数学杂志  2023, Vol. 43 Issue (6): 487-500   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
李春晖
王丹
刘淑丽
李金红
王晓丽
基于Bell多项式的一类(3+1)维变系数广义浅水波方程的可积性研究
李春晖, 王丹, 刘淑丽, 李金红, 王晓丽    
齐鲁工业大学(山东省科学院)数学与统计学院, 山东 济南 250353
摘要:本文基于Bell多项式研究了一类(3+1)维变系数广义浅水波方程的可积性问题.首先, 引入变量变换, 借助Bell多项式与Hirota双线性算子之间的关系, 导出方程的Hirota双线性形式, 求出方程的N-孤子解, 并对单孤子、双孤子和三孤子在不同情形下的传播进行图像模拟; 其次, 基于双线性方程, 结合Bell多项式获得方程的双线性Bäcklund变换; 然后, 通过Hopf-Cole变换, 将双线性Bäcklund变换线性化, 求出方程的Lax对; 最后, 利用级数展开法得到方程的无穷守恒律.从而证明该方程具有可积性.
关键词广义浅水波方程    Bell多项式    Bäcklund变换    Lax对    无穷守恒律    
INTEGRABILITY OF A VARIABLE-COEFFICIENT GENERALIZED (3+1) DIMENSIONAL SHALLOW WATER WAVE EQUATION VIA BELL POLYNOMIALS
LI Chun-hui, WANG Dan, LIU Shu-li, LI Jin-hong, WANG Xiao-li    
School of Mathematics and Statistics, Qilu University of Technology (Shandong Academy of Sciences), Shandong Jinan 250353
Abstract: In this paper, we focus on a(3+1)dimensional variable-coefficient generalized shallow water wave equation based on Bell polynomials.Firstly, the transformation of variables is introduced, and the Hirota bilinear form of the equation is derived by the relation between Bell polynomial and the Hirota bilinear operator.The N-soliton solution of the equation is obtained, and the propagation of single soliton, double soliton and triple soliton in different cases are simulated.Furthermore, the bilinear Bäcklund transformation is obtained based on the bilinear equation and Bell polynomials.Then, through the Hopf-Cole transformation, the bilinear Bäcklund transformation is linearized, and the Lax pair of the equation is obtained.Finally, the infinite conservation law of the equation is obtained by using the series expansion method.Thus, the integrability of the equation is proved.
Keywords: generalized shallow water wave equation     Bell polynomials     Bäcklund transformation     Lax pair     infinite conservation law    
1 引言

研究非线性演化方程的可积性能够很好地描述通讯物理、流体力学和海洋工程等领域的物理现象. 目前研究可积性的方法有反散射方法[1]、Hirota双线性方法[2-5]、Riemann-Hilbert方法[6]、Bäcklund变换法[7-9]和Darboux变换法[10]等. 本文主要基于1971年日本物理学家Hirota提出的双线性方法来研究方程的可积性. 该方法的难点之一是构造非线性演化方程的双线性形式, 而Bell多项式理论为其提供了简捷有效的途径[11-14]. Bell多项式的概念是1934年由美国数学家Bell首次提出的, 一直是组合学界的热门课题之一. 直到1996年, Lambert、Gilson和Nimmo等人[15]建立了Bell多项式和Hirota双线性算子之间的联系, 为非线性演化方程的精确解、双线性Bäckulund变换、Lax对和无穷守恒律等可积性质的研究提供了便捷的方法[16-23]. 近年来, Bell多项式方法被广泛应用于非线性演化方程的可积性研究中.

本文基于Bell多项式研究一类变系数广义浅水波方程

$ \begin{equation} \begin{aligned} m_1u_{xt}+m_2u_{yt}+m_3u_{xy}&+m_4u_{xxxy}+m_5(u_{xy}u_x+u_{xx}u_y)+m_6u_{xz}=0, \end{aligned} \end{equation} $ (1.1)

其中$ m_i=m_i(t)(i=1, \cdots, 6) $为任意实函数.

$ m_1=0, m_2=1, m_3=0, m_4=1, m_5=-3, m_6=-1 $时, 方程(1.1)化简为

$ \begin{equation} u_{yt}+u_{xxxy}-3u_xu_{xy}-3u_{xx}u_y-u_{xz}=0, \end{equation} $ (1.2)

该方程是Kadomtsev–Petviashvili(KP)族中的第二个方程[24]. KP族是可积方程族的一个典范, 包含无限个可积非线性微分方程, 在可积系统理论中起着重要作用. 文献[25]-[28]分别讨论了方程(1.2)的多孤子解、类孤子解、有理解, 新的类孤子解, Grammian和Pfaffian解以及精确周期波解.

$ m_1=0, m_2=2, m_3=0, m_4=1, m_5=3, m_6=-3 $时, 方程(1.1)化简为

$ \begin{equation} 2u_{yt}+u_{xxxy}+3u_xu_{xy}+3u_{xx}u_y-3u_{xz}=0, \end{equation} $ (1.3)

该方程是典型的(3+1)维Jimbo–Miwa(JM)方程[29], 在物理中用来描述具有弱色散的三维非线性波的传播. 研究(3+1)维JM方程对于观察各种类型的高色散孤子非常重要. 文献[30]-[32]分别讨论了方程(1.3)的精确解、行波解、孤子分子和相互作用解, 文献[33]讨论了方程(1.3) 的Bäcklund变换、Lax系统、无穷守恒律和多孤子解.

在诸多领域, 变系数非线性演化方程比常系数非线性演化方程能够提供更多的信息, 更有效地描述实际现象. 本文基于方程(1.2)和(1.3)研究更一般的变系数方程(1.1), 结构如下: 第二部分基于Bell多项式获得方程(1.1)的双线性形式和孤子解, 并分析孤子解的传播与演化; 第三部分在某种约束下, 利用Bell多项式构造方程(1.1)的双线性Bäcklund变换和Lax对; 第四部分利用级数展开求出方程(1.1)的无穷守恒律; 最后给出本文的结论.

2 双线性形式和孤子解
2.1 预备知识

定义2.1.1 [34]  多维Bell多项式也称$ Y $-多项式, 定义为$ {Y_{{n_1}{x_1}, \cdots, {n_l}{x_l}}}(f)\equiv{Y_{{n_1}, \cdots, {n_l}}}({f_{{r_1}{x_1}, \cdots, {r_l}{x_l}}})=e^{-f}{\partial _{{x_1}}^{{n_1}}\cdots\partial _{{x_l}}^{{n_l}}}e^{f} \nonumber, $其中, $ f=f(x_1, \cdots, x_n) $是具有$ n $个独立变量的函数, $ l $是任意的非负整数. $ {f_{{r_1}{x_1}, \cdots, {r_l}{x_l}}}={\partial_{{x_1}}^{{r_1}}\cdots\partial_{{x_l}}^{{r_l}}}f({r_1}=0, \cdots, n_1;\cdots;{r_l}=0, \cdots, n_l) $. 当$ f=f(x, y) $时, 对应的$ Y $-多项式为

$ \begin{eqnarray} \begin{split} \nonumber &Y_{x}(f)=f_x, \;\;Y_{2x}(f)=f_{2x}+f_x^2, \\ &Y_{x, y}(f)=f_{x, y}+{f_x}{f_y}, \;\; Y_{3x}(f)=f_{3x}+3{f_{2x}}{f_x}+f_{x}^3, \\ &\cdots \end{split} \end{eqnarray} $

定义2.1.2 [34]  多维双Bell多项式也称$ {\cal Y} $-多项式, 定义为

$ \begin{equation} {{\cal Y}_{{n_1}{x_1}, ..., {n_l}{x_l}}}(v, w)= {Y_{{n_1}{x_1}, ..., {n_l}{x_l}}}\left( f \right)\left| {_{{f_{{r_1}{x_1}, ..., {r_l}{x_l}}}= \left\{ \begin{array}{l} {v_{{r_1}{x_1}, ..., {r_l}{x_l}}}, {r_1} + \cdots + {r_l}\mbox{为奇数}, \\ {w_{{r_1}{x_1}, ..., {r_l}{x_l}}}, {r_1} + \cdots + {r_l}\mbox{为偶数}, \end{array} \right.}} \right. \nonumber \end{equation} $

其中, $ v=v(x_1, \cdots, x_n) $$ w=w(x_1, \cdots, x_n) $是具有$ n $个独立变量的函数. 当$ v=v(x, y) $, $ w=w(x, y) $时, 对应的$ {\cal Y} $-多项式为

$ \begin{eqnarray} \begin{split} &{\cal Y}_{x}(v, w)=v_x, \;\;{\cal Y}_{2x}(v, w)=w_{2x}+v_x^2, \\ &{\cal Y}_{x, y}(v, w)=w_{x, y}+{v_x}{v_y}, \;\;{\cal Y}_{3x}(v, w)=v_{3x}+3{v_x}{w_{2x}}+v_x^3, \\ &\cdots \end{split} \end{eqnarray} $ (2.1)

性质2.1.1 [14]  $ {\cal Y} $-多项式和Hirota双线性$ D $-算子的关系为

$ \begin{equation} {{\cal Y}_{{n_1}{x_1}, \cdots {n_l}{x_l}}}(v = \ln f/g, w = \ln fg) = {(fg)^{ - 1}}D_{{x_1}}^{{n_1}} \cdots D_{{x_l}}^{{n_l}}f \cdot g, \end{equation} $ (2.2)

其中$ {n_1} + {n_2} + \cdots + {n_l} \ge 1 $, Hirota双线性$ D $-算子定义为

$ \begin{equation} D_{{x_1}}^{{n_1}} \cdots D_{{x_l}}^{{n_l}}f \cdot g \equiv {({\partial _{{x_1}}} - {\partial _{x_1'}})^{{n_1}}} \cdots {({\partial _{{x_l}}} - {\partial _{x_l'}})^{{n_l}}}f({x_1}, \cdots , {x_l})g(x_1', \cdots , x_l')\left| {_{x_1' = {x_1}, \cdots , x_l' = {x_l}}} \right. \nonumber. \end{equation} $

特别地, 当$ f = g $时, 式(2.2)被化为

$ \begin{equation} \label{ident1} \begin{array}{l} {f^{ - 2}}D_{{x_1}}^{{n_1}} \cdots D_{{x_l}}^{{n_l}}f \cdot f = {{\cal Y}_{{n_1}{x_1}, \cdots {n_l}{x_l}}}(0, w = 2\ln f) \\ \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;= \left\{ \begin{array}{c} 0, \;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;\;{n_1} + \cdots + {n_l}\mbox{为奇数}, \\ {P_{{n_1}{x_1}, \cdots , {n_l}{x_l}}}(w), {n_1} + \cdots + {n_l}\mbox{为偶数}, \\ \end{array} \right. \nonumber \\ \end{array} \end{equation} $

其中$ P_{{n_1}{x_1}, \cdots , {n_l}{x_l}}(w)={\cal Y}_{{n_1}{x_1}, \cdots {n_l}{x_l}}(0, w = 2\ln f) $称为$ P $-多项式. 例如$ w=w(x, y, t) $时, $ P $-多项式为

$ \begin{eqnarray} \begin{split} &{P_{t, x}}(w) = {w_{t, x}}, \;\;{P_{2x}}(w) = {w_{2x}}, \;\;{P_{3x, y}}(w) = {w_{3x, y}} + 3{w_{2x}}{w_{x, y}}, \\ &\cdots \end{split} \end{eqnarray} $ (2.3)
2.2 双线性形式

定理2.2.1  做变换$ u={\frac{3m_4}{m_5}}(2 \ln f)_x-\phi(y, z)(m_5\neq0) $, 则方程(1.1)的双线性形式为

$ \begin{equation} \begin{aligned} \left[m_1{D_x}{D_t}+m_2{D_y}{D_t}+m_3{D_x}{D_y}+m_4D_x^3{D_t}-\phi_y(y, z) m_5D_x^2+m_6{D_x}{D_z}\right]f\cdot f=0, \end{aligned} \end{equation} $ (2.4)

其中$ f $是关于$ x, y, z $$ t $的函数, $ \phi(y, z) $是关于$ y $$ z $的函数.

  引入辅助变量$ q $$ \phi(y, z) $, 其中$ q $是关于$ x, y, z $$ t $的实函数, $ \phi(y, z) $是关于$ y $$ z $的函数, 并且令

$ \begin{equation} u=cq_x-\phi(y, z), \end{equation} $ (2.5)

其中$ c $是待确定的常数. 将方程(2.5)带入方程(1.1), 并且等式两边对$ x $积分一次得

$ \begin{equation} E(q)\equiv m_1q_{x, t}+m_2q_{y, t}+m_3q_{x, y}+m_4q_{3x, y}+cm_5q_{2x}q_{x, y}-\phi_y(y, z) m_5q_{2x}+m_6q_{x, z}=0. \end{equation} $ (2.6)

$ c=\frac{3m_4}{m_5}(m_5\neq0) $, 上式化为

$ \begin{equation} E(q)\equiv m_1q_{x, t}+m_2q_{y, t}+m_3q_{x, y}+m_4(q_{3x, y}+3q_{2x}q_{x, y})-\phi_y(y, z) m_5q_{2x}+m_6q_{x, z}=0. \end{equation} $ (2.7)

并根据式(2.3), 则方程(2.7)可转化为如下$ P $-多项式形式

$ \begin{equation} \begin{aligned} m_1P_{x, t}(q)+m_2P_{y, t}(q)&+m_3P_{x, y}(q)+m_4P_{3x, y}(q)-\phi_y(y, z) m_5P_{2x}(q)+m_6P_{x, z}(q)=0. \end{aligned} \end{equation} $ (2.8)

作变量变换

$ \begin{equation} q=2 \ln f, \end{equation} $ (2.9)

且由性质(2.1.1), 可得方程(1.1)的双线性形式

$ \begin{equation} \nonumber \begin{aligned} \left[m_1{D_x}{D_t}+m_2{D_y}{D_t}+m_3{D_x}{D_y}+m_4D_x^3{D_t}-\phi_y(y, z) m_5D_x^2+m_6{D_x}{D_z}\right]f\cdot f=0. \end{aligned} \end{equation} $

证毕.

2.3 孤子解

定理2.3.1   方程(1.1)有N-孤子解

$ \begin{equation} u = \frac{6m_4}{m_5}{\left[\ln \left(\sum\limits_{\mu = 0, 1}^{} {{e^{\sum\limits_{j = 1}^n {{\mu _i}{\xi _i} + \sum\limits_{1 \le j < s}^{} {{\mu _j}{\mu _s}{A_{js}}} } }}} \right)\right]_x}-\phi(y, z), \end{equation} $ (2.10)

其中对$ \mu $的求和应取$ \mu_j=0, 1(j=1, 2, \cdots) $所有可能的组合,

$ \begin{equation} \nonumber \xi _j=w_jt+k_jx+p_jy+l_jz+{\xi _j}^{(0)}, \end{equation} $
$ \begin{equation} \nonumber \begin{aligned} e^{A_{js}}=&-\frac{m_1(w_j-w_s)(k_j-k_s)+m_2(w_j-w_s)(p_j-p_s)+m_3(k_j-k_s)(p_j-p_s)}{symb_{js}}\\ &-\frac{m_4(k_j-k_s)^3(p_j-p_s)-\phi_y(y, z)m_5(k_j-k_s)^2+m_6(k_j-k_s)(l_j-l_s)}{symb_{js}}, \\ symb_{js}=&m_1(w_j+w_s)(k_j+k_s)+m_2(w_j+w_s)(p_j+p_s)+m_3(k_j+k_s)(p_j+p_s)\\ &+m_4(k_j+k_s)^3(p_j+p_s)-\phi_y(y, z)m_5(k_j+k_s)^2+m_6(k_j+k_s)(l_j+l_s), \end{aligned} \end{equation} $

这里

$ \begin{equation} \nonumber w_j=-\frac{m_3k_jp_j+m_4{k_j}^3p_j-\phi_y(y, z)m_5{k_j}^2+m_6k_jl_j}{m_1k_j+m_2p_j}, \end{equation} $

$ m_1k_j+m_2p_j\neq0 $, 且$ k_j $$ p_j $$ l_j $$ {\xi _j}^{(0)}(j=1, 2, \cdots, n) $是任意常数.

  将式(2.4)中的$ f $按参数展开成级数

$ \begin{equation} f=1+f^{(1)}\varepsilon+f^{(2)}\varepsilon^2+\cdots+f^{(j)}\varepsilon^j+\cdots, \end{equation} $ (2.11)

将展开式(2.11)带入双线性方程(2.4), 比较$ \varepsilon $的同次幂系数有

$ \begin{equation} m_1f_{tx}^{(1)}+m_2f_{ty}^{(1)}+m_3f_{xy}^{(1)}+m_4f_{xxxy}^{(1)}-\phi_y(y, z)m_5f_{xx}^{(1)} +m_6f_{xz}^{(1)}=0, \end{equation} $ (2.12)
$ \begin{equation} \begin{aligned} &2[m_1f_{tx}^{(2)}+m_2f_{ty}^{(2)}+m_3f_{xy}^{(2)}+m_4f_{xxxy}^{(2)}-\phi_y(y, z)m_5f_{xx}^{(2)}+m_6f_{xz}^{(2)}] \\ =&-\left[m_1D_xD_t+m_2D_yD_t+m_3D_xD_y+m_4D_x^3D_t-\phi_y(y, z)m_5D_x^2+m_6D_xD_z\right]f^{(1)}\cdot f^{(1)}, \end{aligned} \end{equation} $ (2.13)
$ \begin{equation} \begin{aligned} &2[m_1f_{tx}^{(3)}+m_2f_{ty}^{(3)}+m_3f_{xy}^{(3)}+m_4f_{xxxy}^{(3)}-\phi_y(y, z)m_5f_{xx}^{(3)}+m_6f_{xz}^{(3)}] \\ &=-\left[m_1D_xD_t+m_2D_yD_t+m_3D_xD_y+m_4D_x^3D_t-\phi_y(y, z)m_5D_x^2+m_6D_xD_z\right]f^{(1)}\cdot f^{(2)}, \end{aligned} \end{equation} $ (2.14)
$ \begin{equation} \vdots \nonumber \end{equation} $

(ⅰ) 单孤子解

方程(2.12)是关于$ f^{(1)} $的一个线性微分方程, 容易得到指数形式解

$ \begin{equation} f^{(1)}=e^{\xi _1}, \;\;\;\xi _1=w_1t+k_1x+p_1y+l_1z+{\xi ^{(0)}_1}, \end{equation} $ (2.15)

其中

$ \begin{equation} \nonumber w_1=-\frac{m_3k_1p_1+m_4{k_1}^3p_1-\phi_y(y, z)m_5{k_1}^2+m_6k_1l_1}{m_1k_1+m_2p_1}, \end{equation} $

这里$ m_1k_1+m_2p_1\neq0 $, 且$ k_1 $$ p_1 $$ l_1 $$ {\xi ^{(0)}_1} $是任意常数.

$ f^{(j)}=0(j=2, 3, \cdots) $, 则双线性方程(2.4)有解

$ \begin{equation} f=1+e^{\xi_1}, \end{equation} $ (2.16)

注意这里扰动参数$ \varepsilon $可被吸收到任意常数$ {\xi ^{(0)}_1} $中. 进而方程(1.1)的单孤子解为

$ \begin{equation} u=\frac{6m_4}{m_5}{\Big[\ln (1 + {e^{{\xi _1}}})\Big]_x}-\phi(y, z). \end{equation} $ (2.17)

(ⅱ) 双孤子解

由于(2.12)是关于$ f^{(1)} $的线性微分方程, 所以$ f^{(1)} $有叠加解

$ \begin{equation} f^{(1)}=e^{\xi_1}+e^{\xi_2}, \;\;\xi_j=w_jt+k_jx+p_jy+l_jz+\xi_j^{(0)}(j=1, 2), \end{equation} $ (2.18)

其中

$ \begin{equation} \nonumber w_j=-\frac{m_3k_jp_j+m_4{k_j}^3p_j-\phi_y(y, z)m_5{k_j}^2+m_6k_jl_j}{m_1k_j+m_2p_j}(j=1, 2). \end{equation} $

将(2.18)代入(2.13)得

$ \begin{equation} \begin{aligned} &m_1f_{tx}^{(2)}+m_2f_{ty}^{(2)}+m_3f_{xy}^{(2)}+m_4f_{xxxy}^{(2)}-\phi_y(y, z)m_5f_{xx}^{(2)}+m_6f_{xz}^{(2)}\\ &=-\left[{m_1(w_1-w_2)(k_1-k_2)+m_2(w_1-w_2)(p_1-p_2)+m_3(k_1-k_2)(p_1-p_2)}\right.\\ &\left.{+m_4(k_1-k_2)^3(p_1-p_2)-\phi_y(y, z)m_5(k_1-k_2)^2+m_6(k_1-k_2)(l_1-l_2)}\right]e^{\xi_1+\xi_2}. \end{aligned} \end{equation} $ (2.19)

解得

$ \begin{equation} f^{(2)}=e^{\xi_1+\xi_2+A_{12}}, \end{equation} $ (2.20)

其中

$ \begin{equation} \nonumber \begin{aligned} e^{A_{12}}=&-\frac{m_1(w_1-w_2)(k_1-k_2)+m_2(w_1-w_2)(p_1-p_2)+m_3(k_1-k_2)(p_1-p_2)}{symb_{12}}\\ &-\frac{m_4(k_1-k_2)^3(p_1-p_2)-\phi_y(y, z)m_5(k_1-k_2)^2+m_6(k_1-k_2)(l_1-l_2)}{symb_{12}}, \\ symb_{12}=&m_1(w_1+w_2)(k_1+k_2)+m_2(w_1+w_2)(p_1+p_2)+m_3(k_1+k_2)(p_1+p_2)\\ &+m_4(k_1+k_2)^3(p_1+p_2)-\phi_y(y, z)m_5(k_1+k_2)^2+m_6(k_1+k_2)(l_1+l_2). \end{aligned} \end{equation} $

$ f^{(3)}=f^{(4)}=f^{(5)}=\cdots=0 $, 双线性方程(2.4)有解

$ \begin{equation} f=1+e^{\xi_1}+e^{\xi_2}+e^{\xi_1+\xi_2+A_{12}}, \end{equation} $ (2.21)

从而方程(1.1)的双孤子解为

$ \begin{equation} u=\frac{6m_4}{m_5}{\Big[\ln (1 + {e^{\xi _1}}+e^{\xi_2}+e^{\xi_1+\xi_2+A{12}})\Big]_x}-\phi(y, z). \end{equation} $ (2.22)

(ⅲ) 三孤子解

类似地可得方程(1.1)的三孤子解为

$ \begin{align} u = &\frac{6m_4}{m_5}{\Big[\ln (1 + {e^{{\xi _1}}} + {e^{{\xi _2}}} +{e^{{\xi _3}}} + {e^{{\xi _1} + {\xi _2} + {A_{12}}}}+{e^{{\xi _1} + {\xi _3} + {A_{13}}}}+{e^{{\xi _2} + {\xi _3} + {A_{23}}}}+{e^{{\xi _1} + {\xi _2} + {\xi _3} {A_{12}}+{A_{13}}+{A_{23}}}})\Big]_x} \\ &-\phi(y, z), \end{align} $ (2.23)

其中

$ \begin{equation} \nonumber \begin{aligned} e^{A_{js}}=&-\frac{m_1(w_j-w_s)(k_j-k_s)+m_2(w_j-w_s)(p_j-p_s)+m_3(k_j-k_s)(p_j-p_s)}{symb_{js}}\\ &-\frac{m_4(k_j-k_s)^3(p_j-p_s)-\phi_y(y, z)m_5(k_j-k_s)^2+m_6(k_j-k_s)(l_j-l_s)}{symb_{js}}, \\ symb_{js}=&m_1(w_j+w_s)(k_j+k_s)+m_2(w_j+w_s)(p_j+p_s)+m_3(k_j+k_s)(p_j+p_s)\\ &+m_4(k_j+k_s)^3(p_j+p_s)-\phi_y(y, z)m_5(k_j+k_s)^2+m_6(k_j+k_s)(l_j+l_s), \end{aligned} \end{equation} $
$ \begin{equation} \nonumber \;\;\;\;\;\;\;\;\;\;\;\;\;\;(j<s, j, s=1, 2, 3). \end{equation} $

(ⅳ) N-孤子解

归纳可得方程(1.1)的N-孤子解为

$ \begin{equation} \nonumber u = \frac{6m_4}{m_5}{\left[\ln \left(\sum\limits_{\mu = 0, 1}^{} {{e^{\sum\limits_{j = 1}^n {{\mu _i}{\xi _i} + \sum\limits_{1 \le j < s}^{} {{\mu _j}{\mu _s}{A_{js}}} } }}} \right)\right]_x}-\phi(y, z), \end{equation} $

其中对$ \mu $的求和应取$ \mu_j=0, 1(j=1, 2, \cdots) $所有可能的组合,

$ \begin{equation} \nonumber \xi _j=w_jt+k_jx+p_jy+l_jz+{\xi _j}^{(0)}, \end{equation} $
$ \begin{equation} \nonumber \begin{aligned} e^{A_{js}}=&-\frac{m_1(w_j-w_s)(k_j-k_s)+m_2(w_j-w_s)(p_j-p_s)+m_3(k_j-k_s)(p_j-p_s)}{symb_{js}}\\ &-\frac{m_4(k_j-k_s)^3(p_j-p_s)-\phi_y(y, z)m_5(k_j-k_s)^2+m_6(k_j-k_s)(l_j-l_s)}{symb_{js}}, \\ symb_{js}=&m_1(w_j+w_s)(k_j+k_s)+m_2(w_j+w_s)(p_j+p_s)+m_3(k_j+k_s)(p_j+p_s)\\ &+m_4(k_j+k_s)^3(p_j+p_s)-\phi_y(y, z)m_5(k_j+k_s)^2+m_6(k_j+k_s)(l_j+l_s), \end{aligned} \end{equation} $

这里

$ \begin{equation} \nonumber w_j=-\frac{m_3k_jp_j+m_4{k_j}^3p_j-\phi_y(y, z)m_5{k_j}^2+m_6k_jl_j}{m_1k_j+m_2p_j}, \end{equation} $

$ m_1k_j+m_2p_j\neq0 $, 且$ k_j $$ p_j $$ l_j $$ {\xi _j}^{(0)}(j=1, 2, \cdots, n) $是任意常数.

接下来以$ \phi_y(y, z)=2 $为例, 讨论孤子解的传播和相互作用.

图 1图 6给出了系数$ m_4 $为常数($ m_4=1 $)和$ t $的线性函数($ m_4=t $)时, 单孤子解、双孤子解和三孤子解在$ x-t $$ y-t $$ z-t $平面上的传播. 从图中可以看出, 单孤子解是亮孤子解, 波在传播过程中振幅保持不变; 双孤子解和三孤子解中, 孤波发生碰撞后会快速恢复到原来的形状, 保持振幅不变. 同时, 当系数$ m_4 $为常数时, 波的传播方向不改变; 当系数$ m_4 $$ t $的线性函数时, 波的传播方向发生改变.

图 1 方程(1.1)的单孤子解(2.17), 其中取$ m_i=1(i=1, 2, 3, 4, 6) $, $ m_5=2 $, $ k_1=1 $, $ p_1=1 $, $ l_1=1 $

图 2 方程(1.1)的单孤子解(2.17), 其中取$ m_i=1(i=1, 2, 3, 6) $, $ m_4=t $, $ m_5=2t $, $ k_1=1 $, $ p_1=1 $, $ l_1=1 $

图 3 方程(1.1)的双孤子解(2.22),其中取$ m_i=1(i=1, 2, 3, 4, 6) $, $ m_5=2 $, $ k_1=2 $, $ p_1=1 $, $ l_1=2 $, $ k_2=-1 $, $ p_2=3 $, $ l_2=4 $

图 4 方程(1.1)的双孤子解(2.22), 其中取$ m_i=1(i=1, 2, 3, 6) $, $ m_4=t $, $ m_5=2t $, $ k_1=2 $, $ p_1=1 $, $ l_1=2 $, $ k_2=-1 $, $ p_2=3 $, $ l_2=4 $

图 5 方程(1.1)的三孤子解(2.23), 其中取$ m_i=1(i=1, 2, 3, 4, 6) $, $ m_5=2 $, $ k_1=1 $, $ p_1=3 $, $ l_1=2 $, $ k_2=-1 $, $ p_2=2 $, $ l_2=-3 $, $ k_3=2 $, $ p_3=3 $, $ l_3=-4 $

图 6 方程(1.1)的三孤子解(2.23), 其中取$ m_i=1(i=1, 2, 3, 6) $, $ m_4=t $, $ m_5=2t $, $ k_1=1 $, $ p_1=3 $, $ l_1=2 $, $ k_2=-1 $, $ p_2=2 $, $ l_2=-3 $, $ k_3=2 $, $ p_3=3 $, $ l_3=-4 $
3 双线性Bäcklund变换和Lax对
3.1 双线性Bäcklund变换

定理3.1.1  假设$ f $$ g $是双线性方程(2.4)的两个解, 则方程(1.1)的双线性Bäcklund变换为

$ \begin{align} &(D_xD_y-\lambda D_x)f \cdot g=0, \end{align} $ (3.1a)
$ \begin{align} &\big[m_1D_t+m_3D_y-m_5\phi_y(y, z)D_x+m_6D_z\big]f \cdot g=0, \end{align} $ (3.1b)
$ \begin{align} &(m_2D_t+m_4D_x^3)f \cdot g=0. \end{align} $ (3.1c)

  设$ q=2\ln f $$ \tilde{q}=2\ln g $均是方程(2.7)的解, 引进两个新的独立变量

$ \begin{equation} w=\frac{\tilde{q}+q}{2}=\ln (fg), v=\frac{\tilde{q}-q}{2}=\ln {\frac{f}{g}}, \end{equation} $ (3.2)

$ \begin{equation} q=w-v, \tilde{q}=w+v, \end{equation} $ (3.3)

且由式(2.7)知, $ q $$ \tilde{q} $满足二场条件

$ \begin{eqnarray} \begin{split} E(\tilde{q})-E(q)&=E(w+v)-E(w-v)\\ &=2m_1v_{xt}+2m_2v_{yt}+2m_3v_{xy}+m_4\Big[2v_{3x, y}+3(2w_{2x}v_{xy}+2w_{xy}v_{2x})\Big]\\ &\;\;\;-2m_5\phi_y(y, z)v_{2x}+2m_6v_{xz}. \end{split} \end{eqnarray} $ (3.4)

根据式(2.1), 条件(3.4)可以重新写作

$ \begin{eqnarray} \begin{split} \frac{1}{2}\Big[E(\tilde{q})-E(q)\Big]&=\partial x\Big[m_1{\cal Y}_t(v, w)+m_3{\cal Y}_y(v, w)-m_5\phi_y(y, z){\cal Y}_x(v, w)+m_6{\cal Y}_z(v, w)\Big]\\ &\;\;\;+\partial y\Big[m_2{\cal Y}_t(v, w)+m_4{\cal Y}_{3x}(v, w)\Big]+3m_4R(v, w), \end{split} \end{eqnarray} $ (3.5)

其中

$ \begin{equation} R(v, w)=Wr\big[{\cal Y}_{x, y}(v, w), {\cal Y}_x(v, w)\big]. \end{equation} $ (3.6)

这里$ Wr $代表$ Wronski $行列式, $ Wr\big[{\cal Y}_{x, y}(v, w), {\cal Y}_x(v, w)\big]=w_{xy}v_{2x}-w_{2x, y}v_x-v^2_xv_{xy} $.

引入限制条件

$ \begin{equation} {\cal Y}_{x, y}(v, w)-\lambda {\cal Y}_x(v, w)=0, \end{equation} $ (3.7)

其中$ \lambda $是常数, 则$ R(v, w)=0 $, 从而方程(1.1)有以下$ {\cal Y} $-多项式型的$ B\ddot{a}cklund $变换

$ \begin{align} &{\cal Y}_{x, y}(v, w)-\lambda {\cal Y}_x(v, w)=0, \end{align} $ (3.8a)
$ \begin{align} &m_1{\cal Y}_t(v, w)+m_3{\cal Y}_y(v, w)-m_5\phi_y(y, z){\cal Y}_x(v, w)+m_6{\cal Y}_z(v, w)=0, \end{align} $ (3.8b)
$ \begin{align} &m_2{\cal Y}_t(v, w)+m_4{\cal Y}_{3x}(v, w)=0. \end{align} $ (3.8c)

结合$ {\cal Y} $-多项式和Hirota双线性$ D $-算子的关系式(2.2), 可得到方程(1.1)的双线性$ B\ddot{a}cklund $变换

$ \begin{align} &(D_xD_y-\lambda D_x)f \cdot g=0, \end{align} $ (3.9a)
$ \begin{align} &\big[m_1D_t+m_3D_y-m_5\phi_y(y, z)D_x+m_6D_z\big]f \cdot g=0, \end{align} $ (3.9b)
$ \begin{align} &(m_2D_t+m_4D_x^3)f \cdot g=0. \end{align} $ (3.9c)

证毕.

3.2 Lax对

定理3.2.1  通过Hopf-Cole变换, 方程(1.1)的Lax对为

$ \begin{align} &\psi_{x, y}-\lambda \psi_{x}+\frac{m_5}{3m_4}\big[u_y+\phi_y(y, z)\big]\psi=0, \end{align} $ (3.10a)
$ \begin{align} &(m_1+m_2)\psi_t+m_5\big[u_x-\phi_y(y, z)\big]\psi_x+m_4\psi_{3x}+m_3\psi_y+m_6\psi_z=0, \end{align} $ (3.10b)

其中$ \lambda $是任意参数.

  做Hopf-Cole变换$ v=\ln \psi $, 结合式(2.1)和(3.3)可得

$ \begin{eqnarray} \begin{split} &{{\cal Y}_t}(v, w)={{\cal Y}_t}(v, v+q)={\psi_t}/\psi, \quad {{\cal Y}_x}(v, w)={{\cal Y}_x}(v, v+q)={\psi_x}/\psi, \\ &{{\cal Y}_y}(v, w)={{\cal Y}_y}(v, v+q)={\psi_y}/\psi, \quad {{\cal Y}_z}(v, w)={{\cal Y}_z}(v, v+q)={\psi_z}/\psi, \\ &{{\cal Y}_{x, y}}(v, w)={{\cal Y}_{x, y}}(v, v+q)=q_{x, y}+{\psi_{x, y}}/\psi, \\ &{{\cal Y}_{3x}}(v, w)={{\cal Y}_{3x}}(v, v+q)=3q_{2x}{\psi_x}/\psi+{\psi_{3x}}/\psi. \end{split} \end{eqnarray} $

从而$ {\cal Y} $-多项式型的$ B\ddot{a}cklund $变换(3.8)被线性化为含有参数$ \lambda $的线性系统

$ \begin{align} &\psi_{x, y}-\lambda \psi_{x}+q_{x, y}\psi=0, \end{align} $ (3.11a)
$ \begin{align} &m_1\psi_t+m_3\psi_y-m_5\phi_y(y, z)\psi_x+m_6\psi_z=0, \end{align} $ (3.11b)
$ \begin{align} &m_2\psi_t+3m_4q_{2x}\psi_x+m_4\psi_{3x}=0. \end{align} $ (3.11c)

从而

$ \begin{align} &\psi_{x, y}-\lambda \psi_{x}+q_{x, y}\psi=0, \\ &(m_1+m_2)\psi_t+\big[3m_4q_{2x}-m_5\phi_y(y, z)\big]\psi_x+m_4\psi_{3x}+m_3\psi_y+m_6\psi_z=0. \end{align} $ (3.12)

通过变换(2.5), 求得方程(1.1)的Lax对

$\begin{align} &\psi_{x, y}-\lambda \psi_{x}+\frac{m_5}{3m_4}\big[u_y+\phi_y(y, z)\big]\psi=0, \\ &(m_1+m_2)\psi_t+m_5\big[u_x-\phi_y(y, z)\big]\psi_x+m_4\psi_{3x}+m_3\psi_y+m_6\psi_z=0. \end{align}$

容易验证相容性条件$ \psi_{xyt}=\psi_{txy} $可以推出方程(1.1), 证毕.

4 无穷守恒律

定理4.1  通过级数展开, 方程(1.1)有以下无穷守恒律

$ \begin{equation} {\cal J}_{n, t}+{\cal F}_{n, x}+m_4{\cal G}_{n, y}+m_6{\cal H}_{n, z}=0, \;\;n=1, 2, 3, \cdots, \end{equation} $ (4.1)

其中守恒密度$ {\cal J}_n $的显式公式为

$ \begin{equation} {\cal J}_n=m_1({\partial^{-1}_y{{\cal I}_{n, x}}})+m_2{{\cal I}_n}, n=1, 2, \cdots. \end{equation} $ (4.2)

第一个连带流$ {\cal F}_n $的递推公式为

$ \begin{equation} {\cal F}_n=m_3{\cal I}_n+m_4{{\cal I}_{n, 2x}}-m_5\phi _y(y, z)(\partial^{-1}_y {\cal I}_{n, x}), n=1, 2, \cdots. \end{equation} $ (4.3)

第二个连带流$ {\cal G}_n $的递推公式为

$ \begin{eqnarray} \begin{split} {\cal G}_1=&\frac{m_5}{m_4}u_x({\partial^{-1}_y {\cal I}_{1, x}}), \\ {\cal G}_2=&3(\partial^{-1}_y {\cal I}_{1, 2x})(\partial^{-1}_y {\cal I}_{1, x})+\frac{m_5}{m_4}u_x({\partial^{-1}_y {\cal I}_{2, x}}), \\ {\cal G}_n=&3\sum\limits_{k=1}^{n-1}({\partial^{-1}_y {\cal I}_{k, 2x}})({\partial^{-1}_y {\cal I}_{n-k, x}})+\frac{m_5}{m_4}u_x({\partial^{-1}_y {\cal I}_{n, x}}) \\ &+\sum\limits_{i+j+k=n}(\partial^{-1}_y {\cal I}_{i, x})(\partial^{-1}_y {\cal I}_{j, x})(\partial^{-1}_y {\cal I}_{k, x}), n=3, 4, \cdots. \end{split} \end{eqnarray} $ (4.4)

剩余连带流$ {\cal H}_n $的递推公式为

$ \begin{equation} {\cal H}_n=\partial^{-1}_y {\cal I}_{n, x}, n=1, 2, \cdots. \end{equation} $ (4.5)

这里守恒密度$ {\cal I}_n $的递归关系如下

$ \begin{eqnarray} \begin{split} {\cal I}_1&=-q_{2y}=-\frac{m_5}{3m_4}\big[\partial^{-1}_x \big(u_{2y}+\phi_{2y}(y, z)\big)\big], \\ {\cal I}_2&=-{\cal I}_{1, y}+\lambda {\cal I}_1=q_{3y}-\lambda q_{2y}\\ &=\frac{m_5}{3m_4}\big[\partial^{-1}_x\big(u_{3y}+\phi_{3y}(y, z)\big)\big]-\frac{\lambda m_5}{3m_4}\big[\partial^{-1}_x\big(u_{2y}+\phi_{2y}(y, z)\big)\big], \\ {\cal I}_n&=-{\cal I}_{n-1, y}+\lambda {\cal I}_{n-1}-\sum\limits_{k=1}^{n-2}\partial^{-1}_x\partial y({\cal I}_k \partial^{-1}_y {{\cal I}_{n-k-1, x}}), n=3, 4, \cdots. \end{split} \end{eqnarray} $ (4.6)

  由关系$ \partial x{\cal Y}_t(v)=\partial t{\cal Y}_x(v)=v_{x, t} $, $ \partial x{\cal Y}_z(v)=\partial z{\cal Y}_x(v)=v_{x, z} $, 则式(3.8)可以化为

$ \begin{align} &w_{xy}+v_xv_y-\lambda v_x=0, \end{align} $ (4.7a)
$ \begin{align} &\partial t(m_1v_x+m_2v_y)+\partial x\big[m_3v_y+m_4v_{2x, y}-m_5\phi_y(y, z)v_x\big]+m_4\partial y(3w_{2x}v_x+{v_x}^3) \\ &+m_6\partial z(v_x)=0. \end{align} $ (4.7b)

引进一个新的势函数

$ \begin{equation} \eta=\frac{\tilde{q}_y-q_y}{2}, \end{equation} $ (4.8)

则由关系式(3.2)可得

$ \begin{equation} v_y=\eta, w_y=q_y+\eta. \end{equation} $ (4.9)

将(4.9)式代入(4.7)式, 得到一个Riccati型方程

$ \begin{equation} q_{xy}+\eta_x+\eta(\partial^{-1}_y\eta_x)-\lambda(\partial^{-1}_y\eta_x)=0, \end{equation} $ (4.10)

和一个离散型方程

$ \begin{eqnarray} \begin{split} &\partial t[m_1(\partial^{-1}_y\eta_x)+m_2\eta]+\partial x\big[m_3\eta+m_4\eta_{2x}-m_5\phi_y(y, z)(\partial^{-1}_y\eta_x)\big] \\ &+m_4\partial y \big[3(\partial^{-1}_y\eta_{xx}+q_{2x})(\partial^{-1}_y\eta_x)+(\partial^{-1}_y\eta_x)^3 \big]+m_6\partial z(\partial^{-1}_y\eta_x)=0. \end{split} \end{eqnarray} $ (4.11)

将展开式

$ \begin{equation} \eta=\varepsilon+\sum\limits_{n=1}^{\infty}{\cal I}_n(q, q_x, q_{2x}, \cdots)\varepsilon^{-n}, \end{equation} $ (4.12)

代入方程(4.10)并令$ \varepsilon $各幂次系数为零

$ \begin{eqnarray} \begin{split} \varepsilon^0&:\partial^{-1}_y{\cal I}_{1, x}+q_{x, y}=0, \\ \varepsilon^{-1}&:{\cal I}_{1, x}+\partial^{-1}_y{\cal I}_{2, x}-\lambda(\partial^{-1}_y{\cal I}_{1, x})=0, \\ \varepsilon^{-2}&:{\cal I}_{2, x}+\partial^{-1}_y{\cal I}_{3, x}+{\cal I}_1(\partial^{-1}_y{\cal I}_{1, x})-\lambda(\partial^{-1}_y{\cal I}_{2, x})=0, \\ \varepsilon^{-3}&:{\cal I}_{3, x}+\partial^{-1}_y{\cal I}_{4, x}+{\cal I}_1(\partial^{-1}_y{\cal I}_{2, x})+{\cal I}_2(\partial^{-1}_y{\cal I}_{1, x})-\lambda(\partial^{-1}_y{\cal I}_{3, x})=0, \\ &\vdots \end{split} \end{eqnarray} $ (4.13)

得到守恒密度的递推关系式(4.6). 再将展开式(4.12)代入式(4.11), 有

$ \begin{eqnarray} \begin{split}\nonumber &\partial t[m_1(\partial^{-1}_y\sum\limits_{n=1}^{\infty}{\cal I}_{n, x}\varepsilon^{-n})+m_2\varepsilon+m_2\sum\limits_{n=1}^{\infty}{\cal I}_n\varepsilon^{-n}]+\partial x[m_3\varepsilon+m_3\sum\limits_{n=1}^{\infty}{\cal I}_n\varepsilon^{-n} \\ &+m_4\sum\limits_{n=1}^{\infty}{\cal I}_{n, 2x}\varepsilon^{-n}-m_5\phi_y(y, z)(\partial^{-1}_y\sum\limits_{n=1}^{\infty}{\cal I}_{n, x}\varepsilon^{-n})] \\ &+m_4\partial y[3(\partial^{-1}_y\sum\limits_{n=1}^{\infty}{\cal I}_{n, 2x}\varepsilon^{-n})(\partial^{-1}_y\sum\limits_{n=1}^{\infty}{\cal I}_{n, x}\varepsilon^{-n})+3q_{2x}(\partial^{-1}_y\sum\limits_{n=1}^{\infty}{\cal I}_{n, x}\varepsilon^{-n})+(\partial^{-1}_y\sum\limits_{n=1}^{\infty}{\cal I}_{n, x}\varepsilon^{-n})^3] \\ &+m_6\partial z[\partial^{-1}_y\sum\limits_{n=1}^{\infty}{\cal I}_{n, x}\varepsilon^{-n}]=0, \end{split} \end{eqnarray} $

比较$ \varepsilon $的幂次系数, 可得无穷守恒律

$ \begin{equation} \nonumber {\cal J}_{n, t}+{\cal F}_{n, x}+m_4{\cal G}_{n, y}+m_6{\cal H}_{n, z}=0, \;\;n=1, 2, 3, \cdots. \end{equation} $

在无穷守恒律(4.1)中, 守恒密度的显式公式$ {\cal J}_n $由式(4.2)给出, 第一连带流$ {\cal F}_n $、第二连带流$ {\cal G}_n $和剩余连带流$ {\cal H}_n $分别由式(4.3)、(4.4)和(4.5)给出. 证毕.

5 结论

本文研究了(3+1)维变系数广义浅水波方程(1.1)的可积性, 主要研究结果如下: 基于Bell多项式方法, 获得了方程(1.1)的双线性表达式(2.4); 进而推导出了方程(1.1)的单孤子解(2.17)、双孤子解(2.22)、三孤子解(2.23)以及N-孤子解(2.10), 通过图像发现孤波在传播过程中振幅保持不变, 且系数$ m_4 $影响孤波的传播方向; 除此之外, 运用Bell多项式方法构造了方程(1.1)的双线性Bäcklund变换(3.1)、Lax对(3.10)和无穷守恒律(4.1). 我们希望本文的论述能用于研究数学、物理等其他领域中的类似问题.

参考文献
[1] Ablowitz M J, Clarkson P A. Solitons, nonlinear evolution equations and inverse scattering[M]. Cambridge: Cambridge University Press, 1991.
[2] Hirota R. Direct methods in soliton theory[M]. Cambridge: Cambridge University Press, 2004.
[3] Ma W X, You Y C. Solving the Korteweg-de Vries equation by its bilinear form: Wronskian solutions[J]. Transactions of the American mathematical society, 2005, 357(5): 1753–1778.
[4] Wang H, Li B. Solitons for a generalized variable-coefficient nonlinear Schrödinger equation[J]. Chin. Phys. B., 2011, 20(4): 040203. DOI:10.1088/1674-1056/20/4/040203
[5] Zuo J M, Zhang Y M. The Hirota bilinear method for the coupled Burgers equation and the high-order Boussinesq-Burgers equation[J]. Chin. Phys. B., 2011, 20(1): 010205. DOI:10.1088/1674-1056/20/1/010205
[6] Wang D S, Wang X L. Long-time asymptotics and the bright N-soliton solutions of the Kundu-Eckhaus equation via the Riemann-Hilbert approach[J]. Nonlinear Anal: Real., 2018, 41(41): 334–361.
[7] Gao L N, Zi Y Y, Yin Y H, Ma W X, Lü X. Bäcklund transformation, multiple wave solutions and lump solutions to a (3 + 1)-dimensional nonlinear evolution equation[J]. Nonlinear Dyn., 2017, 89(3): 2233–2240. DOI:10.1007/s11071-017-3581-3
[8] 郭婷婷. (2+1)维KdV方程的Bäcklund变换和无穷守恒律[J]. 中北大学学报(自然科学版), 2017, 38(03): 277–281.
[9] 郝晓红, 程智龙. (2+1)维AKNS方程的可积性研究[J]. 动力学与控制学报, 2018, 16(03): 201–205.
[10] Du Z, Tian B, Chai H P, Sun Y. Darboux transformations, solitons, breathers and rogue waves for the modified Hirota equation with variable coefficients in an inhomogeneous fiber[J]. Opt Quant Electron., 2018, 50(2): 83. DOI:10.1007/s11082-017-1251-9
[11] 余兰, 张玉平, 魏光美. 贝尔多项式方法在6阶KdV方程的应用[J]. 北京信息科技大学学报(自然科学版), 2018, 33(02): 16–19.
[12] 韩鹏飞. 贝尔多项式与非线性发展方程的可积性与相关问题研究[D]. 呼和浩特: 内蒙古师范大学数学科学学院, 2021.
[13] 刘运. Bell多项式在合流范德蒙矩阵求逆中的应用[D]. 杭州: 浙江工商大学统计与数学学院, 2018.
[14] Fan E G, Chow K W. Darboux covariant Lax pairs and infinite conservation laws of the (2+1)-dimensional breaking soliton equation[J]. J. Math. Phys., 2011, 52(2): 023504. DOI:10.1063/1.3545804
[15] Gilson C R, Lambert F, Nimmo J J C, Willox R. On the combinatorics of the hirota D-operators[J]. Proc. R. Soc. Lond. Ser. A Math. Phys. Eng. Sci., 1996, 452(1945): 223–234. DOI:10.1098/rspa.1996.0013
[16] 郝晓红, 程智龙. 一类广义浅水波KdV方程的可积性研究[J]. 数学物理学报, 2019, 39(A3): 451–460.
[17] Huang Q M, Gao Y T, Jia S L, Wang Y L, Deng G F. Bilinear Bäcklund transformation, soliton and periodic wave solutions for a (3+1)-dimensional variable-coefficient generalized shallow water wave equation[J]. Nonlinear Dyn., 2017, 87(4): 2529–2540. DOI:10.1007/s11071-016-3209-z
[18] Tian S F, Zhang H Q. On the integrability of a generalized variable-coefficient Kadomtsev-Petviashvili equation[J]. J. Phys. A: Math. Theor., 2012, 45(5): 055203. DOI:10.1088/1751-8113/45/5/055203
[19] 刘娜. 非线性偏微分方程的可积性和非线性波的研究[D]. 济南: 山东师范大学数学与统计学院, 2020.
[20] 魏薇薇. 基于Bell多项式方法的孤子方程可积性的研究[D]. 金华: 浙江师范大学数学与计算机科学学院, 2012.
[21] 李晓东. 非线性偏微分方程的Hirota方法和Bell多项式研究[D]. 宁波: 宁波大学数学与统计学院, 2014.
[22] 肖子崇. 孤立子方程的Bell多项式解法[D]. 上海: 复旦大学数学科学学院, 2012.
[23] 程腾飞. Hirota方法和Bell多项式在孤子方程中的应用[D]. 金华: 浙江师范大学数学与计算机科学学院, 2011.
[24] Jimbo M, Miwa T. Solitons and infinite dimensional Lie algebras[J]. Publications of the Research Institute for Mathematical Sciences, 1983, 19(3): 943–1001. DOI:10.2977/prims/1195182017
[25] Zeng Z F, Liu J G, Nie B. Multiple-soliton solutions, soliton-type solutions and rational solutions for the (3+1)-dimensional generalized shallow water equation in oceans[J]. Estuaries and Impoundments. Nonlinear Dyn., 2016, 86(1): 667–675. DOI:10.1007/s11071-016-2914-y
[26] Tian B, Gao Y T. Beyond travelling waves: a new algorithm for solving nonlinear evolution equations[J]. Comput. Phys. Commun., 1996, 95(2-3): 139–142. DOI:10.1016/0010-4655(96)00014-8
[27] Tang Y N, Ma W X, Xu W. Grammian and Pfaffian solutions as well as Pfaffianization for a (3 + 1)-dimensional generalized shallow water equation[J]. Chin. Phys. B, 2012, 21(7): 85–91.
[28] Wu J Z, Xing X Z, Geng X G. Generalized bilinear differential operators application in a (3 + 1)-dimensional generalized shallow water equation[J]. Adv. Math. Phys. 2015, 2015: 291804.
[29] Darvishi M T, Najafi M. Some complexiton type solutions of the (3+ 1)-dimensional Jimbo-Miwa equation[J]. International Journal of Computational and Mathematical Sciences, 2012, 6(1): 25–27.
[30] Ma W X, Lee J H. A transformed rational function method and exact solutions to the 3+1dimensional Jimbo-Miwa equation[J]. Chaos, Solitons Fractals, 2009, 42(3): 1356–1363. DOI:10.1016/j.chaos.2009.03.043
[31] Zayed E M E. New traveling wave solutions for higher dimensional nonlinear evolution equations using the (G'/G)-expansion method[J]. Journal of Physics A: Mathematical and Theoretical., 2009, 42(19): 195202. DOI:10.1088/1751-8113/42/19/195202
[32] Ma H C, Huang H Y, Deng A P. Soliton molecules and some interaction solutions for the (3+1)-dimensional Jimbo-Miwa equation[J]. Journal of Geometry and Physics, 2021, 170(170): 104362.
[33] Singh M, Gupta R K. Bäcklund transformations, Lax system, conservation laws and multisoliton solutions for Jimbo-Miwa equation with Bell-polynomials[J]. Communications in Nonlinear Science and Numerical Simulation, 2016, 37: 362–373. DOI:10.1016/j.cnsns.2016.01.023
[34] Liu N. Bäcklund transformation and multi-soliton solutions for the(3+1)-dimensional BKP equation with Bell polynomials and symbolic computation[J]. Nonlinear Dyn., 2015, 82(1-2): 311–318. DOI:10.1007/s11071-015-2159-1