \in {\Omega _n}\]，定义 \[\begin{gathered} \alpha F + \beta G = (\alpha f + \beta G,\{ \alpha {a_i} + \beta {b_i}\} ) \hfill \ F \cdot G = (f \cdot g,\{ \sum\limits_{j = 0}^i {{a_j}} {b_{i - j}}\} ),\overline F = (\overline f ,\{ {\overline a _i}\} ) \hfill \\ \end{gathered} \] 显然\[{\Omega _n}\]是一个交换代数，它的子代数\[{\omega _n}\]定义为 \[{\omega _n} = \{ F = (f,\{ {a_i}\} ) \in {\Omega _n}|\]在0点的一个与F有关的邻域中，成立\[{\text{|f(t) - }}\sum\limits_{i = 0}^{2n} {a{t^i}} | \leqslant {M_F}|t{|^{2n + 1}},{M_F}\]与F有关} 定义 设A是\[{\Pi _k}\]上的自共轭算子，C（A）={0}，r(0)=n,对\[F = (f,\{ {a_i}\} ) \in {\omega _n}\]，定义 \[\begin{gathered} FA{\text{ = }}\int_{{\text{ - }}\infty }^\infty {|f(t) - \sum\limits_{i = 0}^{2n} {{a_i}} } {t^i}{|^2}d{E_t} + \sum\limits_{i = 0}^{2n} {{a_i}} {A^i} \hfill \ DF(A)) = D({A^{2n}}) \cap \{ x \in {\Pi _k}\int_{{\text{ - }}\infty }^\infty {|f(t) - \sum\limits_{i = 0}^{2n} {{a_i}} } {t^i}{|^2}d{\left\| {{E_t}x} \right\|^2} < \infty \hfill \\ \end{gathered} \] 如果f解析，\[F = (f,\{ \frac{{{f^{(i)}}(0)}}{{i!}}\} )\]，那么可得F（A）=f（A）。 定理8 设A是有界自共轭算子，C（A）={0}，r（0）=n，\[G \in {\omega _n}\]，那么 \[\begin{gathered} \overline F (A) = {[F(A)]^ + },(\alpha F + \beta G)(A) = \alpha F(A) + \beta G(A) \hfill \ (FG)(A) = F(A)G(A). \hfill \\ \end{gathered} \] 定理9 设A是\[{\Pi _k}\]上的自共轭算子，C（A）={0}，r（0）=n，\[{F_1} = ({f_1},\{ {a_i}\} ) \in {\Omega _n}\]，\[{F_2} = ({f_2},\{ {a_i}\} ) \in {\omega _n},{f_1},{f_2}\]在\[( - \infty ,\infty )\]连续，在\[\sigma (A)\]上恒等，那么\[{F_1}(A) = {F_2}(A)\]。 定理10 设A是\[{\Pi _k}\]上自共轭算子C（A）={0}，r（0）=n，\[F = (f,\{ {a_i}\} ) \in {\Omega _n}\]f是连续函数，那么\[\sigma (F(A)) = \{ f(t)|t \in \sigma (A)\} \]。 在定理11中，我们建立了F（A）的三角模型并由此证明当\[F = \overline F \]时，\[C(F(A)) = \{ f(t)|t \in C(A)\} \] 定理12 设A施可析\[{\Pi _k}\]空间上的自共轭算子，C（A）={0}，r（0）=n，与0相应的根子空间非退化，T是稠定闭算子，那么\[T \in {\{ A\} ^{'}}\]的充要条件是存在\[F \in {\Omega _n}\]，使T=F（A）。这里\[{\{ A\} ^{'}} = \{ T|\]对满足\[BA \subset AB\]的有界算子B，均有\[BT \subset TB\]}     

A Hypothesis Testing Problem in the Linear Model (in English)

本文讨论了多元线性模型中的一个假设检验问题。假定 $\[{E(Y) = A\theta + B\eta }\]$ $Y的各行独立、正太、同协差阵V$ 现在要检验假设H_0:存在矩阵C使$\theta= C\eta$ 是否成立。首先可将问题化为法式的形式，对法式分两种情况进行讨论： （一）$[V = {\sigma ^2}I,{\sigma ^2}\]$未知，此时可求出 \theta,C，\sigma ^2的最大似然估计（当 H^0成立时）是 $[\left\{ {\begin{array}{*{20}{c}} {\hat \theta = {{({I_p} + \hat C'\hat C)}^{ - 1}}({y_1} + \hat C'{y_2})}\{\hat C = - {{({{T'}_{22}})}^{ - 1}}{{T'}_{12}}}\{{{\hat \sigma }^2} = \frac{1}{{nk}}(\sum\limits_{j = p + 1}^{p + q} {\lambda _j^* + \sum\limits_{j = 1}^k {{d_j})} } } \end{array}} \right.\]$ 其中y_1,y_2是法式 $[E\left( {\begin{array}{*{20}{c}} {{y_1}}\{{y_2}}\{{y_3}} \end{array}} \right) = \left( {\begin{array}{*{20}{c}} \theta \\eta \0 \end{array}} \right)\begin{array}{*{20}{c}} p\q\{n - (p + q)} \end{array}\]$ 中的资料阵y_1,y_2,d_1,\cdots，d_k是y^'_3y_3的全部特征根，$[\lambda _1^* \ge \cdots \lambda _{p + q}^*\]$是$[\left( {\begin{array}{*{20}{c}} {{y_1}}\{{y_2}} \end{array}} \right)\left( {\begin{array}{*{20}{c}} {{{y'}_1}}&{{{y'}_2}} \end{array}} \right)\]$的全部特征根，相应特征向量依$\lambda^*_i$的大小顺序从左到右排成矩阵T，T的分块子阵是T_ij，即 $[T = \left( {\begin{array}{*{20}{c}} {{T_{11}}}&{{T_{12}}}\{{T_{21}}}&{{T_{22}}} \end{array}} \right)\begin{array}{*{20}{c}} p\q \end{array}\]$ 对H_0的广义似然比检验是 $[\Lambda = \sum\limits_{j = p + 1}^k {{\lambda _j}/\sum\limits_{j = 1}^k {{d_j}} } \]$ $=lambda_1 \geq \lambda_2 \geq \cdots \geq \lambda_k$是$y_1^'y_1+y_2^'y_2$的全部特征根。 （二）一般情形V未知，此时 \theta,C的估计量同前，可求出 $[\hat V = \frac{1}{n}({y_2}^\prime {T_{22}}{T_{22}}^\prime {y_2} + {y_2}^\prime {y_2})\]$ H_0相应的Lawley不变检验是 $[\sum\limits_{j = p + 1}^k {{\beta _j}} \ge {\alpha _1}\]$ 其中 $\beta_1 \geq \beta_2 \geq \cdots \beta_k$是$y'_1y_1+y'_2y_2$的相应于$y'_sy_s$的全部特征根。 有关$\Lambda \$的以及$[\sum\limits_{j = p + 1}^k {{\beta _j}} \]$的极限分布将在另外的文章中讨论。     

Cn中单位球上$mu$-Bloch函数的等价刻画

设$\mu$是$[0,1)$上的正规函数, 给出了${\bf C}^{\it n}$中单位球$B$上$\mu$-Bloch空间$\beta_{\mu}$中函数的几种刻画. 证明了下列条件是等价的: (1) $f\in \beta_{\mu}$; \ (2) $f\in H(B)$且函数$\mu(|z|)(1-|z|^{2})^{\gamma-1}R^{\alpha,\gamma}f(z)$ 在$B$上有界; (3) $f\in H(B)$ 且函数${\mu(|z|)(1-|z|^{2})^{M_{1}-1}\frac{\partial^{M_{1}} f}{\partial z^{m}}(z)}$ 在$B$上有界, 其中$|m|=M_{1}$; (4) $f\in H(B)$ 且函数${\mu(|z|)(1-|z|^{2})^{M_{2}-1}R^{(M_{2})}f(z)}$ 在$B$上有界.     

分数次积分交换子在非齐Morrey空间上的紧性

本文主要建立由分数次积分$I_{\gamma}$与函数$b\in\mathrm{Lip}_{\beta}(\mu)$生成的交换子$[b, I_{\gamma}]$在以满足几何双倍与上部双倍条件的非齐度量测度空间为底空间的Morrey空间上紧性的充要条件.在假设控制函数$\lambda$满足逆双倍条件下,证明了交换子$[b,I_{\gamma}]$为从Morrey空间$M^{p}_{q}(\mu)$到$M^{s}_{t}(\mu)$紧性当且仅当$b\in\mathrm{Lip}_{\beta}(\mu)$.     

一类分数阶微分方程多点边值问题解的存在性

本文讨论下面一类分数阶微分方程多点边值问题 $$\align &D^{\alpha}_{0+}u(t) = f(t, u(t),~D^{\alpha-1}_{0+}u(t), D^{\alpha-2}_{0+}u(t), D^{\alpha-3}_{0+}u(t)),~~t\in(0,1), \\&I^{4-\alpha}_{0+}u(0) = 0, ~D^{\alpha-1}_{0+}u(0)=\displaystyle{\sum_{i=1}^{m}}\alpha_{i}D^{\alpha-1}_{0+}u(\xi_{i}),\\&D^{\alpha-2}_{0+}u(1)=\sum\limits_ {j=1}^{n}\beta_{j} D^{\alpha-2}_{0+}u(\eta_{j}),~D^{\alpha-3}_{0+}u(1)-D^{\alpha-3}_{0+}u(0)=D^{\alpha-2}_{0+}u(\frac{1}{2}),\endalign$$其中$3<\alpha \leq 4$是一个实数.通过应用Mawhin重合度理论和构建适当的算子,得到了该边值问题解的存在性结果.     

SECOND INITIAL-BOUNDARY VALUE PROBLEMS FOR QUASI-LINEAR HYPERBOLIC-PARABOLIC COUPLED SYSTEMS

On a rectangular domain \[R(\delta ) = \{ 0 \leqslant t \leqslant \delta ,0 \leqslant x \leqslant 1\} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} (1)\] We consider the second initial-boundary value problem for the quasi-linear hyperbolic- parabolic coupled system \[{\begin{array}{*{20}{c}} {\sum\limits_{j = 1}^n {{\zeta _{ij}}(t,x,u,v)(\frac{{\partial {u_j}}}{{\partial t}} + {\lambda _l}(t,x,u,v,{v_x})\frac{{\partial {u_j}}}{{\partial x}})} } \\ { = {\zeta _l}(t,x,u,v)(\frac{{\partial v}}{{\partial t}} + {\lambda _l}(t,x,u,v,{v_x})\frac{{\partial v}}{{\partial x}})} \\ { + {\mu _l}(t,x,u,v,{v_x}),(l = 1,...,n)} \\ {\frac{{\partial v}}{{\partial t}} - a(t,x,u,v,{v_x})\frac{{{\partial ^2}v}}{{\partial {x^2}}} = b(t,x,u,v,{v_x})} \end{array}}\] without loss of generatity,the initial conditions may be written as \[t = 0,{u_j} = 0,(j = 1,...,n),v = 0\] and we can suppose that \[\left\{ {\begin{array}{*{20}{c}} {a(0,x,0,0,0) \equiv 1} \\ {b(0,x,0,0,0) \equiv 0} \\ {{\zeta _{ij}}(0,x,0,0) \equiv {\delta _{lj}} = \left\{ {\begin{array}{*{20}{c}} {1,if{\kern 1pt} {\kern 1pt} {\kern 1pt} l = j} \\ {0,if{\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} l \ne j} \end{array}} \right.} \end{array}} \right.\] The boundary conditions are as follows: \[\begin{gathered} on{\kern 1pt} {\kern 1pt} {\kern 1pt} x = 1,\left\{ {\begin{array}{*{20}{c}} {{u_{\bar r}} = {G_{\bar r}}(t,u,v),(\bar r = 1,...,h;h \leqslant n)} \\ {\frac{{\partial v}}{{\partial x}} = {F_ + }(t,u,v);} \end{array}} \right. \hfill \ on{\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} x = 0,{\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} \left\{ {\begin{array}{*{20}{c}} {{u_{\hat s}} = {{\hat G}_{\hat s}}(t,u,v),(\hat s = m + 1,...,n;m \geqslant 0)} \\ {\frac{{\partial v}}{{\partial x}} = {F_ - }(t,u,v){\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} } \end{array}} \right. \hfill \\ \end{gathered} \] Uf = Q-f(t> u, x), (r = 1> k^n), We assume that the following conditions are satisfied: (1) the orientability condition \[\begin{gathered} {\lambda _{\bar r}}(0,1,0,0,0) < 0,{\lambda _s}(0,1,0,0,0) > 0,\left( {\begin{array}{*{20}{c}} {\bar r = 1,...,h} \\ {s = h + 1,...,n} \end{array}} \right) \hfill \ {\lambda _{\bar r}}(0,0,0,0,0) < 0,{\lambda _{\hat s}}(0,0,0,0,0) > 0,\left( {\begin{array}{*{20}{c}} {\hat r = 1,...,m} \\ {\hat s = m + 1,...,n} \end{array}} \right) \hfill \\ \end{gathered} \] (2) the compatibility condition \[\begin{gathered} \frac{{\partial {G_{\bar r}}}}{{\partial t}}(0,0,0) + \sum\limits_{j = 1}^n {\frac{{\partial {G_{\bar r}}}}{{\partial {u_j}}}} (0,0,0){\mu _j}(0,1,0,0,0) = {\mu _{\bar r}}(0,1,0,0,0) \hfill \ \frac{{\partial {{\hat G}_{\hat s}}}}{{\partial t}}(0,0,0) + \sum\limits_{j = 1}^n {\frac{{\partial {{\hat G}_{\hat s}}}}{{\partial {u_j}}}} (0,0,0){\mu _j}(0,0,0,0,0) = {\mu _{\hat s}}(0,0,0,0,0) \hfill \ (\bar r = 1,...,h;\hat s = m + 1,...,n);{F_ \pm }(0,0,0) = 0 \hfill \\ \end{gathered} \] (3) the condition of characterizing number \[\begin{gathered} \sum\limits_{j = 1}^n {\left| {\frac{{\partial {G_{\bar r}}}}{{\partial {u_j}}}(0,0,0)} \right|} < 1 \hfill \ \sum\limits_{j = 1}^n {\left| {\frac{{\partial {{\hat G}_{\hat s}}}}{{\partial {u_j}}}(0,0,0)} \right|} < 1(\bar r = 1,...,h,\hat s = m + 1,...,n \hfill \\ \end{gathered} \] (4)The smoothness condition: the coefficients of the system and the boundary conditions are suitably smooth. By means of certain a priori estimations for the solution of the heat equation and the linear hyperbolic system, using an iteration method and Leray-Schauder fixed point theorem, we have proved Theorem 1. Under the preceding hypotheses, for the second initial-boundary value problem (2)—(4), (6), (7), there exists uniquely a classical solution on R(8) where \[\delta \]>0 is suitably small. Theorem 2. In theorem the 1,condition of characterizing number (13) may be ameliorated as the following solvable condition; \[\left\{ {\begin{array}{*{20}{c}} {\det |({\delta _{\bar rr'}} - \frac{{\partial {G_{\bar r}}}}{{\partial {u_{r'}}}}(0,0,0)| \ne 0,(\bar r,r' = 1,...,h)} \\ {\det |({\delta _{\hat s\hat s'}} - \frac{{\partial {G_{\hat s}}}}{{\partial {u_{\hat s'}}}}(0,0,0)| \ne 0,(\hat s,\hat s' = m + 1,...,n)} \end{array}} \right.\] i.e,the boundary condition (6),(7)may be written as \[\begin{gathered} on{\kern 1pt} {\kern 1pt} {\kern 1pt} x = 1,\left\{ {\begin{array}{*{20}{c}} {{u_{\bar r}} = {H_{\bar r}}(t,{u_s},v),} \\ {\frac{{\partial v}}{{\partial x}} = {F_ + }(t,u,v);} \end{array}} \right.\left( {\begin{array}{*{20}{c}} {\bar r = 1,...,h} \\ {s = h + 1,...,n} \end{array}} \right) \hfill \ on{\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} x = 0,{\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} \left\{ {\begin{array}{*{20}{c}} {{u_{\hat s}} = {H_{\hat s}}(t,{u_{\hat r}},v){\kern 1pt} ,} \\ {\frac{{\partial v}}{{\partial x}} = {F_ - }(t,u,v){\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} } \end{array}} \right.\left( {\begin{array}{*{20}{c}} {\hat r = 1,...,m} \\ {\hat s = m + 1,...,n} \end{array}} \right) \hfill \\ \end{gathered} \]     

单位正方形上的一些振荡积分及其应用

设Q²=[0, 1]²是Eulid空间$\R^2$上的单位正方形, ${\mathcal{T}}_{\alpha,\beta}$是如下定义在Schwartz函数类${\mathcal{S}}(\R^3)$上振荡奇异积分算子
${\mathcal{T}}_{\alpha, \beta}f(x,y,z)=\int_{Q^2}f(x-t,y-s,z-t^ks^j)e^{-it^{-\beta_1}s^{-\beta_2}}t^{-1-\alpha_1} s^{-1-\alpha_2}dtds.
$
本文首先建立了该算子的L^p有界性, 然后利用这些结果获得了乘积空间上的一些奇异积分算子的(p, p)有界性.     

一类带p-Laplacian的Sturm-Liouville型2m点边值问题三个正解的存在性

该文考虑了下面的具一维$p$\,-Laplacian算子的多点边值问题 $ \left\{ \begin{array}{rl} &;\disp (\phi_{p}(x'(t)))'+h(t)f(t,x(t),x'(t))=0,\hspace{3mm}01,~\alpha_{i}>0,~\beta_{i}>0,~0<\sum\limits_{i=1}^{m-1}\alpha_{i}\xi_{i}\leq1,~ 0<\sum\limits_{i=1}^{m-1}\beta_{i}(1-\eta_{i})\leq1,~0=\xi_{0} <\xi_{1}<\xi_{2}<\cdots<\xi_{m-1}<\eta_{1}<\eta_{2}<\cdots<\eta_{m-1}<\eta_{m}=1,~i=1,2,\cdots,m-1.$ 通过运用锥上的不动点定理, 该文得到了至少三个正解的存在性. 有趣的是文中的边界条件是一个新型的Sturm-Liouville型边界条件, 这类边值问题到目前为止还很少被研究.     

一类带Hardy项的椭圆方程的无穷多解

假设 $\Omega=B_R:=\{x\in \mathbb{R}^N:|x|0$, $ N \geq 7$, $ 2^*=\frac{2N}{N-2}$, 我们得到了如下半线性问题无穷多解的存在性: $\left\{ \begin{array}{ll} -\Delta u=\frac{\mu}{|x|^2}u+|u|^{2^*-2}u+\la u, &; x\in\Omega, \\ u=0, &; x\in \partial\Omega. \end{array} \right.$ 其中$\lambda \in \mathbb{R}, \mu \in \mathbb{R}$. 这些解由不同的节点来区分.     

ON MILIN-LEBDEV INEQUALITIES

Let \[\varphi (x) = \sum\limits_{k = 1}^\infty {{A_k}} {x^k},\Phi (x) = {e^{\varphi (x)}} = \sum\limits_{k = 1}^\infty {{D_k}} {x^k}\] \[\begin{gathered} \frac{1}{{{{(1 - x)}^\lambda }}} = \sum\limits_{k = 1}^\infty {{d_k}} (\lambda ){x^k} \hfill \ {\overline \Delta _n}(\lambda ) = {\lambda ^{2 - p}}\sum\limits_{k = 1}^\infty {{k^{p - 1}}} \mathop {|{A_k}|}\nolimits_{}^p - \sum\limits_{k = 1}^\infty {\frac{1}{k}} \hfill \\ \end{gathered} \] Milin-Lebedey proved that \[\sum\limits_{k = 0}^\infty {\frac{{|{D_k}{|^p}}}{{d_k^{p - 1}(\lambda )}}} \leqslant \exp \{ {\lambda ^{1 - p}}\sum\limits_{k = 1}^\infty {{k^{p - 1}}} |{A_k}{|^p}\} \] where p>l and \[\lambda \]>0. In this paper, we have proved the following theorems； Theorem 1. Let \[p \geqslant 1,\lambda > 0\] and \[F(x) = \sum\limits_{k = 0}^\infty {\frac{{|{D_k}{|^p}}}{{d_k^p(\lambda )}}} {x^p}\exp \{ - {\lambda ^{1 - p}}\sum\limits_{k = 1}^\infty {{k^{p - 1}}|{A_k}{|^p}{x^k}} \} (2)\] then F(x) is a decreasing function of x on [0, 1]. This theorem is stronger than the result (1). Theorem 2. Let \[p \geqslant 2,\lambda > 0\] and \[{{\bar Q}_n}(\lambda ) = \frac{1}{{n + 1}}\sum\limits_{k = 0}^n {\frac{{|{D_k}{|^p}}}{{d_k^p(\lambda )}}\exp } \{ - \frac{1}{{n + 1}}\sum\limits_{v = 1}^n {\overline {{\Delta _p}} } (\lambda )\} \] then \[{{\bar Q}_n}(\lambda )\] is a decreasing fimctLon of n(n=l, 2，...)In the case p=2 this is contained in the Miiin-Lebedev's result.     

ON THE CHARACTERISTIC FUNCTION OF SEMI-HYPONORMAL OPERATOR

本文继[3]之后，研究拟亚正常算子和半亚正常算子的特征函数.设\[A = U|A{|_r}\]是\[H{\kern 1pt} {\kern 1pt} \] 上拟亚正常算子，\[U\]是酉算子，\[B = |A{|_ + } - |A{|_ - }\],作算子\[A\]的特征函数\[W(\lambda ,A) = I - {B^{\frac{1}{2}}}{(\lambda I - {A_ - })^{ - 1}}U{B^{\frac{1}{2}}}\] 定理1 设\[A = U|A{|_r}\]及\[{A^'} = {U^'}|{A^'}{|_r}\]为\[\varphi - \]拟亚正常算子而且都是简单的.又设 \[U\]与\[{U^'}|\]是酉算子.如果有酉算\[T\]将\[H\]映照成\[{H^'}\]而且\[|{A^'}{|_ \pm } = T|A{|_ \pm }{T^{ - 1}}\]，\[{U^'} = TU{T^{ - 1}}\]那末必有\[{\cal B}(A)\]到\[{\cal B}({A^'}){\kern 1pt} \]上的酉算子\[S{\kern 1pt} {\kern 1pt} \]使当\[\lambda \notin \sigma ({A_ - }) = \sigma (A_ - ^')\]时\[W(\lambda ,{A^'}) = SW(\lambda ,A){S^{ - 1}}\]反之亦真. 下面设\[A\]是半亚正常的.又设\[{\cal D}\]为一辅助的希尔伯特空间，\[K\]为\[{\cal D}\]到\[{\kern 1pt} H\]中的线 性算子使\[Q = |A{|_{\rm{r}}} - |A{|_l} = K{K^*}{\kern 1pt} {\kern 1pt} \]，当\[\lambda \in \rho (A)\]，\[|Z| \ne 1\]时作 \[Y(z,\lambda ) = I - {\kern 1pt} {\kern 1pt} z{K^*}{(I - z{U^*})^{ - 1}}{(A - \lambda I)^{ - 1}}K\] 定理2设\[A = U|A{|_r}{\kern 1pt} {\kern 1pt} {\kern 1pt} \]及\[{A^'} = {U^'}|{A^'}{|_r}{\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} \]分别是\[H\]与\[{H^'}{\kern 1pt} {\kern 1pt} {\kern 1pt} \]中的半亚正常算子，\[U\]与 \[{U^'}\]是酉算子而且\[A\]与\[{A^'}\]都是简单的.如果存在\[{\cal D} \to {{\cal D}^'}{\kern 1pt} \]上的酉算子\[S\]使 \[{Y^'}(z,\lambda ) = SY(z,\lambda ){S^{ - 1}}\] 那末必有由\[H\]到\[{H^'}{\kern 1pt} {\kern 1pt} {\kern 1pt} \]上的酉算子\[T\]使(1)成立，反之亦真. 定理3 若\[K\]是希尔伯特-许密特算子则\[Y(z,\lambda )\]的行列式（当\[|Z| \ne 1\]时）存在， 且\[\det (Y(z,\lambda )) = \det ((I - z{U^*})(A - \lambda I){(I - z{U^*})^{ - 1}}{(A - \lambda I)^{ - 1}})\] 下面只考虑奇型积分模型这时\[W(\lambda ,A)\]成为乘法算子，\[(W(\lambda ,A)f)({e^{i\theta }}) = W({e^{i\theta }},\lambda )f({e^{i\theta }})\]其中\[W({e^{i\theta }},\lambda ) = I - \alpha ({e^{i\theta }}){(\lambda {e^{i\theta }}I - \beta ({e^{i\theta }}))^{ - 1}}\alpha ({e^{i\theta }})\] 我们又假设\[A\]是完全非正常的.记\[{Y_ \pm }({e^{i\theta }},\lambda )a = \mathop {\lim }\limits_{r \to 1 \pm 0} Y({e^{i\theta }},\lambda )a\] 定理4设\[\lambda \in \rho (A){\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} \]，\[a \in {\cal D}{\kern 1pt} {\kern 1pt} {\kern 1pt} \]为固定的，那末\[{Y_ \pm }({e^{i\theta }},\lambda )a\]为黎曼-希尔伯特问题 \[{Y_ - }({e^{i\theta }},\lambda )a = W({e^{i\theta }},\lambda ){Y_ + }({e^{i\theta }},\lambda )a\] 的解. 设\[{\cal L}({\cal D}{\kern 1pt} {\kern 1pt} {\kern 1pt} ){\kern 1pt} {\kern 1pt} {\kern 1pt} \]为\[{\cal D}{\kern 1pt} {\kern 1pt} {\kern 1pt} \]上线性有界算子全体所成的Banach空间，\[H_ \pm ^p({\cal L}{\kern 1pt} ({\cal D}{\kern 1pt} {\kern 1pt} ){\kern 1pt} {\kern 1pt} ){\kern 1pt} {\kern 1pt} \]为单位圆 外，内取值于\[{\cal L}({\cal D}{\kern 1pt} {\kern 1pt} {\kern 1pt} ){\kern 1pt} \]的某些解析函数所成的Hardy空间.设\[f({e^{i\theta }})\]是单位圆周上的函 数，如果有\[{u_ \pm } \in H_ \pm ^p({\cal L}{\kern 1pt} ({\cal D}{\kern 1pt} {\kern 1pt} ){\kern 1pt} {\kern 1pt} ){\kern 1pt} {\kern 1pt} (p > 2)\]使\[u_ - ^{ - 1}{\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} {\kern 1pt} \]存在\[{u_ - }{\kern 1pt} {\kern 1pt} {\kern 1pt} {({e^{i\theta }})^{ - 1}}{u_ + }{\kern 1pt} ({e^{i\theta }}) = f({e^{i\theta }})\]则称\[f\]是可分解的.     

THE UNIFORM INTEGRABILITY OF A CLASS OF EXPONENTIAL MARTINGALES

若说\[(\Omega ,\mathcal{F},P)\]为完备概率空间，\[F = {({\mathcal{F}_t})_{t \in [a,b]}}\]为\[\mathcal{F}\]的递增子\[\sigma \]域族，且满足通常 条件，\[b \leqslant \infty \].又\[W = \{ {W_t},0 \leqslant t \leqslant b\} \]为关于F的Wiener过程，\[X = \{ {X_t},0 \leqslant t < b\} \]为 循序讨测过程，且 \[P\{ \int_0^b {X_t^2} dt < \infty \} = 1\]， 则可定义X关于W的Ito随机积分 \[{(X \cdot W)_t} = \int_0^t {{X_s}} d{W_s},0 \leqslant t \leqslant b\] 这时若记 \[{Z_t} = \exp \{ \int_0^t {{X_s}} d{W_s} - \frac{1}{2}\int_0^t {{X_s}^2} ds\} \] 它便是一个指数（局部）鞅.本文的目的在于证明当X为循序可测正态过程时，只要X关于W的积分存在，\[{\text{\{ }}{Z_t}0 \leqslant {\text{t < b\} }}\]总是一致可积的。 引理1若\[\{ {Z_t},0 \leqslant t < b\} \]为实可测正态过程且 \[\int_0^{\text{b}} {\left\| {{X_t}} \right\|} d{m_t} < \infty \] 其中\[\left\| {{X_t}} \right\| = {(E|{X_t}{|^2})^{1/2}}\],\[{m_t}\]为[0,b)上右连续递增函数，则X的几乎所有样本函数关于\[{m_t}\]可积，且其轨道积分 \[\tilde I = \int_0^{\text{b}} {{X_t}} d{m_t}\] 为正态分布随机变量. 引理2若\[X = \{ {X_t},0 \leqslant t < b\} \]为可测正态过程，其几乎所有样本函数关于右连续增函数\[{m_t}\]可积，即 \[P(\int_0^b {|{X_t}} |d{m_t} < \infty ) = 1\] 则按轨道积分 \[\tilde I = \int_0^{\text{b}} {{X_t}} d{m_t}\] 是正态分布随机变量. 引理3 若\[\{ {\xi _n},n \geqslant 1\} \]为正态分布随机变量序列，则 \[\sum\limits_{j = 1}^\infty {E{\xi _i}^2} \leqslant {[Eexp( - \frac{1}{2}\sum\limits_{j = 1}^\infty {{\xi _i}^2} )]^{ - 2}}\] 进而若\[\sum\limits_{j = 1}^\infty {E{\xi _i}^2} < 1\],则 \[E[exp(\frac{1}{2}\sum\limits_{j = 1}^\infty {{\xi _i}^2} )] \leqslant {(1 - \sum\limits_{j = 1}^\infty {E{\xi _i}^2} )^{ - \frac{1}{2}}}\] 引理4若\[{m_s}\]为[0, b)上右连续增函数，又\[X = \{ X_t^{(i)},0 \leqslant t < b,1 \leqslant i < \infty \} \]为正态 过程，则当\[P\{ \sum\limits_{i = 1}^\infty {\int_0^b {{{({X_t}^{(i)})}^2}d{m_t}} } < \infty \} = 1\]时必有 \[\sum\limits_{i = 1}^\infty {\int_0^b {{{({X_t}^{(i)})}^2}d{m_t}} } < \infty \} = 1\] 进而若;\[\sum\limits_{i = 1}^\infty {\int_0^b {{{({X_t}^{(i)})}^2}d{m_t}} } < 1\]，必有 \[Eexp(\frac{1}{2}\sum\limits_{i = 1}^\infty {\int_0^b {{{({X_t}^{(i)})}^2}d{m_s}} } ) \leqslant {(1 - \sum\limits_{j = 1}^\infty {E\int_0^b {{{({X_t}^{(i)})}^2}d{m_s}} } )^{ - \frac{1}{2}}}\] 定理 若\[W = (W_t^{(1)},...,W_t^{(n)},...)\]为一个具有无限个分量的过程，其分量都是连续 正态独立增量过程且满足 \[\begin{gathered} E\{ W_t^{(i)} - W_s^{(i)}\} = 0 \hfill \ E\{ (W_t^{(i)} - W_s^{(i)})(W_t^{(j)} - W_s^{(j)})\} = {\delta _{ij}}(m_t^{(i)} - m_s^{(i)}) \hfill \\ \end{gathered} \] 又\[\{ {f_t} = (f_t^{(1)},...,f_t^{(n)},...)\} \]为循序可测正态过程，若 \[P\{ \sum\limits_{i = 1}^\infty {\int_0^b {{{({f_t}^{(i)})}^2}dm_t^{(i)}} } < \infty \} = 1\] 则 \[{Z_t} = \exp \{ \sum\limits_{i = 1}^\infty {\int_0^b {{f_s}^{(i)}dW_s^{(i)} - \frac{1}{2}\int_0^t {{{({f_s}^{(i)})}^2}dm_s^{(i)}} } } \} ,0 \leqslant t < b\] 是一致可积鞅，特别有\[E{Z_0} = 1\] 利用上述结果及正态过程的Hida-Cramer分解，可以象[1]一样方便地讨论正态测 度的等价性问题并求出其Radon-Nikodym导数.     

最佳逼近与Fourier系数的等价关系

探讨最佳逼近E_n(f)与函数的Fourier系数\!$\hat{f}(n)\in {\bf C},n=0,\pm 1,\pm 2,\ldots$, 在\!$\{\hat{f}(n)\}_{n=0}^{\infty }\linebreak\in $MVBVS^*和$\{\hat{f}(n)+f\left( -n\right) \}_{n=0}^{\infty }\in$ MVBVS^*条件下的等价关系问题, 此地MVBVS^*为所称的强均值有界变差(strong mean value bounded variation)数列的集合.     

构造相应于有限维非退化可解李代数的顶点代数

设g是带有非退化不变对称双线性型的有限维可解李代数, 该文首先应用g的仿射李代数{\heiti $\hat{g}$}的表示理论,构造出一类水平为l的限制$\hat{g}$ -模$V_{\hat{g}}(l,0)$.然后应用顶点算子的局部理论在hom$(V_{\hat{g}}(l,0),V_{\hat{g}}(l,0)((x)))$中 找到一类顶点代数$L_{V_{\hat{g}}(l,0)}$.建立了$L_{V_{\hat{g}}(l,0)}$到 $V_{\hat{g}}(l,0)$的映射,最后证明了这类映射是顶点代数同构.     

与强奇异 Calder\''{o}n-Zygmund算子相关的Toeplitz算子的双权估计

研究了与强奇异Calder\'{o}n-Zygmund算子和加权 Lipschitz函数${\rm Lip}_{\beta_0,\omega}$相关的Toeplitz算子$T_b$的sharp极大函数的点态估计,并证明了Toeplitz算子是从 $L^p(\omega)$到$L^q(\omega^{1-q})$上的有界算子.此外, 建立了与强奇异Calder\'{o}n-Zygmund算子和加权 BMO函数${\rm BMO}_{\omega}$相关的Toeplitz算子$T_b$的sharp极大函数的点态估计,并证明了Toeplitz算子是从 $L^p(\mu)$到$L^q(\nu)$上的有界算子.上述结果包含了相应交换子的有界性.     

ON MILIN-LEBDEV INEQUALITIES

Let \[\varphi (x) = \sum\limits_{k = 1}^\infty {{A_k}} {x^k},\Phi (x) = {e^{\varphi (x)}} = \sum\limits_{k = 1}^\infty {{D_k}} {x^k}\] \[\begin{gathered} \frac{1}{{{{(1 - x)}^\lambda }}} = \sum\limits_{k = 1}^\infty {{d_k}} (\lambda ){x^k} \hfill \ {\overline \Delta _n}(\lambda ) = {\lambda ^{2 - p}}\sum\limits_{k = 1}^\infty {{k^{p - 1}}} \mathop {|{A_k}|}\nolimits_{}^p - \sum\limits_{k = 1}^\infty {\frac{1}{k}} \hfill \\ \end{gathered} \] Milin-Lebedey proved that \[\sum\limits_{k = 0}^\infty {\frac{{|{D_k}{|^p}}}{{d_k^{p - 1}(\lambda )}}} \leqslant \exp \{ {\lambda ^{1 - p}}\sum\limits_{k = 1}^\infty {{k^{p - 1}}} |{A_k}{|^p}\} \] where p>l and \[\lambda \]>0. In this paper, we have proved the following theorems； Theorem 1. Let \[p \geqslant 1,\lambda > 0\] and \[F(x) = \sum\limits_{k = 0}^\infty {\frac{{|{D_k}{|^p}}}{{d_k^p(\lambda )}}} {x^p}\exp \{ - {\lambda ^{1 - p}}\sum\limits_{k = 1}^\infty {{k^{p - 1}}|{A_k}{|^p}{x^k}} \} (2)\] then F(x) is a decreasing function of x on [0, 1]. This theorem is stronger than the result (1). Theorem 2. Let \[p \geqslant 2,\lambda > 0\] and \[{{\bar Q}_n}(\lambda ) = \frac{1}{{n + 1}}\sum\limits_{k = 0}^n {\frac{{|{D_k}{|^p}}}{{d_k^p(\lambda )}}\exp } \{ - \frac{1}{{n + 1}}\sum\limits_{v = 1}^n {\overline {{\Delta _p}} } (\lambda )\} \] then \[{{\bar Q}_n}(\lambda )\] is a decreasing fimctLon of n(n=l, 2，...)In the case p=2 this is contained in the Miiin-Lebedev's result.