数学物理学报, 2019, 39(1): 105-113 doi:

论文

带瞬时脉冲的分数阶非自制发展方程解的存在唯一性

朱波,1, 刘立山2

Existence and Uniqueness of the Mild Solutions for a Class of Fractional Non-Autonomous Evolution Equations with Impulses

Zhu Bo,1, Liu Lishan2

通讯作者: 朱波, E-mail: zhubo207@163.com

收稿日期: 2017-09-14  

基金资助: 山东省高校科技计划项目.  J16LI14
国家自然科学基金.  11871302

Received: 2017-09-14  

Fund supported: the PSDPHESTP.  J16LI14
the NSFC.  11871302

摘要

该文利用广义Banach不动点定理研究了一类带迟滞和瞬时脉冲的分数阶非自治发展方程初值问题解的存在性和唯一性,给出其解的迭代序列和误差估计并讨论了其解是连续依赖于初值的.

关键词: 分数阶非自治发展方程 ; 非瞬时脉冲 ; 预解算子 ; 广义Banach不动点定理

Abstract

In this paper, we consider a class of fractional non-autonomous evolution equations with impulses and delay. By the generalized Banach fixed point theorem, we obtain some new results on the existence and uniqueness of the mild solution. An explicit iterative scheme for the mild solution and an error estimate of the approximation sequence for the initial value problem are also derived. Moreover, the unique mild solution of the problem is continuously dependent on the initial value.

Keywords: Fractional non-autonomous evolution equations ; Non-instantaneous impulse ; Resolvent operator ; Generalized Banach fixed point theorem

PDF (319KB) 元数据 多维度评价 相关文章 导出 EndNote| Ris| Bibtex  收藏本文

本文引用格式

朱波, 刘立山. 带瞬时脉冲的分数阶非自制发展方程解的存在唯一性. 数学物理学报[J], 2019, 39(1): 105-113 doi:

Zhu Bo, Liu Lishan. Existence and Uniqueness of the Mild Solutions for a Class of Fractional Non-Autonomous Evolution Equations with Impulses. Acta Mathematica Scientia[J], 2019, 39(1): 105-113 doi:

1 引言

近几十年来,分数阶微积分在描述分形、多孔介质弥散、粘弹性介质的应力应变等方面的成功应用而得到众多研究者的关注,并且分数阶微积分在物理、工程、金融等领域应用广泛.相比整数阶微积分,分数阶微积分能更好的刻画各种材料和过程的记忆性.而带有脉冲的微分方程在理论和应用方面正在受到国内外研究者们的关注.脉冲微分方程能被用于描述具有突然的或瞬时的变化现象和系统.许多科研人员在抽象空间对带瞬时脉冲的微分方程解的存在性做了大量研究并获得了很多成果.详细内容参见文献[1-11].

本文,我们研究如下的带迟滞和瞬时脉冲的分数阶非自治发展方程初值问题

$ ^{c}D_{t}^{\beta}u(t)=A(t)u(t)+f(t, u(\tau_1 (t)), u(\tau_2 (t)), \cdots , u(\tau_l (t))), \ \ t\in J, t\neq t_j, $

$\Delta u(t_j)=I_j(u(t_j)), t=t_j, j=1, 2, \cdots , m, $

$u(0)=u_0 \in E, $

这里$J=[0, T_0], 0\leq \tau_i(t)\leq t \ (i=1, 2, \cdots , l), \beta \in(0, 1]$, $l$是正整数, $0 <t_1 <t_2 <\cdots <t_k <\cdots <t_m <t_{m+1}=T_0$, $I_j:E\rightarrow E$是脉冲函数, $f:J\times E^l\rightarrow E$是连续函数, $\ ^{c}D_{t}^{\beta}$$\beta$阶Caputo分数阶导数, $\{A(t)\}_{t\in J}$是定义在Banach空间$E$上的一个闭线性算子族.

陈、张和李[12]研究了如下带迟滞的分数阶非自治发展方程

$^{c}D_{t}^{\beta}u(t)=A(t)u(t)+f(t, u(\tau_1 (t)), u(\tau_2 (t)), \cdots, u(\tau_l (t))), \ \ t\in J, $

$u(0)=u_0 \in E, $

利用$k$ -集压缩映像得到温和解的存在性定理.而在定理的条件中需要压缩系数小于1的限制条件.在文献[13-15]中,作者研究了如下带迟滞的分数阶反应扩散方程初边值问题

$^{c}D_{t}^{\beta}u(x, t)-a(t)u_{xx}(x, t)=f(t, u(x, \tau _1(t)), u(x, \tau _2(t)), \cdots , u(x, \tau _l(t))), \ (x, t)\in \Omega\times J, $

$u(x, t)=0, (x, t)\in \partial \Omega \times \mathbb{R} _+, $

$u(x, 0)=\varphi(x), x\in \Omega, $

这里$\beta \in(0, 1]$, $l$是正整数, $f:J\times \mathbb{R} ^l\rightarrow \mathbb{R} $是连续函数, $a(t)$是扩散系数且在$J$上连续, $\tau _i:J\rightarrow J \ (i=1, 2, \cdots , l)$是连续函数且$0\leq\tau _i(t)\leq t \ (i=1, 1, \cdots , l)$, $\Omega\subset \mathbb{R} ^l$有充分光滑的边界, $\partial \Omega, \varphi \in L^2(\Omega)$, $\ ^{c}D_{t}^{\beta}$$\beta$阶Caputo分数阶导数.欧阳[13]给出了问题(1.6)-(1.8)局部解的存在性.不同于文献[13]的方法,朱、刘和吴在文献[14-15]中讨论了问题(1.6)-(1.8)整体解的存在性,并且作者首次将反应扩散方程初边值问题转化为抽象形式的分数阶非自治发展方程(1.4)-(1.5).本文,我们利用广义Banach压缩映像原理讨论了带脉冲的分数阶非自治发展方程解的存在性和唯一性.该方法不需要单独给出条件来保证压缩系数小于1,而已有的利用压缩映像原理证明解的存在性时都需要单独给出条件来保证压缩系数小于1.因此,我们的方法推广和改进了前人的结果.

2 预备知识和引理

显然$PC(J, E)$是一个Banach空间,具有范数

定义2.1[16-17]  设$A(t):D(A)\subset E$是一个闭线性算子, $\beta>0$, $\rho[A(t)]$$A(t)$的预解集,如果存在$\omega\geq 0$和一个强连续函数$U_\beta:\mathbb{R} _+^2\rightarrow B(E)$使得$\{\lambda^\beta:$ Re $\lambda >\omega\}\subset \rho (A)$

成立,我们称$A(t)$生成一个$\beta$ -预解族.这里, $U_\beta(t, s)$被称为由$A(t)$生成的$\beta$ -预解族.

注2.1  根据文献[17-18], $U_\beta(t, s)$有如下性质:

(1) $U_\beta(s, s)=I, U_\beta(t, s)=U_\beta(t, r)U_\beta(r, s)$, $0\leq s\leq r\leq t\leq a$.

(2) $(t, s)\rightarrow U_\beta(t, s)$是强连续的, $0\leq s\leq t\leq a$.

定义2.2   $u\in PC(J, E)$是方程(1.1)-(1.3)的一个温和解当且仅当$u\in PC(J, E)$是下面的脉冲积分方程的一个解

$\begin{eqnarray}u(t)&=&U_\beta(t, 0)u_0+\int_{0}^{t}U_\beta(t, s)f(s, u(\tau_1 (s)), u(\tau_2 (s)), \cdots , u(\tau_l (s))){\rm d}s\\ &&+\sum\limits_{0 <t_j <t} U_\beta(t, t_j)I_j(u(t _j)), t\in J.\end{eqnarray} $

3 主要结论

首先,我们给出本文的假设.

$(H_1)$存在非负常数$a_i\ (i=1, 2, \cdots , l), $$b_j\ (j=1, 2, \cdots , m)$和非负Lebesgue可积函数$g\in L(J, \Bbb R_+)$使得

$\begin{equation}\|f(t, u_1, u_2, \cdots , u_l)-f(t, v_1, v_2, \cdots , v_l)\|\leq g(t)\sum\limits_i^la_i\|u_i-v_i\|, \end{equation} $

$\begin{equation}\|I_j(u)-I_j(v)\|\leq b_j\|u-v\|, ~~ t\in J, u_i, v_i, u, v \in E \ (i=1, 2, \cdots , l, j=1, 2, \cdots , m);\end{equation} $

定理3.1  假设条件$(H_1)$$(H_2)$成立,则问题(1.1)-(1.3)存在一个唯一的温和解$u\in PC(J, E)$.对任意的$z_0\in PC(J, E)$,迭代列$\{z_n(t)\}$定义如下

$\begin{eqnarray}z_n(t)&=&U_\beta(t, 0)u_0+\int_{0}^{t}U_\beta(t, s)f(s, z_{n-1}(\tau_1 (s)), z_{n-1}(\tau_2 (s)), \cdots , z_{n-1}(\tau_l (s))){\rm d}s\\&&+\sum\limits_{0 <t_j <t} U_\beta(t, t_j)I_j(z_{n-1}(t _j)), t\in J, n=1, 2, \cdots , \end{eqnarray} $

$t\in J$一致收敛到$u(t)$,并且对任意实数$r>0$

$ \begin{equation} \|z_n-u\|_{PC}=o(\frac{1}{n^{r+1}}), ~~n\rightarrow +\infty.\end{equation} $

  定义算子

$\begin{eqnarray}Au(t)&=&U_\beta(t, 0)u_0+\int_{0}^{t}U_\beta(t, s)f(s, u(\tau_1 (s)), u(\tau_2 (s)), \cdots , u(\tau_l (s))){\rm d}s\\ &&+\sum\limits_{0 <t_j <t} U_\beta(t, t_j)I_j(u(t _j)).\end{eqnarray}$

显然,算子$A$$PC(J, E)$$PC(J, E)$.

$\begin{equation}L(s)=Mg(s)\sum\limits_i^la_i, ~~s\in J.\end{equation}$

由条件$(H_1)$可知$L\in L(J, \Bbb R_+)$.根据条件$(H_2)$,我们选取$\varepsilon>0$使得

$\begin{equation}\varepsilon+\sum\limits_{j=1}^m Mb_j <1.\end{equation}$

对上面的$\varepsilon>0$,由Lebesgue可积函数的性质,存在一个连续函数$\phi(s)$使得$\int_{0}^{T_0}|L(s)-\phi(s)|{\rm d}s <\varepsilon$.由条件$(H_1)$$(H_2)$和(3.5)式,对任意的$t\in J, u, v \in PC(J, E)$,我们有

$ \begin{eqnarray} &&\|(Au)(t)-(Av)(t)\|\\&\leq&\int_{0}^{t}M\|f(s, u(\tau_1 (s)), \cdots , u(\tau_l (s)))-f(s, v(\tau_1 (s)), \cdots , v(\tau_l (s)))\|{\rm d}s\\&&+\sum\limits_{0 <t_j <t} M\|I_j(u(t _j))-I_j(v(t _j))\|\\&\leq&\int_{0}^{t}Mg(s)\sum\limits_i^la_i\|u(\tau_i(s))-v(\tau_i(s))\|{\rm d}s+\sum\limits_{0 <t_j <t} Mb_j\|u(t _j)-v(t _j)\|\\&\leq&\bigg(\int_{0}^{t}L(s){\rm d}s+\sum\limits_{j=1}^m Mb_j\bigg)\|u-v\|_{PC}\\&\leq&\bigg(\int_{0}^{t}|L(s)-\phi(s)|{\rm d}s+\int_{0}^{t}|\phi(s)|{\rm d}s\bigg)\|u-v\|_{PC}+\sum\limits_{j=1}^m Mb_j\|u-v\|_{PC}\\&\leq&\bigg(\varepsilon+Nt+\sum\limits_{j=1}^m Mb_j\bigg)\|u-v\|_{PC}\leq(b+Nt)\|u-v\|_{PC}\\&=&\left(C_1^0b^1+C_1^1\frac{(Nt)^1}{1!}\right)\|u-v\|_{PC}, \end{eqnarray} $

这里$b=\varepsilon+\sum\limits_{j=1}^m Mb_j <1, N=\max\limits_{t\in J}|\phi(t)|.$

对任意的正整数$n$, $t\in J$,我们将证明下面的不等式

$\begin{equation}\|(A^nu)(t)-(A^nv)(t)\|\leq \left(C_n^0b^n+C_n^1b^{n-1}\frac{(Nt)^1}{1!}+\cdots +C_n^nb^{n-n}\frac{(Nt)^n}{n!}\right)\|u-v\|_{PC}, \end{equation}$

这里$C_n^m=n!/{(m!(n-m)!)}.$$n=1$时, (3.9)式成立.假设$n=k$时(3.9)式成立,即对任意的$t\in J$,有

$\begin{equation}\|(A^ku)(t)-(A^kv)(t)\|\leq \left(C_k^0b^k+C_k^1b^{k-1}\frac{(Nt)^1}{1!}+\cdots +C_k^kb^{k-k}\frac{(Nt)^k}{k!}\right)\|u-v\|_{PC}, \end{equation}$

则由(3.1), (3.2), (3.5), (3.8)式和公式$C_{k+1}^m=C_k^m+C_k^{m-1}$,对任意的$t\in J$,我们有

因此, (3.8)式对$n=k+1$成立.根据数学归纳法知对任意正整数$n$$t\in J$, (3.8)式成立.从而对任意的正整数$n$

$\begin{equation}\|A^nu-A^nv\|_{PC}\leq \left(C_n^0b^n+C_n^1b^{n-1}\frac{d^1}{1!}+\cdots +C_n^nb^{n-n}\frac{d^n}{n!}\right)\|u-v\|_{PC}, \end{equation}$

这里$d=NT_0$.容易看出

因此,我们选取充分大的正整数$K>2$使得

$ \begin{equation} \left(b^{K-1}K\left(\frac{K}{K-1}\right)^{K-1}\right)^{1/K}\equiv\alpha <1.\end{equation} $

对任意正整数$n>2K$,使得$n=Kh+p\ (0\leq p <K)$,这里$K$是上面给出的.显然, $h=[n/K] <[n/2]$.则对任意充分大的正整数$n>2K$,根据下面的Stirling公式

和(3.12)式可得

$\begin{eqnarray}S_1&\equiv &C_n^0b^n+C_n^1b^{n-1}\frac{d^1}{1!}+\cdots +C_n^hb^{n-h}\frac{d^h}{h!}\\&\leq &b^{n-h}C_n^h\left(1+d+\frac{d^2}{2!}+\cdots +\frac{d^h}{h!}\right)=b^{n-h}C_n^hO(1)\\ &=&\frac{O(1)b^{n-h}n^n\sqrt{2\pi n}(1+O(\frac{1}{h}))}{h^h\sqrt{2\pi h}\sqrt{2\pi(n-h)}(n-h)^{n-h}}\\ &=&O\left(\frac{K^h}{\sqrt{h}}\right)\left(\frac{bK}{K-1}\right)^{(n-h)h}=O\left(\frac{(b^{K-1}K(\frac{K}{K-1})^{K-1})^h}{\sqrt{h}}\right)\\ &=&O\left(\frac{\alpha^{Kh}}{\sqrt{h}}\right)=O\left(\frac{\alpha^{n}}{\sqrt{n}}\right).\end{eqnarray} $

另一方面,不失一般性,我们假设$d=NT_0>1$.根据Stirling公式和$C_n^{[n/2]}=O(2^n/\sqrt{n})$,我们可得

$\begin{eqnarray}S_2&\equiv &C_n^{h+1}b^{n-h-1}\frac{d^{h+1}}{(h+1)!}+\cdots +C_n^nb^{n-n}\frac{d^n}{h!}\\&\leq& \frac{1}{(h+1)!}C_n^{[n/2]}(b^{n-h-1}d^{h+1}+b^{n-n}d^n)\\&=&b^{n-h}C_n^hO(1)\\&=&\frac{O(\frac{2^n}{\sqrt{n}})e^{h+1}(b^{n-h-1}d^{h+1}+b^{n-n}d^n)}{\sqrt{2\pi (h+1)}(h+1)^{h+1}(1+O(\frac{1}{h+1}))}\\ &\leq&\frac{O(\frac{2^n}{\sqrt{n}})e^{h+1}(1+d+d^2+\cdots +d^{h+1}+\cdots +d^n)}{\sqrt{2\pi (h+1)}(h+1)^{h+1}}\\ &\leq &\frac{O(1)2^ne^{h+1}d^{n+1}}{\sqrt{n}\sqrt{h+1}(h+1)^{h+1}}\leq \frac{O(1)2^ne^{h+1}d^{n+1}}{h^{h+2}}\\ &=&o\left(\frac{1}{h^{r+1}}\right)=o\left(\frac{1}{n^{r+1}}\right) \ \ (n\rightarrow+\infty), \end{eqnarray} $

这里$r>0$是任意实数.由(3.11), (3.13)和(3.14)式可得

从而,对任意固定的$r>0$,存在一个正整数$n_0$使得对任意的$u, v\in PC(J, E)$

根据广义Banach压缩映像原理,算子$A$有唯一的不动点$u\in PC(J, E)$.因此, $u\in PC(J, E)$是问题(1.1)-(1.3)的唯一温和解.并且对任意的$z_0\in PC(J, E)$,由(3.3)式定义的迭代列$z_n=Az_{n-1}\ (n=1, 2, 3, \cdots ) $收敛到问题(1.1)-(1.3)的唯一温和解$u\in PC(J, E)$ (参见文献[19]中定理2.2.1的证明).

对任意的$r>0$,由$u$是算子$A$的唯一不动点和(3.3)式得

同上面的证明过程一样,我们可得

因此,对任意的$r>0$我们得到(3.4)式.

定理3.2  假设条件$(H_1)$$(H_2)$成立, $u(t), \overline{u}(t)\in PC(J, E)$分别是问题(1.1)-(1.3)和下面方程(3.15)-(3.17)的唯一解

$\begin{equation}{} ^{c}D_{t}^{\beta}u(t)=A(t)u(t)+f(t, u(\tau_1 (t)), u(\tau_2 (t)), \cdots , u(\tau_l (t))), \ \ t\in J, t\neq t_j, \end{equation}$

$\begin{equation}\Delta u(t_j)=I_j(u(t_j)), t=t_j, j=1, 2, \cdots , m, \end{equation}$

$\begin{equation}u(0)=\overline{u}_0 \in E, \end{equation}$

则存在一个常数$C>0$使得

$\begin{equation}\|\overline{u}-u\|_{PC}\leq C\|\overline{u}_0-u_0\|.\end{equation} $

  由定义2.2知$u(t)$满足

$\begin{eqnarray}u(t)&=&U_\beta(t, 0)u_0+\int_{0}^{t}U_\beta(t, s)f(s, u(\tau_1 (s)), u(\tau_2 (s)), \cdots , u(\tau_l (s))){\rm d}s\\ &&+\sum\limits_{0 <t_j <t} U_\beta(t, t_j)I_j(u(t _j)), t\in J.\end{eqnarray}$

类似的, $\overline{u}(t)$满足

$ \begin{eqnarray}\overline{u}(t)&=&U_\beta(t, 0)\overline{u}_0+\int_{0}^{t}U_\beta(t, s)f(s, \overline{u}(\tau_1 (s)), \overline{u}(\tau_2 (s)), \cdots , \overline{u}(\tau_l (s))){\rm d}s\\ &&+\sum\limits_{0 <t_j <t} U_\beta(t, t_j)I_j(\overline{u}(t _j)), t\in J.\end{eqnarray} $

类似于(3.8)式,根据(2.2), (3.20)式和条件$(H_1)$我们有

对任意正整数$n$,根据归纳法重复这一过程可得

由(3.13)和(3.14)式和上面的不等式,对$r>0$

上面不等式令$n\rightarrow \infty$,可得

由(3.13)和(3.14)式可知

这里$0 <\alpha <1, r>0$,无穷级数$\Sigma_{j=1}^\infty\{O(\alpha^j/\sqrt{j})+o({1}/{j^{r+1}})\}$是收敛的.因此,存在一个常数$C>0$使得

定理3.2证毕.

注3.1  由(3.9)式可知当$\overline{u}_0\rightarrow u_o$时有$\|\overline{u}\rightarrow u\|_{PC}\rightarrow0.$这意味着在条件$(H_1)$$(H_2)$成立时,问题(1.1)-(1.3)的唯一解是连续依赖于初值$\overline{u}_0$$u_0$.

4 小结

本文利用广义Banach不动点定理讨论了带迟滞和脉冲的分数阶非自治发展方程解的存在唯一性,给出其解的迭代序列及误差估计.最后讨论了初值扰动时解的扰动问题.可以证明其解是连续依赖于初值的.

参考文献

Arthi G , Park J H , Jung H Y .

Existence and exponential stability for neutral stochastic integro-differential equations with impulses driven by a fractional Brownian motion

Commun Nonlinear Sci Numer Simul, 2016, 32: 145- 157

DOI:10.1016/j.cnsns.2015.08.014      [本文引用: 1]

Chauhan A , Daba J .

Local and global existence of mild solution to an impulsive fractional functional integro-differential equation with nonlocal condition

Commun Nonlinear Sci Numer Simul, 2014, 19 (4): 821- 829

DOI:10.1016/j.cnsns.2013.07.025     

Chadha A , Pandey D N .

Existence results for an impulsive neutral stochastic fractional integro-differential equation with infinite delay

Nonlinear Anal, 2015, 128: 149- 175

DOI:10.1016/j.na.2015.07.018     

Dabas J , Chauhan A .

Existence and uniqueness of mild solution for an impulsive neutral fractional integrodifferential equations with infinity delay

Math Comput Modell, 2013, 57 (3): 754- 763

URL    

Ge F D , Zhou H C , Kou C H .

Approximate controllability of semilinear evolution equations of fractional order with nonlocal and impulsive conditions via an approximating technique

Appl Math Comput, 2016, 275: 107- 120

URL    

Lin Z , Wang J R , Wei W .

Multipoint BVPs for generalized impulsive fractional differential equations

Appl Math Comput, 2015, 258: 608- 616

URL    

Pierri M , O'Regan D , Rolnik V .

Existence of solutions for semilinear differential equations with not instantaneous impulses

Appl Math Comput, 2013, 219: 6743- 6749

URL    

Tomar N K , Dabas J .

Controllability of impulsive fractional order semilinear evolution equations with nonlocal conditions

J Non Evol Equ Appl, 2012, 5: 57- 67

URL    

Yang X J , Li C D , Huang T W , Song Q K .

Mittag-Leffler stability analysis of nonlinear fractional-order systems with impulses

Appl Math Comput, 2017, 293: 416- 422

URL    

Yan Z M .

Existence of solutions for nonlocal impulsive partial functional integrodifferential equations via fractional operators

J Comput Appl Math, 2011, 235 (8): 2252- 2262

DOI:10.1016/j.cam.2010.10.022     

Zhang G L , Song M H , Liu M Z .

Exponential stability of the exact solutions and the numerical solutions for a class of linear impulsive delay differential equations

J Comput Appl Math, 2015, 285: 32- 44

DOI:10.1016/j.cam.2015.01.034      [本文引用: 1]

Chen P Y , Zhang X P , Li Y X .

Study on fractional non-autonomous evolution equations with delay

Comput Math Appl, 2017, 73 (5): 794- 803

DOI:10.1016/j.camwa.2017.01.009      [本文引用: 1]

Ouyang Z G .

Existence and uniqueness of the solutions for a class of nonlinear fractional order partial differential equations with delay

Comput Math Appl, 2011, 61 (4): 860- 870

DOI:10.1016/j.camwa.2010.12.034      [本文引用: 3]

Zhu B , Liu L S , Wu Y H .

Local and global existence of mild solutions for a class of nonlinear fractional reaction-diffusion equations with delay

Appl Math Lett, 2016, 61: 73- 79

DOI:10.1016/j.aml.2016.05.010      [本文引用: 1]

Zhu B , Liu L S , Wu Y H .

Existence and uniqueness of global mild solutions for a class of nonlinear fractional reaction-diffusion equations with delay

Comput Math Appl, 2016

DOI:10.1016/j.camwa.2016.01.028      [本文引用: 2]

Araya D , Lizama C .

Almost automorphic mild solutions to fractional differential equations

Nonlinear Anal, 2008, 69 (11): 3692- 3705

DOI:10.1016/j.na.2007.10.004      [本文引用: 1]

Debbouche A , Baleanu D .

Controllability of fractional evolution nonlocal impulsive quasilinear delay integro-differential systems

Comput Math Appl, 2011, 62 (3): 1442- 1450

DOI:10.1016/j.camwa.2011.03.075      [本文引用: 2]

Pazy A . Semigroups of Linear Operators and Applications to Partial Differential Equations. New York: Applied Mathematical Sciences, 1983

[本文引用: 1]

Guo D J , Lakshamikantham V , Liu X Z . Nonlinear Integral Equations in Abstract Spaces. Dordrecht: Kluwer Academic, 1996

[本文引用: 1]

/