数学物理学报  2015, Vol. 35 Issue (5): 867-877   PDF (365 KB)    
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
魏利
Agarwal Ravi P
含有广义p-Laplacian算子的双曲型非线性微分方程的单调性方法
魏利[0, 1], Agarwal Ravi P2,3    
1 河北经贸大学数学与统计学院 石家庄 050061;
2 Department of Mathematics, Texas A & M University-Kingsville, Kingsville, TX 78363, USA;
3 Department of Mathematics, Faculty of Science, King Abdulaziz University, Jeddah 21589, Saudi Arabia
摘要: 利用伪单调算子和极大单调算子值域的扰动结果,得到了含有广义p-Laplacian算子、具混合边值条件的双曲型非线性微分方程存在唯一解的抽象结论,是对含有广义p-Laplacian算子的非线性椭圆方程和抛物方程相关研究工作的推广.运用了一些新的证明技巧.而且,展示了这个唯一解与某极大单调算子零点之间的关系.
关键词: 伪单调算子     p-Laplacian算子     次微分     双曲型非线性微分方程     混合边界    
Monotonicity Method for Hyperbolic Type Nonlinear Differential Equation with Generalized p-Laplacian
Wei Li[0, 1], Agarwal Ravi P2,3    
1 School of Mathematics and Statistics, Hebei University of Economics and Business, Shijiazhuang 050061;
2 Department of Mathematics, Texas A & M University-Kingsville, Kingsville, TX 78363, USA;
3 Department of Mathematics, Faculty of Science, King Abdulaziz University, Jeddah 21589, Saudi Arabia
Abstract: By using the results on the perturbations of the ranges for pseudo-monotone operators and maximal monotone operators, we present the abstract result for the existence of the unique solution of hyperbolic type nonlinear differential equation involving the generalized p-Laplacian with mixed boundaries. It is an extension of the corresponding studies on nonlinear elliptic equation and nonlinear parabolic equation with generalized p-Laplacian. Some new techniques are employed in the discussion. Moreover, the connections between the unique solution studied in this paper and the zero point of a maximal monotone operator is exemplified.
Key words: Pseudo-monotone operator     p-Laplacian     Subdifferential     Hyperbolic type nonlinear differential equation     Mixed boundaries    
1引言和预备知识

本文假设$X$为实Banach空间,其对偶空间$X' $严格凸. 用$"\rightarrow"$和$"w-\lim"$分别表示空间中的强、弱收敛. 对$X$中任意子集$G$,分别用$int G$和$\overline G$表示其内部和强闭包. 称$X$中子集$G_1$和$G_2$几乎相等:若 $\overline{G_1}= \overline{G_2}$且$int G_1 = int G_2,$并记之为$G_1 \simeq G_2.$称 $T:X \rightarrow X'$为$X$上的hemi连续映射[1]:若 $w-\lim\limits_{t \rightarrow 0}T(x+ty) =Tx,$ $\forall x,y \in X.$

称函数$\Phi$为$X$上的正则凸函数[1]: 若$\Phi : X \rightarrow (-\infty,+\infty]$,不恒等于 $+\infty$且满足$\Phi((1-\lambda)x+\lambda y)\leq (1-\lambda)\Phi(x)+\lambda \Phi(y),$ $\forall x ,y \in X$及 $0 \leq \lambda \leq 1.$ 称函数$\Phi: X \rightarrow (-\infty,+\infty]$为$X$上的下半连续函数[1]: 若 $\liminf\limits_{y \rightarrow x}\Phi(y)\geq \Phi(x),$ $\forall x \in X$.

对$X$上的正则凸函数$\Phi$和$x \in X,$定义$\partial \Phi(x)$为

$$\partial \Phi(x)= \{ x^* \in X' : \Phi(x)\leq \Phi(y)+ (x-y,x^*),y \in X\}. $$

并称$x^*$为$\Phi$在$x$点的次梯度,$\partial \Phi(x)$ 为$\Phi$在$x$点的次微分[1].

正规对偶映射$J :X \rightarrow 2^{X' }$ 定义为

$$ J(x)=\{f\in X' : (x,f) = \|x\|\cdot\|f\|,\,\|f\|=\|x\|\}, \quad x \in X, $$

其中 $(\cdot,\cdot)$表示$X$和$X' $中元素的广义对偶对.因$X$严格凸,故$J$为单值的[2].

令$A:X\rightarrow 2^{X' }$为多值映射,$A$的图像$G(A)$定义为

$$G(A)= \{[u,w]|u\in D(A),w\in Au \}. $$

称$A:X\rightarrow 2^{X' }$为单调算子[2]:若$G(A)$为$X\times X' $中单调集,即:$ (u_{1}-u_{2},w_{1}-w_{2})\ge 0 ,$ $\forall [u_{i},w_{i}]\in G(A),i= 1,2.$称$A$为严格单调算子:若$(u_{1}-u_{2},w_{1}-w_{2})\ge 0$ 中等号成立的充要条件是$u_1 = u_2.$ 单调算子 $A$称为极大单调的:若 $G(A)$不真含于$X\times X' $的任何单调子集中. $A$为极大单调算子当且仅当 $R(A + \lambda J) = X' ,$ $\forall \lambda > 0.$ 称$A$为强迫的:若 $ \lim\limits_{n\rightarrow +\infty}\frac{(x_{n},x^*_{n})}{\|x_{n}\|}= \infty,$ $\forall [x_{n},x^*_{n}]\in G(A)$具 $ \lim\limits_{n\rightarrow +\infty } \|x_{n}\|= +\infty$.

令$A: X \rightarrow 2^{X' } $为极大单调算子且$[0, 0] \in G(A),$则方程$J(u_t-u)+ t Au_t \ni 0$有唯一解$u_t \in D(A),$ $\forall u \in X$和$t > 0.$ $A$的豫解式$J_t^A$和Yosida逼近$A_t$分别定义为[2]

$$ J_t^A u = u_t; \qquad A_t u =-\frac{1}{t}J(u_t-u),$$ $\forall u \in X$和 $\forall t > 0$ (因此,$[J_t^A u,A_t u]\in G(A)$).

定义 1.1 (Pascali和Sburlan[2]) 令$C$为$X$的闭凸子集,$A: C \rightarrow 2^{X' }$为多值映射.称$A$为伪单调算子:若 (i) $\forall x \in C,$ $Ax$为$X' $中非空闭凸子集; (ii)若$C$中序列$\{x_n\}$弱收敛到$x \in C ,$ $f_n \in Ax_n$且满足 $\limsup\limits_{n\rightarrow \infty}(x_n-x,f_n) \leq 0,$则 $\forall y \in C,$ 存在$f(y) \in Ax $满足$(x- y,f(y))\leq \liminf\limits_{n \rightarrow \infty} (x_n-x,f_n);$ (iii) 对$X$的任意有限维子空间$F$,算子$A: C \bigcap F \rightarrow X' $在弱拓扑意义下连续.

引理 1.1 (Pascali和Sburlan[2]) 若 $A : X \rightarrow X'$极大单调且$D(A) = X,$则$A$伪单调.

引理 1.2 (Pascali和Sburlan[2]) 若 $A: X \rightarrow 2^{X' }$处处有定义、单调且hemi连续,则$A$ 极大单调.进一步,若$A$还是强迫的,则$R(A)= X' .$

引理 1.3 (Pascali和Sburlan[2]) 若$\Phi : X \rightarrow (-\infty,+\infty]$ 为正则、凸、下半连续函数,则$\partial \Phi : X \rightarrow X' $极大单调.

引理 1.4 (Pascali和Sburlan[2]) 若 $A_1$和 $ A_2$均为$X$中极大单调算子且$(intD(A_1)) \bigcap $ $D(A_2)\neq \emptyset,$则$A_1 + A_2$极大单调.

引理 1.5 (Lions[3]) 假设$X$及其对偶空间$X' $均为严格凸Banach空间, $S: D(S) \subset X \rightarrow X' $为闭线性算子且$S^*$为$S$的对偶算子.若$(u,Su) \geq 0$,$\forall u \in D(S)$且$(v, S^*v) \geq 0$,$\forall v \in D(S^*),$则$S$为极大单调算子且具有稠密定义域.

定理 1.1 (Calvert和Gupta[4])假设$X$为实Banach空间, 其对偶空间$X' $严格凸. $J :X\rightarrow X' $为$X$ 上对偶映射且满足:存在函数 $\eta : X \rightarrow [0,+\infty)$使得 $\forall u,v \in X,$

$$\|Ju-Jv\| \leq \eta(u-v). $$ (1.1)

假设$A, B_{1}:X\rightarrow 2^{X} $为增生映射且满足

(i)或者$A$和$C_{1}$均满足以下条件$(1.2)$或者$D(A)\subset D(C_{1})$ 且 $C_{1}$满足以下条件 $(1.2),$ 即 对$u\in D(A),v \in Au,$存在常数$C(a,f)$使得

$$ (v-f,J(u-a)) \geq C(a,f); $$ (1.2)

(ii) $A+C_{1}$为m增生且有界逆紧映射.假设$C_{2}:X\rightarrow X$为有界、 连续映射且满足 $\forall y\in X$,存在常数 $C(y)$使得 $ (C_{2}(u+y),Ju)\ge-C(y),$ $\forall u\in X$.那么有以下结论成立

(a) $\overline {[R(A)+R(C_{1})]}\subset \overline {R(A+C_{1}+C_{2})}$.

(b) $int[R(A)+R(C_{1})]\subset intR(A+C_{1}+C_{2})$.

定理 1.2 (Brezis[5]) 假设$T: X \rightarrow X' $为有界伪单调算子,$K$为$X$中 闭凸子集,$\Phi$为$K$上定义的下半连续凸函数、 不恒等于$ +\infty$并满足$\forall v \in K,\Phi(v) \in (-\infty,+\infty].$ 若存在$v_0 \in K$具$\Phi(v_0)<+\infty,$满足

$$ \frac{(v-v_0,Tv) + \Phi(v)}{\|v\|} \rightarrow \infty, $$

当$\|v\|\rightarrow \infty,v \in K,$则存在$u \in K$满足

$$(u-v,Tu) \leq \Phi(v)- \Phi(u),~\forall v \in K.$$

定理 1.3 (Chen和Luo[6]) 假设$X$为实自反Banach空间, $X$和$X' $均严格凸.令 $S :D(S) \subset X\rightarrow X' $为线性极大单调算子且 $T: X \rightarrow X' $为伪单调、强迫算子. 那么$\forall f \in X' ,$存在$u \in D(S)$满足 $S u + T u = f.$

定理 1.4 (Reich[7]) 假设$X$为实自反Banach空间,$X$和$X' $均严格凸.令$J$ 为正规对偶映射,$A$和$B$为$X$中极大单调算子.若存在$0 \leq k <1$及$C_1, C_2 > 0 $满足

$$(a,J^{-1}(B_t v))\geq-k \|B_t v\|^2-C_1 \|B_t v\|-C_2,$$ (1.3)

$\forall v \in D(A),a \in Av$及$t > 0,$其中 $B_t$为$B$的Yosida逼近. 则 $R(A)+R(B)\simeq R(A+B)$.

含有广义$p$-Laplacian算子的非线性椭圆边值问题与实际问题密切相关, 所以对这类问题的研究吸引了数学家们的注意,见文献[8, 9, 10, 11, 12]等. 2008年,利用定理 1.1,Wei和Agarwal[12] 研究了以下含有广义$p$-Laplacian算子的非线性椭圆边值问题

$$ \left\{ \begin{array}{ll} -{\rm div}[(C(x)+|Du|^2)^{\frac{p-2}{2}}\nabla u] +\varepsilon |u|^{q-2}u+g(x,u(x))= f(x),~&\mbox{a.e. }~x \in \Omega ,\\ -\langle \vartheta,(C(x)+|Du|^2)^{\frac{p-2}{2}}Du \rangle \in \beta_{x}(u(x)),&\mbox{a.e. } x\in \Gamma, \end{array} \right. $$ (1.4)

其中$D:= (\frac{\partial }{\partial x_1}, \frac{\partial }{\partial x_2},\cdots,\frac{\partial }{\partial x_N})$,$x = (x_1,x_2,\cdots,x_N)\in \Omega,$ $0\leq C(x)\in L^p(\Omega)$,$\varepsilon$为非负常数且 $\vartheta$为$\Gamma$的外法向导数.文献[12]证明了问题(1.4) 在$L^s(\Omega)$中存在解,其中$\frac{2N}{N+1} < p\leq s <+\infty,~1\leq q<+\infty$当$p\geq N,$ $1\leq q \leq \frac{Np}{N-p}$当 $p

2010年,Wei,Agarwal和Wong[13] 把对椭圆边值问题的研究推广为以下含有广义$p$-Laplacian算子、具有混合边值条件的 非线性抛物方程

$$ \left\{ \begin{array}{ll} \frac{\partial u}{\partial t}-{\rm div}[(C(x,t)+|Du|^2)^{\frac{p-2}{2}}Du] + \varepsilon |u|^{p-2}u = f(x,t),~& (x,t)\in \Omega \times (0,T), \\[2mm] -\langle \vartheta,(C(x,t)+|D u|^2)^{\frac{p-2}{2}}Du \rangle \in \beta(u)-h(x,t),& (x,t)\in \Gamma \times (0,T),\\ u(x,0)= u(x,T) ,~~~&\mbox{ a.e. } x \in \Omega. \end{array} \right. $$ (1.5)

采用了将方程(1.5)分解成以下两个辅助方程的技巧: (i)具有 Dirichlet边值条件的抛物方程

$$ \left\{ \begin{array}{ll} \frac{\partial u}{\partial t}- {\rm div}[(C(x,t)+|Du|^2)^{\frac{p-2}{2}}Du] + \varepsilon |u|^{p-2}u = f(x,t),~& (x,t)\in \Omega \times (0,T), \\[2mm] \gamma u = w,~ & (x,t)\in \Gamma \times (0,T),\\ u(x,0)= u(x,T) ,~~~& \mbox{ a.e. }\ x \in \Omega. \end{array} \right. $$ (1.6)

和(ii)具有Neumann边值条件的抛物方程

$$ \left\{ \begin{array}{ll} \frac{\partial u}{\partial t}-{\rm div}[(C(x,t)+|Du|^2)^{\frac{p-2}{2}}Du] + \varepsilon |u|^{p-2}u = f(x),~& (x,t)\in \Omega \times (0,T), \\ -\langle \vartheta,(C(x,t)+|Du|^2)^{\frac{p-2}{2}}Du \rangle \in \beta(u)-h(x,t), & (x,t)\in \Gamma \times (0,T). \end{array} \right. $$ (1.7)

利用定理1.2和1.3,文献[13]证明了方程$(1.6)$存在唯一解.利用定理1.4,证明了方程(1.7)在 $L^p(0,T; $ $W^{1,p}(\Omega))$中存在唯一解,从而方程(1.5) 在$L^p(0,T; W^{1,p}(\Omega))$中存在唯一解,其中$2 \leq p <+\infty$.然而验证不等式(1.3)成立并不是一件容易的事情.

能否借鉴处理椭圆方程和抛物方程的思想处理双曲型方程? $|u|^{q-2}u$能否被更一般化的函数 $g(x,u,\frac{\partial u}{\partial t},Du)$所取代 ?能否将$2 \leq p < +\infty $推广为$\frac{2N}{N+1} < p < +\infty$,$N \geq 1$?能否简化证明过程?

基于以上问题,本文将研究以下含有广义$p$-Laplacian算子、具有混合边值条件的双曲型方程

$$ \left\{ \begin{array}{ll} \frac{\partial^2 u}{\partial t^2}+{\rm div}[(C(x,t)+|D u|^2)^{\frac{p-2}{2}}Du]-g(x,u,\frac{\partial u}{\partial t }, Du) = f(x,t),&(x,t) \in \Omega \times (0,T),\\ -\langle \vartheta,(C(x,t)+|D u|^2)^{\frac{p-2}{2}}Du \rangle \in \beta_{x}(u(x,t)),~~ &(x,t) \in \Gamma \times (0,T),\\[2mm] u(x,0) = u(x,T),& x \in \Omega,\\[2mm] \frac{\partial u}{\partial t}(x,0) = \frac{\partial u}{\partial t}(x,T),& x \in \Omega. \end{array} \right. $$ (1.8)
2主要结论

以下假设$ N \geq 1,$ $\frac{2N}{N+1} < p < +\infty,$ $\frac{1}{p}+\frac{1}{p'}=1$. 在方程(1.8)中,$\Omega$为$R^{N}$中的有界锥形区域且其边界$\Gamma\in C^{1}$ (见文献[10]). $D:= (\frac{\partial }{\partial x_1},\frac{\partial }{\partial x_2},\cdots,\frac{\partial }{\partial x_N})$,$x = (x_1,x_2, \cdots,x_N)\in \Omega,$ $T$为正常数,$\vartheta$为$\Gamma$的外法向导数. $0 \leq C(x,t) \in L^p(\Omega \times (0,T))$且$f(x,t) \in L^{p'}(0,T; (W^{1,p}(\Omega))').$假设Green公式成立.

$\beta_{x}$为 $ \varphi _{x}$的次微分(i.e. $ \beta_x = \partial\varphi _{x}$),其中 $\varphi _{x}= \varphi ( x,\cdot):R\rightarrow R$是正则、凸、下半连续函数具$\varphi _{x}(0) =0,$ $\forall x\in \Gamma,$ $\varphi : \Gamma \times R\rightarrow R$为一给定函数.假设$0\in \beta_{x}(0) $且 $\forall t\in R,$函数 $x\in \Gamma \rightarrow (I+\lambda \beta _{x})^{-1}(t)\in R $对 $\forall \lambda > 0$可测. 假设$g:\Omega \times R^{N+2} \rightarrow R$满足以下条件

(a) Carathéodory条件.

映射$ x \rightarrow g(x,r)$在$\Omega$上可测,$\forall r \in R^{N+2};$映射$ r \rightarrow g(x,r)$在$R^{N+2}$上连续,$\forall x\in \Omega;$

(b)增长性条件.

$$g(x,r_1,\cdots,r_{N+2})\leq h(x,t)+ b |r_1|^{\frac{p}{p'}}, $$其中$ (r_1,r_2,\cdots,r_{N+2})\in R^{N+2} ,$ $h(x,t)\in L^{p'}(0,T; (W^{1,p}(\Omega))')$且 $ b$为正常数;

(c)单调性条件. $g$关于$r_1$单调,即

$$ (g(x,r_1,\cdots,r_{N+2})-g(x,t_1,\cdots,t_{N+2}))(r_1-t_1) \geq 0,$$

$\forall x \in \Omega,$$(r_1,\cdots,r_{N+2}),(t_1,\cdots,t_{N+2})\in R^{N+2}.$

引理 2.1定义映射$\Phi : L^p(0,T; W^{1,p}(\Omega))\rightarrow R$为

$$\Phi(u) = \int_0^T \int_\Gamma \varphi_x(u|_\Gamma(x,t)){\rm d}\Gamma(x){\rm d}t,$$

$\forall u \in L^p(0,T; W^{1,p}(\Omega)),$则$\Phi$是$L^p(0,T; W^{1,p}(\Omega))$上的正则凸、 下半连续函数.进而引理1.3蕴含 $\partial \Phi$极大单调.

类似于文献[4,引理3.1],可知结论成立.

引理 2.2 (Wei和Agarwal[12]) 令$X_0$

为$W^{1,p}(\Omega)$中所有常值函数构成的闭线性子空间.令$X$ 为商空间$W^{1,p}(\Omega)/X_0.$对$u\in W^{1,p}(\Omega),$定义 $P:W^{1,p}(\Omega) \rightarrow X_0$为 $Pu=\frac{1}{meas(\Omega)}\int_\Omega u {\rm d}x.$ 则存在常数$C>0$满足 $\forall u\in W^{1,p}(\Omega),$

$$\|u-Pu\|_{p}\leq C\|Du\|_{(L^{p}(\Omega))^N}.$$

引理 2.3 定义映射$B: L^p(0,T; W^{1,p}(\Omega))\rightarrow L^{p'}(0,T;(W^{1,p}(\Omega))' )$为

$$(w,Bu) = \int_0^T\int_{\Omega}\langle (C(x,t)+|Du|^2)^{\frac{p-2}{2}}Du,Dw \rangle {\rm d}x{\rm d}t, $$

$\forall u,w\in L^p(0,T; W^{1,p}(\Omega)).$ 则$B$严格单调、伪单调且强迫.

这里 $\langle \cdot,\cdot \rangle$和 $|\cdot|$分别表示$R^N$中的内积和范数.

类似于文献[13,引理5.1]可证,$B$处处有定义、严格单调且hemi连续. 因此引理1.1和1.2蕴含$B$极大单调且伪单调.

鉴于$B$中不含有$\int_0^T\int_{\Omega}|u|^{p-2}uv{\rm d}x{\rm d}t,$ 即不同于文献[13,引理5.1],下面证明$B$是强迫的.

为此首先证明对$u\in L^p(0,T; W^{1,p}(\Omega)),$

\begin{equation} \|u\|_{L^p(0,T; W^{1,p}(\Omega))}\leq k_1 \bigg(\int_0^T\int_{\Omega}|Du|^p {\rm d}x{\rm d}t\bigg)^{\frac{1}{p}}+ k_2, \end{equation} (2.1)

其中$k_1$和$k_2$为正常数.

事实上,由引理2.2知对$u \in L^p(0,T; W^{1,p}(\Omega)),$

$$ \bigg\|u- \frac{1}{meas(\Omega)}\int_\Omega u {\rm d}x\bigg\|_{L^p(\Omega)} \leq C\bigg(\int_{\Omega}|Du|^p {\rm d}x\bigg)^{\frac{1}{p}},$$

其中$ C $为正常数.

因此

\begin{eqnarray*} && \bigg\|u-\frac{1}{meas(\Omega)}\int_\Omega u {\rm d}x\bigg\|^p_{W^{1,p}(\Omega)} \\ &=& \bigg\|u-\frac{1}{meas(\Omega)}\int_\Omega u {\rm d}x\bigg\|^p_{L^{p}(\Omega)} + \bigg\|D(u-\frac{1}{meas(\Omega)}\int_\Omega u {\rm d}x)\bigg\|^p_{(L^{p}(\Omega))^N} \\\ &\leq & (C^p +1)\int_{\Omega}|Du|^p {\rm d}x.\end{eqnarray*}

因为

$$\bigg\|u- \frac{1}{meas(\Omega)}\int_\Omega u {\rm d}x\bigg\|_{W^{1,p}(\Omega)} \geq \|u\|_{W^{1,p}(\Omega)}-\bigg\|\frac{1}{meas(\Omega)}\int_\Omega u {\rm d}x\bigg\|_{W^{1,p}(\Omega)},$$

所以

$$ \|u\|_{W^{1,p}(\Omega)}\leq \bigg\|u- \frac{1}{meas(\Omega)}\int_\Omega u {\rm d}x\bigg\|_{W^{1,p}(\Omega)}+Const. $$

从而

\begin{eqnarray*} \|u\|_{L^p(0,T; W^{1,p}(\Omega))} &\leq & \bigg\|u-\frac{1}{meas(\Omega)}\int_\Omega u {\rm d}x\bigg\|_{L^p(0,T; W^{1,p}(\Omega))}+ k_2\\ & \leq & (C^p +1)^{\frac{1}{p}}\bigg(\int_0^T\int_{\Omega}|Du|^p {\rm d}x{\rm d}t\bigg)^{\frac{1}{p}}+ k_2. \end{eqnarray*}

若令$k_1 = (C^p + 1)^{\frac{1}{p}}$,则(2.1)式成立.

因此对 $u\in L^p(0,T; W^{1,p}(\Omega)),$令 $\|u\|_{L^p(0,T; W^{1,p}(\Omega))}\rightarrow +\infty,$由(2.1)式有

$$ \frac{(u,Bu)}{\|u\|_{L^p(0,T; W^{1,p}(\Omega))}}\geq \frac{\int_0^T\int_{\Omega}|Du|^p {\rm d}x{\rm d}t} {\|u\|_{L^p(0,T; W^{1,p}(\Omega))}} \rightarrow +\infty. $$

由此可知$B$为强迫的,进而结论成立.

定义 2.1 定义$S: D(S) = \{ u\in L^p(0,T; W^{1,p}(\Omega)): \frac{\partial u}{\partial t}\in L^{p'}(0,T; (W^{1,p}(\Omega))'),u(x,0) = u(x,T),\frac{\partial^2 u}{\partial t^2} \in L^{p'}(0,T; (W^{1,p}(\Omega))'), \frac{\partial u}{\partial t}(x,0) = \frac{\partial u}{\partial t}(x,T)\} \rightarrow L^{p'}(0,T; (W^{1,p}(\Omega))')$为

$$Su(x,t) =-\frac{\partial^2 u}{\partial t^2}.$$

引理 2.4 映射 $S$线性极大单调.

$\forall u(x,t),w(x,t) \in D(S),$分部积分有

\begin{eqnarray*} && (w,Su) + (u,\frac{\partial^2 w}{\partial t^2}) \\ & =&-\int_0^T\int_\Omega \frac{\partial^2 u}{\partial t^2}w(x,t) {\rm d}x{\rm d}t + \int_0^T\int_\Omega u(x,t) \frac{\partial^2 w}{\partial t^2} {\rm d}x{\rm d}t \\ & =&-\int_\Omega \frac{\partial u}{\partial t}(x,T)w(x,T) {\rm d}x + \int_\Omega \frac{\partial u}{\partial t}(x,0)w(x,0) {\rm d}x + \int_0^T\int_{\Omega}\frac{\partial u}{\partial t}\frac{\partial w}{\partial t}{\rm d}x{\rm d}t \\ && + \int_\Omega \frac{\partial w}{\partial t}(x,T)u(x,T) {\rm d}x- \int_\Omega \frac{\partial w}{\partial t}(x,0)u(x,0) {\rm d}x- \int_0^T\int_{\Omega}\frac{\partial u}{\partial t}\frac{\partial w}{\partial t}{\rm d}x{\rm d}t = 0. \end{eqnarray*}

因此$S^*w =-\frac{\partial^2 w}{\partial t^2},$其中 $D(S^*) = \{w\in L^p(0,T; W^{1,p}(\Omega)): \frac{\partial w}{\partial t}\in L^{p'}(0,T; (W^{1,p}(\Omega))'),$ $ w(x,0) = w(x,T), \frac{\partial^2 w}{\partial t^2} \in L^{p'}(0,T; (W^{1,p}(\Omega))'),\frac{\partial w}{\partial t}(x,0) = \frac{\partial w}{\partial t}(x,T)\} .$

对$u(x,t) \in D(S),$

\begin{eqnarray*} &&\int_0^T\int_\Omega \frac{\partial^2 u}{\partial t^2}u(x,t) {\rm d}x{\rm d}t \\ & =& \int_\Omega u(x,T)\frac{\partial u}{\partial t}(x,T) {\rm d}x- \int_\Omega u(x,0)\frac{\partial u}{\partial t}(x,0) {\rm d}x- \int_0^T\int_\Omega \bigg(\frac{\partial u}{\partial t}\bigg)^2 {\rm d}x{\rm d}t\\ & =&-\int_0^T\int_\Omega\bigg (\frac{\partial u}{\partial t}\bigg)^2 {\rm d}x{\rm d}t \leq 0. \end{eqnarray*}

从而$(u,Su) \geq 0.$类似地,$(w,S^*w) \geq 0,$ $w \in D(S^*).$ 于是引理1.5蕴含$S$线性极大单调.

引理 2.5定义映射$C : L^p(0,T; W^{1,p}(\Omega)) \rightarrow L^{p'}(0,T;(W^{1,p}(\Omega))' )$为

$$ (v,Cu) = \int_0^T\int_\Omega g(x,u,\frac{\partial u}{\partial t},Du)v(x,t) {\rm d}x{\rm d}t, $$

$\forall u(x,t),v(x,t) \in L^p(0,T; W^{1,p}(\Omega)),$则$C$处处有定义、单调且 hemi连续.进而引理1.2蕴含$C$极大单调.

第1步.对$u(x,t) \in L^p(0,T; W^{1,p}(\Omega)),$ $x \rightarrow g(x,u,\frac{\partial u}{\partial t},Du)$在 $\Omega$上可测.

因$u(x,t),\frac{\partial u}{\partial x_i} \in L^p(\Omega),i = 1 ,2,\cdots,N,$故$x \rightarrow (u,\frac{\partial u}{\partial x_1},\cdots, \frac{\partial u}{\partial x_N} )$在$\Omega$上可测.又因 $g$满足Carathéodory条件,故$x \rightarrow g(x,u,\frac{\partial u}{\partial t},Du)$在$\Omega$上可测.

第2步. $C$处处有定义.

对$ u,v\in L^p(0,T; W^{1,p}(\Omega)),$有

\begin{eqnarray*} |(v,Cu)|& \leq & \int_0^T\int_{\Omega}|h(x,t)||v(x,t)|{\rm d}x{\rm d}t + b \int_0^T\int_{\Omega}|u(x,t)|^{\frac{p}{p'}}|v(x,t)|{\rm d}x{\rm d}t \\ & \leq & (\|h(x,t)\|_{L^{p'}(0,T; (W^{1,p}(\Omega))')}+ b \|u\|^{\frac{p}{p'}}_{L^p(0,T; W^{1,p}(\Omega))})\|v\|_{L^p(0,T; W^{1,p}(\Omega))}, \end{eqnarray*}

从而$C$处处有定义.

第3步. $C$是hemi连续的. 只需证明对$u,v,w\in L^p(0,T; W^{1,p}(\Omega))$和$s\in [0,1],$ $(w,C(u+sv)-Cu) \rightarrow 0,$当$s\rightarrow 0$.

因为$g$在$\Omega$上可测且满足Carathéodory条件,所以利用Lebesgue控制收敛定理有

\begin{eqnarray*} 0 &\leq & \lim_{s \rightarrow 0}|(w,C(u+sv)-Cu)| \\ &\leq& \int_0^T\int_{\Omega}\lim_{s \rightarrow 0}\bigg|g\bigg(x,u+sv, \frac{\partial u}{\partial t} +s\frac{\partial v}{\partial t},Du + s Dv\bigg)-g\bigg (x,u,\frac{\partial u}{\partial t},Du\bigg)\bigg||w| {\rm d}x{\rm d}t\\ & =& 0, \end{eqnarray*}

于是$C$是hemi连续的.

第4步.因为$g(x,r_1,\cdots ,r_{N+2})$关于 $r_1$单调,所以$C$单调.

引理 2.6 $\forall u,v \in L^{p}(0,T; W^{1,p}(\Omega)),$

$$ (v, \partial\Phi(u)) = \int_0^T\int_\Omega \beta_x(u|_\Gamma(x,t)) v |_\Gamma(x,t) {\rm d}\Gamma(x){\rm d}t. $$

而且 $0 \in \partial\Phi(0).$

主要证明思想来源于文献[4,命题3.2(ii)].为文章的完整性, 证明如下.

定义映射$G:L^{p}(0,T;L^{p}(\Gamma))\rightarrow L^{p'}(0,T; L^{p'}(\Gamma))$为$Gu = \beta_{x}(u) ,$ $\forall u\in L^{p}(0,T; L^{p}(\Gamma))$.定义映射 $K:L^{p}(0,T; W^{1,p}(\Omega))\rightarrow L^{p}(0,T; L^{p}(\Gamma)) $为 $ K(v)= v|_{\Gamma}$,$ \forall v\in L^{p}(0,T; W^{1,p}(\Omega)).$于是$K^*GK = \partial \Phi,$这里$\Phi$同于引理 2.1.

事实上,易知 $G$连续.对 $\forall u(x,t),v(x,t) \in L^{p}(0,T; L^p(\Gamma)),$因$\beta_x$单调,故 $(u-v,Gu-Gv)= \int_0^T\int_\Gamma (\beta_x(u)-\beta_x(v)) (u- v){\rm d}\Gamma(x){\rm d}t \geq 0$.因此$G$单调.于是引理1.2蕴含$G : L^{p}(0,T; L^p(\Gamma))\rightarrow L^{p'}(0,T; L^{p'}(\Gamma))$极大单调.

定义$\Psi : L^{p}(0,T; L^p(\Gamma)) \rightarrow R$为$\Psi(u)= \int_0^T \int_{\Gamma}\varphi _{x}(u) {\rm d}\Gamma(x){\rm d}t,$则易知$\Psi$在$L^{p}(0,T; L^{p}(\Gamma))$ 上正则凸、下半连续.进而引理1.3蕴含$\partial \Psi : L^{p}(0,T; L^p(\Gamma))\rightarrow L^{p'}(0,T; L^{p'}(\Gamma))$极大单调.因为$\Psi(u)-\Psi(v)= \int_0^T \int_{\Gamma}[\varphi_{x}(u)-\varphi_{x}(v)]{\rm d}\Gamma(x){\rm d}t \geq \int_0^T\int_{\Gamma}\beta_{x}(v)(u-v){\rm d}\Gamma(x){\rm d}t = (Gv,u-v),$ $\forall u(x,t),v(x,t) \in L^{p}(0,T; L^p(\Gamma)),$所以$Gv \in \partial \Psi(v).$因此$G = \partial \Psi.$

因$K$和$G$均连续,故 $K^*GK: L^{p}(0,T; W^{1,p}(\Omega))\rightarrow L^{p'}(0,T; (W^{1,p}(\Omega))' ) $极大单调.对$ u,v\in L^{p}(0,T; W^{1,p}(\Omega))$,

\begin{eqnarray*} \Phi(v)-\Phi(u) &=& \Psi(Kv)-\Psi(Ku)\\ & =& \int_0^T\int_{\Gamma}[\varphi_{x}(v|_{\Gamma}(x,t))-\varphi_{x}(u|_{\Gamma}(x,t))]{\rm d}\Gamma(x){\rm d}t \\ & \geq &\int_0^T\int_{\Gamma}\beta_{x}(u|_{\Gamma}(x,t))(v|_{\Gamma}(x,t)-u|_{\Gamma}(x,t)){\rm d}\Gamma(x){\rm d}t \\ & =& (GKu,Kv-Ku) = (K^*GKu,v-u). \end{eqnarray*}

于是$K^*GK \subset \partial\Phi,$从而$K^*GK = \partial\Phi.$

进而 $\forall u,v \in L^{p}(0,T; W^{1,p}(\Omega)),$

$$ (v, \partial\Phi(u)) = \int_0^T\int_\Omega \beta_x(u|_\Gamma(x,t)) v |_\Gamma(x,t) {\rm d}\Gamma(x){\rm d}t. $$

又因$0 \in \beta_x(0)$,故$0 \in \partial\Phi(0).$

定理 2.1非线性问题(1.8)在$L^{p}(0,T; W^{1,p}(\Omega))$中存在唯一解.

由引理2.1,2.3,2.5,1.1和1.4,可知$B + \partial \Phi+ C:L^{p}(0,T; W^{1,p}(\Omega))\rightarrow L^{p'}(0,T;$ $ (W^{1,p}(\Omega))')$伪单调、极大单调且强迫.利用引理2.4和定理1.3,有

$$R(S + B+ \partial \Phi+ C) = L^{p'}(0,T; (W^{1,p}(\Omega))'). $$

于是对$f(x,t) \in L^{p'}(0,T; (W^{1,p}(\Omega))'),$存在 $u(x,t)\in L^{p}(0,T; W^{1,p}(\Omega))$满足

\begin{equation} Su + Bu+ \partial \Phi (u) + Cu =-f(x,t). \end{equation} (2.2)

并且$u(x,t)$还是唯一的.

事实上,若$u(x,t)$和$v(x,t)$均满足(2.2)式,则由$B$单调可知

$$(u-v,(S+C+\partial \Phi)u-(S+C+\partial \Phi)v) =-(u-v,Bu-Bv)\leq 0.$$

因$S+C+\partial \Phi$也单调,故 $(u-v,Bu-Bv) = 0.$再由$B$严格单调,$u(x,t) = v(x,t).$

由$S$的定义知 $u(x,0) = u(x,T)$且$\frac{\partial u}{\partial t} (x,0)= \frac{\partial u }{\partial t}(x,T),$ $\forall x \in \Omega.$

对$\forall w \in L^p(0,T; W^{1,p}(\Omega)),$利用 (2.2)式,有

$$ (u-w,Su) + (u-w,Bu) + (u-w,\partial \Phi(u))+(u-w,Cu) + (u-w,f(x,t))= 0. $$

由次微分的定义

\begin{equation} (u-w ,Su) + (u-w,Bu) +(u-w,Cu)+ \Phi(u)-\Phi(w) + (u-w,f(x,t))\leq 0. \end{equation} (2.3)

因此

\begin{eqnarray} &&-\int_0^T\int_{\Omega}\frac{\partial^2 u}{\partial t^2} (u-w) {\rm d}x {\rm d}t + \int_0^T \int_{\Omega}\langle (C(x,t)+|D u|^2)^{\frac{p-2}{2}}D u,D(u-w) \rangle {\rm d}x{\rm d}t \nonumber\\ & &+ \int_0^T\int_{\Omega}f(x,t) (u-w) {\rm d}x{\rm d}t+\Phi(u)-\Phi(w) +\int_0^T \int_{\Omega} g(x,u,\frac{\partial u}{\partial t},Du) (u-w){\rm d}x{\rm d}t\nonumber\\ &\leq & 0. \end{eqnarray} (2.4)

令$w = u \pm \psi,$其中$\psi\in C_0^\infty(\Omega\times (0,T)),$则

\begin{eqnarray*} &&-\int_0^T\int_{\Omega}\frac{\partial^2 u}{\partial t^2} \psi {\rm d}x {\rm d}t+ \int_0^T \int_{\Omega}\langle (C(x,t)+|Du|^2)^{\frac{p-2}{2}}Du,D\psi\rangle {\rm d}x{\rm d}t \\ &&+\int_0^T \int_{\Omega} g(x,u,\frac{\partial u}{\partial t},Du)\psi {\rm d}x{\rm d}t + \int_0^T\int_{\Omega}f(x,t) \psi {\rm d}x{\rm d}t = 0. \end{eqnarray*}

由广义函数的定义

\begin{equation}\frac{\partial^2 u}{\partial t^2}+{\rm div}[(C(x,t)+|Du|^2)^{\frac{p-2}{2}}Du] - g(x,u,\frac{\partial u}{\partial t},Du) = f(x,t) ~~{\rm a.e.} x\in \Omega \times (0,T). \end{equation} (2.5)

利用(2.4)式和Green公式有

\begin{eqnarray*} & &-\int_0^T\int_{\Omega}\frac{\partial^2 u}{\partial t^2} (w-u ) {\rm d}x{\rm d}t-\int_0^T \int_{\Omega} div[(C(x,t)+|Du|^2)^{\frac{p-2}{2}}Du] (w-u){\rm d}x{\rm d}t \\ && +\int^T_0 \int_\Gamma \langle \vartheta, (C(x,t)+|Du|^2)^{\frac{p-2}{2}}Du \rangle (w-u)|_{\Gamma} {\rm d}\Gamma(x){\rm d}t \\ && + \int_0^T\int_\Omega g(x,u,\frac{\partial u}{\partial t},Du) (w-u) {\rm d}x{\rm d}t + \Phi(w)-\Phi(u) + \int_0^T\int_{\Omega}f(x,t) (w-u) {\rm d}x{\rm d}t \geq 0. \end{eqnarray*}

利用(2.5)式,

$$ \Phi(w)-\Phi(u) \geq-\int^T_0 \int_\Gamma \langle \vartheta, (C(x,t)+|Du|^2)^{\frac{p-2}{2}}Du\rangle (w-u)|_{\Gamma} {\rm d}\Gamma(x){\rm d}t. $$

因此

$$-\langle \vartheta,(C(x,t)+|Du|^2)^{\frac{p-2}{2}}Du\rangle \in \partial \Phi(u(x,t)), $$

从而引理2.6蕴含 $-\langle \vartheta, (C(x,t)+|Du|^2)^{\frac{p-2}{2}}Du\rangle \in \beta_x(u(x,t))$ a.e. $(x,t) \in \Gamma \times (0,T). $

至此证明了方程(1.8)在$L^p(0,T; W^{1,p}(\Omega))$中存在唯一解.

推论 2.1 若令$C(x,t) \equiv 0$且 $g(x,u,\frac{\partial u}{\partial t },Du) \equiv |u|^{p-2}u,$ 则方程(1.8)退化成以下含有$p$-Laplacian 算子的双曲方程

$$\left\{\begin{array}{ll} \frac{\partial^2 u}{\partial t^2 }+\triangle_p u-|u|^{p-2}u = f(x,t), ~&(x,t) \in \Omega \times (0,T),\\ -\langle \vartheta,|D u|^{p-2}Du\rangle \in \beta_{x}(u(x,t)),~& (x,t) \in \Gamma \times (0,T),\\[2mm] u(x,0) = u(x,T),&x \in \Omega, \\[2mm] \frac{\partial u}{\partial t}(x,0) = \frac{\partial u}{\partial t}(x,T),& x \in \Omega, \end{array}\right.$$ (2.6)

由定理 2.1,方程(2.6)在 $L^{p}(0,T; W^{1,p}(\Omega))$中存在唯一解.

注2.1下面内容是对上面内容的推广.

引理 2.7 定义$\widetilde{B} : L^{p}(0,T; W^{1,p}(\Omega)) \rightarrow L^{p'}(0,T; (W^{1,p}(\Omega))')$为 $\widetilde{B}u \equiv Bu + f(x,t)$,对$\forall u \in L^{p}(0,T; W^{1,p}(\Omega)).$则$\widetilde{B}$ 极大单调.

类似于引理2.3可证 $\widetilde{B}$处处有定义、单调且 hemi连续.因此$\widetilde{B}$极大单调.

引理 2.8 定义算子$F : L^{p}(0,T; W^{1,p}(\Omega)) \rightarrow L^{p'}(0,T; (W^{1,p}(\Omega))')$为

$$F u = \widetilde{B}u + \partial \Phi(u) + Su + C u,$$

对$\forall u \in L^{p}(0,T; W^{1,p}(\Omega)),$则$F$极大单调.

由引理2.1,2.4,2.5,2.7和 1.4可知结论成立.

定义2.2 称 $F^{-1}0 : = \{u(x,t) \in L^{p}(0,T; W^{1,p}(\Omega)): Fu(x,t) = 0\}$为 $F$的零点集.若$u(x,t) \in F^{-1}0,$ 则称$u(x,t)$为极大单调算子$F$的零点.

定理2.2 $u(x,t) \in F^{-1}0$当且仅当$u(x,t) \in L^{p}(0,T; W^{1,p}(\Omega))$为方程(1.8)的唯一解.

令 $u(x,t)$为方程(1.8)的解,则对 $\forall v \in L^{p}(0,T; W^{1,p}(\Omega)),$由Green公式和引理2.6,有

\begin{eqnarray*} (v,F u) &=& \int_0^T\int_{\Omega} \langle (C(x,t)+|D u|^2)^{\frac{p-2}{2}}D u ,D v \rangle {\rm d}x{\rm d}t + \int_0^T\int_{\Omega}f(x,t)v(x,t) {\rm d}x{\rm d}t \\ && + \int_0^T\int_{\Omega}g(x,u,\frac{\partial u}{\partial t},Du)v(x,t) {\rm d}x {\rm d}t +(v, \partial\Phi(u))-\int_0^T\int_{\Omega}\frac{\partial^2 u}{\partial t^2}v {\rm d}x{\rm d}t \\ & =&-\int_0^T\int_{\Omega}div[(C(x,t)+|D u|^2)^{\frac{p-2}{2}}Du] v {\rm d}x{\rm d}t\\ && + \int_0^T\int_{\Gamma}\langle \vartheta ,(C(x,t)+|D u|^2)^{\frac{p-2}{2}}Du\rangle v|_{\Gamma}{\rm d}\Gamma(x){\rm d}t \\ &&+ \int_0^T\int_{\Omega}f(x,t)v(x,t) {\rm d}x{\rm d}t +\int_0^T\int_{\Omega}g(x,u,\frac{\partial u}{\partial t},Du)v(x,t) {\rm d}x{\rm d}t\\ &&+ \int_0^T\int_{\Gamma} \beta_x(u|_{\Gamma})v|_{\Gamma}{\rm d}\Gamma(x){\rm d}t -\int_0^T\int_{\Omega}\frac{\partial^2 u}{\partial t^2}v {\rm d}x{\rm d}t \\ & =& \int_0^T \int_{\Gamma}\langle \vartheta ,(C(x,t)+|D u|^2)^{\frac{p-2}{2}}Du \rangle v|_{\Gamma}{\rm d}\Gamma(x){\rm d}t + \int_0^T\int_{\Gamma} \beta_x(u|_{\Gamma})v|_{\Gamma}{\rm d}\Gamma(x){\rm d}t \\ & =&-\int_0^T\int_{\Gamma} \beta_x(u|_{\Gamma})v|_{\Gamma}{\rm d}\Gamma(x){\rm d}t+ \int_0^T\int_{\Gamma} \beta_x(u|_{\Gamma})v|_{\Gamma}{\rm d}\Gamma(x){\rm d}t = 0. \end{eqnarray*}

因此$u \in F^{-1}0.$

若$u(x,t) \in F^{-1}0,$则对$\forall \varphi \in C^{\infty}_0(\Omega \times (0,T)),$由引理2.6,有

\begin{eqnarray*} 0 &=&-\int_0^T\int_{\Omega} \frac{\partial^2 u}{\partial t^2} \varphi {\rm d}x{\rm d}t + \int_0^T\int_\Omega \langle (C(x,t)+|D u|^2)^{\frac{p-2}{2}}D u,D \varphi \rangle {\rm d}x{\rm d}t \\ & & + \int_0^T\int_\Omega f \varphi {\rm d}x{\rm d}t + \int_0^T\int_\Omega g(x,u,\frac{\partial u}{\partial t},Du)\varphi {\rm d}x{\rm d}t, \end{eqnarray*}

因此 $\frac{\partial^2 u}{\partial t^2} +{\rm div}[(C(x,t)+|Du|^2)^{\frac{p-2}{2}}Du]-g(x,u,\frac{\partial u }{\partial t},Du) = f(x,t)$ a.e. $(x,t)\in \Omega \times (0,T)$.

重复定理2.1最后一部分的证明过程,有 $-\langle \vartheta, (C(x,t)+|Du|^2)^{\frac{p-2}{2}}Du\rangle \in \beta_x(u(x,t)).$再由$S$的定义可知 $u(x,0) = u(x,T)$且 $\frac{\partial u}{\partial t} (x,0) = \frac{\partial u}{\partial t}(x,T),$ $\forall x\in \Omega,$于是$u(x,t)$为方程(1.8)的唯一解.

注 2.2 我们知道研究极大单调算子零点的迭代格式是应用数学领域的 又一个活跃的研究方向,研究的基础奠定在极大单调算子的零点集是非空集的前提下, 但鲜有人给出极大单调算子的零点集 非空的例子 (见文献[14, 15, 16]等). 这种假设合理吗?定理2.2便给出了合理性的解释.

参考文献
[1] Barbu V. Nonlinear Semigroups and Differential Equations in Banach Spaces. Leyden:Noordhoff, 1976
[2] Pascali D, Sburlan S. Nonlinear Mappings of Monotone Type. The Netherlands:Sijthoff and Noordhoff, 1978
[3] Lions J L. Quelques Methods de Resolution des Problems aux Limites Nonlineaires. Paris:Dunod Gauthier-Villars, 1969
[4] Calvert B D, Gupta C P. Nonlinear elliptic boundary value problems in Lp-spaces and sums of ranges of accretive operators. Nonlinear Analysis, 1978, 2:1-26
[5] Brezis H. Equations et inequalitions nonlinearies dans les espaces vectoriels an dualite. Ann Inst Fourier Grenoble, 1968, 18(1):115-175
[6] Chen Z C, Luo T. The initial-boundary value problem for quasilinear integro-differential equations. International J Differential Equa Appl, 2002, 6:299-306
[7] Reich S. The range of sums of accretive and monotone operators. J Math Anal Appli, 1979, 68:310-317
[8] Wei L, Zhou H Y. Research on the existence of solution of equation involving p-Laplacian operator. Applied Mathematics A J Chinese Univ(Ser B), 2006, 21:191-202
[9] 魏利. 与广义p-Laplace算子相关的非线性边值问题在Ls(Ω)空间中解的存在性. 数学物理学报, 2010, 30(A):1111-1116
[10] Wei L, He Z. The applications of theories of accretive operators to nonlinear elliptic boundary value problems in Lp-spaces. Nonlinear Analysis, 2001, 46:199-211
[11] Wei L, Zhou H Y. The existence of solutions of nonlinear boundary value problem involving the p-Laplacian operator in Ls-spaces. J Systems Science and Complexity, 2005, 18:511-521
[12] Wei L, Agarwal R P. Existence of solutions to nonlinear Neumann boundary value problem with generalized p-Laplacian operator. Comput Math Appl, 2008, 56:530-541
[13] Wei L, Agarwal R P, Wong P J Y. Existence of solutions to nonlinear parabolic boundary value problems with generalized p-Laplacian operator. Advances in Math Sci Appli, 2010, 20(2):423-445
[14] Kamimura S, Takahashi W. Approximating solutions of maximal monotone operators in Hilbert spaces. J Approx Theory, 2000, 106(2):226-240
[15] Rockafellar R T. Monotone operators and the proximal point algorithm. SIAM J Control Optim, 1976, 14:877-898
[16] Ibaraki T, Takahashi W. Weak and strong convergence theorems for new resolvents of maximal monotone operators in Banach spaces. Adv Math Econ, 2007, 10:51-64