数学物理学报  2017, Vol. 37 Issue (2): 313-325   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
刘晶鑫
吴红霞
曾云波
无色散BKP方程族可积耦合推广及其求解
刘晶鑫1, 吴红霞1, 曾云波2     
1. 集美大学理学院数学系 福建厦门 361021;
2. 清华大学数学科学系 北京 100084
摘要:该文通过对B类Kadomtsev-Petviashvili(B type of Kadomtsev-Petviashvili,简称为BKP)方程族基于特征函数及共轭特征函数表示的对称约束取无色散极限,得到无色散BKP(dispersionless BKP,简称为dBKP)方程族的对称约束;其次,基于dBKP方程族的对称约束,考察了dBKP方程族的推广问题.通过计算推广的dBKP方程族的零曲率方程,该文导出了第一、二类型的带自相容源的dBKP方程(dispersionless BKP equation with self-consistent sources,简称为dBKPESCS)及其相应的守恒方程.最后,利用速端变换及约化的方法求解了第一型dBKPESCS.
关键词无色散BKP方程族    推广    对称约束    带自相容源的无色散BKP方程族    速端解    
On the Ingegrable Coupled Generalization of Dispersionless BKP Hierarchy and Its Solutions
Liu Jingxin1, Wu Hongxia1, Zeng Yunbo2     
1. Department of Mathematics, Jimei University, Fujian Xiamen 361021;
2. Department of Mathematical, Tsinghua University, Beijing 100084
Abstract: The symmetry constraint for dispersionless BKP (dBKP) hierarchy is firstly derived by taking dispersionless limit of that for BKP hierarchy. Then, based on the symmetry constraint for dBKP hierarchy, a new extended dBKP hierarchy is constructed. In addition, the integrability of this new extended dBKP hierarchy is proved by presenting its zero-curvature equation and the related conservation equation. From its zero-curvature equation, two types of dBKP equations with self-consistent sources (dBKPESCS) together with their associated conservation equations are obtained. Hodograph solutions for the first type of dBKPESCS are finally obtained.
Key words: Dispersionless BKP hierarchy     Extended     Symmetry constraint     Dispersionless BKP hierarchy with self-consistent sources     Hodograph solutions    
1 引言

无色散可积系统可通过对普通的可积系统取无色散极限得到, 与普通可积系统相比, 无色散可积系统没有色散项, 即系统的色散效应忽略不计.已有的研究结果表明了无色散可积系统在共形映射、流体力学和拓扑场论中都有着重要的应用[1-10].在Lax方程中, Lax算子被无色散方程族中的相空间函数所取代.此外, 泊松括号替代交换子, 并且守恒方程扮演Lax对的角色 (哈密顿雅可比方程类型).需要注意的是, 这些无色散可积系统可以通过twistorial方法[8-9], 速端约化方法[4]和拟经典极限$\overline{\partial}$-方法[10]等方法进行求解.

众所周知, 无色散Kadomtsev-Petviashvili (dispersionless Kadomtsev-Petviashvili, 简称为dKP) 方程族和无色散modified Kadomtsev-Petviashvili (dispersionless modified Kadomtsev-Petviashvili, 简称为dmKP) 方程族是两个重要的无色散可积系统, 已经在无色散Sato理论中得到广泛研究. 1992年, Dasgupta和Roy在文献[11]中探讨了dBKP方程.结果表明, 无色散BKP方程可通过对BKP方程取无色散极限得到, 该文献还探讨了dBKP方程的约化速端解.此后, dBKP方程族的研究也得到广泛进行并取得了一些重要的进展, 比如:约化速端解[12]$w$-无穷对称[11]、推广[13]$\overline{\partial}$-dressing方法[14]等.我们注意到文献[13]所给出的推广的无色散BKP方程族其实是潜伏于Takasaki所提出的D型的Landau-Ginzburg模型的一种可积系统.

作为孤立子方程族的可积推广, 带自相容源的孤立子方程一直是数学物理的研究热点 (参见文献[15-18]).众所周知, 存在两种类型的带自相容源孤子方程 (soliton equation with self-consistent sources, 简称为SESCS):第一型SESCS是由添加了附加项的孤立子方程及其特征值问题耦合而成; 而第二型SESCS却是由添加了附加项的孤立子方程及其特征函数的演化方程所构成. 2006年, 曾云波教授等通过把约束的孤子方程族看做是带源孤子方程族的静态方程, 相继探讨了第一型的带自相容源的dKP、dmKP方程族并得到了它们相应的约化速端解[19-20].然而, 我们很容易发现带自相容源的无色散BKP方程族 (dispersionless B type of KP hierarchy with self-consistent sources, 简称为dBKPHSCS), 作为dBKPH的一个新的可积推广, 至今仍然尚未被研究.

在本文中, 我们将给出dBKP方程族的另一种新的推广.首先通过对BKP方程族基于特征函数及共轭特征函数表示的对称约束取无色散极限, 得到dBKP方程族的对称约束.受文献[21]启发, 我们通过引入一个新$\tau_{k}$流, 从而推广了无色散BKP方程族.与此同时我们通过给出其零曲率方程及其相应的守恒方程证明了推广的dBKP方程族 (extended dBKPH, 简称为exdBKPH) 的可积性.再由exdBKPH的零曲率形式, 我们导出了第一型和第二型的dBKPESCS以及它们相应的守恒方程.因此, 我们说这个新exdBKPH可以提供一种有效的和普遍的方法, 去导出$(2+1) $维带自相容源的无色散孤子方程以及它们相应的守恒方程.最后, 应用约化速端方法求解出第一型dBKPESCS.本文的结果不但是前人研究的一个重要补充, 而且也提供了dBKP方程族一个更完整的刻画.

2 无色散BKP方程族

让我们回顾一下BKP方程族和dBKP方程族.由文献[22]知, KP方程族的Lax方程为

$ \widetilde{L}_{t_n}=[\widetilde{B}_n, \widetilde{L}], $ (2.1a)

其中

$ \widetilde{L}=\partial+u_1\partial^{-1}+u_2\partial^{-2}+\cdots $ (2.1b)

是拟微分算子, $\partial$表示为$\frac{\partial}{\partial_x}, u_i, i=1, 2, \cdots, $$t=(t_1, t_2, t_3, \cdots)$$t_1=x$的函数, $\widetilde{B}_n=(\widetilde{L}^n)_{\geq0}$表示$\widetilde{L}^n$的微分部分.由文献[23]知, BKP方程族可由KP方程族添加如下的约束条件得到

$ \widetilde{L}^*=-\partial\widetilde{L}\partial^{-1}, $ (2.1c)

容易证明由条件 (2.1c) 可以推导出$\widetilde{B}_n$, $n=3, 5, \cdots$的常数项和偶数时间变量均可以被消掉 (i, e., $t_2=t_4=\cdots=0)$.此外, 方程族 (2.1c) 也可以推导出

$ u_2=-u'_1, u_4=-2u'_3+u_1^{(3)}, \cdots. $

文献[20]通过对BKP方程族取无色散极限, 得到了dBKP方程族, 其定义如下

$ \partial_{T_n}L=\{B_n, L\}, $ (2.2a)

其中$n$为奇数, $L$为定义如下的Sato函数

$ L=p+\sum^{\infty}_{i=1}U_{2i-1}(T)p^{-2i+1}=p+U_1(T)p^{-1}+U_3(T)p^{-3}+\cdots, $ (2.2b)

其中$L^n$是关于$p, T=(T_1=X, T_3, T_5, \cdots)$的函数, $B_n=(L^n)_{\geq1}$表示$L^n$关于$p$的正次幂部分, 并且定义泊松括号为

$ \{A(p, X), B(p, X)\}=\frac{\partial A}{\partial p}\frac{\partial B}{\partial X}-\frac{\partial A}{\partial X}\frac{\partial B}{\partial p}. $

由方程族 (2.2a) 的交换性可以导出dBKP方程族的零曲率形式

$ \frac{\partial B_m}{\partial T_n}-\frac{\partial B_n}{\partial T_m}+\{B_m, B_n\}=0, $ (2.3)

其中$m, n$均为奇数.当$n=3, m=5$时, 方程族 (2.3) 化为dBKP方程族的最简非平凡流, 即dBKP方程

$ 3U_T-\frac53\partial^{-1}_XU_{YY}-5UU_Y-5U_X\partial^{-1}_XU_Y+15U^2U_X=0, $ (2.4)

其中$U_1=U, T_3=Y, T_5=T$.方程 (2.4) 的守恒方程为

$ \begin{array}{l} p_{Y}=(p^3+3Up)_X, \\ p_{T}=p^5+5Up^3+5(U_3+2U^2)p, \end{array} $ (2.5)

其中$U_3=\frac13\partial^{-1}_XU_Y-U^2$.

3 无色散BKP方程族的对称约束

为了推广dBKP方程族, 我们需要得到dBKP方程族的对称约束.为此, 我们将BKP方程族的对称约束取无色散极限, 从而导出dBKP方程族的对称约束.由文献[24]知BKP方程族的对称约束

$ \widetilde{L}^n=\widetilde{B}_n+\sum^N_{i=1}(r_i\partial^{-1}(q_i)_{x}-q_i\partial^{-1}(r_i)_{x}), $ (3.1)

其中$n, \widetilde{L}, \widetilde{B}_n$如方程族 (2.1) 所定义, 并且$q_i, r_i$满足

$ q_{i, t_n}=\widetilde{B}_n(q_i) , r_{i, t_n}=\widetilde{B}_n(r_i), i=1, \cdots, N. $ (3.2)

下面我们将实施标准的无色散极限过程.令$T_n=\varepsilon t_n$并且考虑当$\varepsilon\rightarrow0$时, $u_n(\frac T\varepsilon)=U_n(T)+O(\varepsilon)$, 则$\widetilde{L}$可化为

$ \widetilde{L}_\varepsilon=\varepsilon\partial+\sum^\infty_{i=1}u_i(\frac T\varepsilon)(\varepsilon\partial)^{-i}=\varepsilon\partial+\sum^\infty_{i=1}(U_i(T)+O(\varepsilon))(\varepsilon\partial)^{-i} , \partial=\partial_X, X=\varepsilon x, $ (3.3)

与此同时方程族 (3.1) 亦可变为

$ \widetilde{L}^n_\varepsilon=\widetilde{B}_{\varepsilon n}+\sum^N_{i=1} \bigg\{r_i(\frac T\varepsilon)(\varepsilon\partial)^{-1}(\varepsilon[q_i(\frac T\varepsilon)]_{X})-q_i(\frac T\varepsilon)(\varepsilon\partial)^{-1}(\varepsilon[r_i(\frac T\varepsilon)]_{X})\bigg\}, \widetilde{B}_{\varepsilon n}=(\widetilde{L}_\varepsilon^n)_{\geq1}, $ (3.4)

其中$q_i(\frac T\varepsilon)$$r_i(\frac T\varepsilon)$分别满足

$ \varepsilon\bigg[q_i(\frac T\varepsilon)\bigg]_{T_n}=\widetilde{B}_{\varepsilon n}q_i(\frac T\varepsilon), \varepsilon\bigg[r_i(\frac T\varepsilon)\bigg]_{T_n}=\widetilde{B}_{\varepsilon n}r_i(\frac T\varepsilon), i=1, \cdots, N. $ (3.5)

考虑

$ \begin{array}{l} q_i(\frac T\varepsilon)\sim\exp\bigg[\frac {S(T, \lambda_i)}\varepsilon+\alpha_{i1}+O(\varepsilon)\bigg], \varepsilon\rightarrow0, \\ \varepsilon\big[r_i(\frac T\varepsilon)\big]_X\sim \exp\bigg[-\frac {S(T, \lambda_i)}\varepsilon+\alpha_{i2}+O(\varepsilon)\bigg], \varepsilon\rightarrow0. \end{array} $ (3.6)

我们易得到

$ \begin{array}[b]{rl} r_i(\frac T\varepsilon)(\varepsilon\partial)^{-1}(\varepsilon[q_i(\frac T\varepsilon)]_X)=&\frac {\exp(-\frac{S(T, \lambda_i)}\varepsilon+\alpha_{i2}+O(\varepsilon))\cdot \exp(\frac {S(T, \lambda_i)}\varepsilon+\alpha_{i1}+O(\varepsilon))}{-\frac {\partial S(T, \lambda_i)}{\partial X}}\\ &\times\frac {\partial S(T, \lambda_i)}{\partial X}\times\bigg[(\varepsilon\partial)^{-1}-(\frac {\partial S(T, \lambda_i)}{\partial X}+O(\varepsilon))(\varepsilon\partial)^{-2}+\cdots\\ &+(-1)^n((\frac {\partial S(T, \lambda_i)}{\partial X})^n+O(\varepsilon))(\varepsilon\partial)^{-(n+1)}+\cdots\bigg], \end{array} $ (3.7a)

$ \begin{array}[b]{rl} q_i(\frac T\varepsilon)(\varepsilon\partial)^{-1}(\varepsilon[r_i(\frac T\varepsilon)]_X)=& \exp(\alpha_{i1}+\alpha_{i2})\times\bigg[(\varepsilon\partial)^{-1}+(\frac {\partial S(T, \lambda_i)}{\partial X}+O(\varepsilon))(\varepsilon\partial)^{-2}\\ &+\cdots+((\frac {\partial S(T, \lambda_i)}{\partial X})^n+O(\varepsilon))(\varepsilon\partial)^{-(n+1)}+\cdots\bigg]. \end{array} $ (3.7b)

对方程族 (3.4) 两边同时取主符号, 并结合 (3.7) 我们即求得dBKP方程族如下的对称约束

$ \begin{eqnarray} L^n &=&B_n+\sum^N_{i=1} \bigg\{-{\rm e}^{\alpha_{i1}+\alpha_{i2}}\bigg[p^{-1}-\frac {\partial S(T, \lambda_i)}{\partial X}p^{-2}+(\frac {\partial S(T, \lambda_i)}{\partial X})^2p^{-3}+\cdots\bigg]\\ &&-{\rm e}^{\alpha_{i1}+\alpha_{i2}}\bigg[p^{-1}+\frac {\partial S(T, \lambda_i)}{\partial X}p^{-2}+(\frac {\partial S(T, \lambda_i)}{\partial X})^2p^{-3}+\cdots\bigg]\bigg\}\\ &=&B_n+\sum^N_{i=1} \bigg(-\frac{\frac12v_i}{p-p_i}-\frac{\frac12v_i}{p+p_i}\bigg)\\ &=&B_n-\sum^N_{i=1}\frac{v_ip}{p^2-p^2_i}, \end{eqnarray}% $ (3.8)

其中$v_i=2{\rm e}^{\alpha_{i1}+\alpha_{i2}}, p_i=\frac {\partial S(T, \lambda_i)}{\partial X}.$

注3.1  在这里我们再次证明了无色散可积方程族的对称约束可以通过相应可积方程族的对称约束取无色散极限得到.当然, 我们发现了利用无色散极限的方法所求得的dBKP方程族的对称约束 (3.8), 与文献[14]通过$\overline{\partial}$-方法导出的相一致.

4 推广的无色散BKP方程族

基于上面我们所导出的dBKP方程族的对称约束, 我们将通过引入$L$新的$\tau_k$流, 来定义一个新的推广的dBKP方程族. $L$$\tau_k$流定义如下

$ L_{\tau_k}=\bigg\{B_k-\sum^N_{i=1}\frac{v_ip}{p^2-p_i^2}, L\bigg\}, $ (4.1a)
$ p_{i, T_n}=[B_n(p_i)]_X, v_{i, T_n}=\bigg[v_i(\frac{\partial B_n(p_i)}{\partial p_i})\bigg]_X, i=1, \cdots, N, $ (4.1b)

其中$n, k$是奇数, $B_n$如 (2.2) 式所定义.

定义4.1  新的推广的dBKP定义为

$ L_{\tau_k}=\bigg\{B_k-\sum^N_{i=1}\frac {v_ip}{p^2-p^2_i}, L\bigg\}, $ (4.2a)
$ L_{T_n}=\{B_n, L\}, n\neq k, $ (4.2b)
$ p_{i, T_n}=(B_n(p)\mid_{p=p_i})_X, $ (4.2c)
$ v_{i, T_n}=\bigg[v_i(\frac {\partial B_n(p)}{\partial p}\mid_{p=p_i}\bigg]_X, $ (4.2d)

其中$n, k$是奇数.

注4.1  我们很容易发现当$v_i=p_i=0$时, (4.2) 式恰好就是dBKP方程族, 因此 (4.2) 式就是dBKP方程族的一种推广.此外, 我们还发现 (4.2) 式是完全不同于文献[13]所给出的exdBKP方程族.在这种意义下, 我们称 (4.2) 式是另一种新的广义dBKP方程族.

为了证明 (4.2a) 式和 (4.2b) 式在 (4.2c) 式和 (4.2d) 式下的交换性, 我们首先证明以下引理.

引理4.1  恒等式$(\frac{v_ip}{p^2-p^2_i})_{T_n}=\{B_n, \frac{v_ip}{p^2-p^2_i}\}_{-}$, 对$i=1, \cdots, N$均成立.

  令

$ B_n=(L^n)_{\geq1}=\sum\limits_{1\leq k\leq n}U_k(T)p^k, $

其中$k$是奇数.则对于$i=1, \cdots, N$我们得到

$ \begin{eqnarray} \bigg\{B_n, \frac{v_ip}{p^2-p^2_i}\bigg\}_{-}&=& \bigg[\frac {\partial B_n}{\partial p}\cdot\frac \partial{\partial X}(\frac{v_ip}{p^2-p^2_i})-\frac {\partial B_n}{\partial X}\times\frac \partial{\partial p}(\frac{v_ip}{p^2-p^2_i})\bigg]_{-}\\ &=&\bigg[ \bigg(\sum\limits_{1\leq k\leq n}kU_kp^{k-1}\bigg) \bigg(\frac{v_{i, X}p}{p^2-p^2_i}+\frac{2v_ip_ip_{i, X}p}{(p^2-p^2_i)^2}\bigg)\\ &&- \bigg(\sum\limits_{1\leq k\leq n}U_{k, X}p^k\bigg) \bigg(\frac{v_i}{p^2-p^2_i}-\frac{2v_ip^2}{(p^2-p^2_i)^2}\bigg)\bigg]_{-}\\ &=&v_{i, X}\sum\limits_{1\leq k\leq n}kU_k(\frac{p^k}{p^2-p^2_i})_{-}-v_i\sum\limits_{1\leq k\leq n}U_{k, X}(\frac{p^k}{p^2-p^2_i})_{-}\\ & &+\sum\limits_{1\leq k\leq n}kU_k\cdot 2v_ip_ip_{i, X}\bigg[\frac{p^k}{(p^2-p^2_i)^2}\bigg]_{-}\\ &&+\sum\limits_{1\leq k\leq n}U_{k, X}\cdot2v_i\bigg[\frac{p^{k+2}}{(p^2-p^2_i)^2}\bigg]_{-}, \end{eqnarray} $ (4.3)

注意到

$ \bigg(\frac{p^k}{p^2-p^2_i}\bigg)_{-}=\frac{pp^{k-1}_i}{p^2-p^2_i}, \bigg[\frac{2p^k}{(p^2-p^2_i)^2}\bigg]_{-}=(k-1)p_i^{k-3}\frac p{p^2-p^2_i}+p_i^{k-1}\frac {2p}{({p^2-p^2_i})^2}, $

$ \begin{eqnarray} (4.3)&=&\frac{v_{i, X}\sum\limits_{1\leq k\leq n}kU_kp_i^{k-1}}{p^2-p^2_i}-\frac{v_i\sum\limits_{1\leq k\leq n}U_{k, X}pp_i^{k-1}}{p^2-p^2_i}\\ &&+\sum\limits_{1\leq k\leq n}kU_kv_ip_ip_{i, X}\bigg[(k-1)p_i^{k-3} \frac p{p^2-p^2_i}+p_i^{k-1}\frac {2p}{(p^2-p^2_i)^2}\bigg]\\ &&+\sum\limits_{1\leq k\leq n}U_{k, X}v_i\bigg[(k+1)p_i^{k-1}\frac p{p^2-p^2_i}+p_i^{k+1}\frac {2p}{(p^2-p^2_i)^2}\bigg]\\ &=&\frac{\Big[v_{i, X}\sum\limits_{1\leq k\leq n}kU_kp_i^{k-1}+v_i\sum\limits_{1\leq k\leq n}kU_{k, X}p_i^{k-1}+v_i\sum\limits_{3\leq k\leq n}k(k-1)U_kp_i^{k-2}p_{i, X}\Big]p}{p^2-p^2_i}\\ &&+\frac{2v_ip_i \Big(\sum\limits_{1\leq k\leq n}U_{k, X}p_i^k+\sum\limits_{1\leq k\leq n}kU_kp_i^{k-1}p_{i, X}\Big)p}{(p^2-p^2_i)^2}. \end{eqnarray} $ (4.4)

此外, 我们发现从 (4.2c) 式和 (4.2d) 式可以得到

$ \begin{eqnarray} \bigg(\frac{v_ip}{p^2-p^2_i}\bigg)_{T_n}&=& \frac{v_{i, T_n}p}{p^2-p^2_i}+\frac{2v_ip_ipp_{i, T_n}}{(p^2-p^2_i)^2}\\ &=&\frac{p\Big[v_i(\frac{\partial B_n(p)}{\partial p})\mid_{p=p_i}\Big]_X}{p^2-p^2_i}+\frac{2v_ip_ip [B_n(p)\mid_{p=p_i}]_X}{(p^2-p^2_i)^2} \\ &=&\frac{\Big[v_{i, X}\sum\limits_{1\leq k\leq n}kU_kp_i^{k-1}+v_i\sum\limits_{1\leq k\leq n}kU_{k, X}p_i^{k-1}+v_i\sum\limits_{3\leq k\leq n}k(k-1)U_kp_i^{k-2}p_{i, X}\Big]p}{p^2-p^2_i}\\ &&+\frac{2v_ip_i\Big(\sum\limits_{1\leq k\leq n}U_{k, X}p_i^k+\sum\limits_{1\leq k\leq n}kU_kp_i^{k-1}p_{i, X}\Big)p}{(p^2-p^2_i)^2}, \end{eqnarray} $ (4.5)

再结合 (4.4) 式和 (4.5) 式, 我们得到

$ \bigg(\frac{v_ip}{p^2-p^2_i}\bigg)_{T_n}= \bigg\{B_n, \frac{v_ip}{p^2-p^2_i}\bigg\}_{-}. $

即证.

定理4.1  在 (4.2c) 和 (4.2d) 式成立的条件下, (4.2a) 和 (4.2b) 式可以导出推广dBKP方程族 (4.2) 的零曲率方程

$ B_{n, \tau_k}- \bigg(B_k-\sum^N_{i=1}\frac{v_ip}{p^2-p^2_i}\bigg)_{T_n}+ \bigg\{B_n, B_k-\sum^N_{i=1} \frac{v_ip}{p^2-p^2_i}\bigg\}=0, $ (4.6a)

其等价形式

$ B_{n, \tau_k}-B_{k, T_n}+\{B_n, B_k\}+\sum^N_{i=1} \bigg[\bigg(\frac{v_ip}{p^2-p^2_i}\bigg)_{T_n}- \bigg\{B_n, \frac{v_ip}{p^2-p^2_i}\bigg\}\bigg]=0, $ (4.6b)

其中$v_i, p_i$如 (4.2c) 式和 (4.2d) 式所定义.在 (4.2c) 式和 (4.2d) 式下, 我们给出了方程族 (4.6a) 或方程族 (4.6b) 的守恒方程

$ p_{T_n}=[B_n(p)]_X, p_{\tau_k}=\bigg[B_k(p)-\sum^N_{i=1}\frac{v_ip}{p^2-p^2_i}\bigg]_X. $ (4.7)

  我们首先证明在 (4.2c, d) 式成立的条件下, (4.2a) 和 (4.2b) 可以导出方程 (4.6a).由 (4.2a), (4.2b) 和引理4.1, 我们得到

$ \begin{eqnarray*} B_{n, \tau_k}&=&(L^n_{\tau_k})_{+}= \bigg\{B_k-\sum^N_{i=1}\frac{v_ip}{p^2-p^2_i}, L^n\bigg\}_{+}\\ &=&\bigg\{B_k-\sum^N_{i=1}\frac{v_ip}{p^2-p^2_i}, B_n\bigg\}_{+}+ \bigg\{B_k-\sum^N_{i=1}\frac{v_ip}{p^2-p^2_i}, (L^n)_{-}\bigg\}_{+}\\ &=&\bigg\{B_k-\sum^N_{i=1}\frac{v_ip}{p^2-p^2_i}, B_n\bigg\}- \bigg\{B_k-\sum^N_{i=1}\frac{v_ip}{p^2-p^2_i}, B_n\bigg\}_{-} +\{B_k, (L^n)_{-}\}_{+}\\ &=&\bigg\{B_k-\sum^N_{i=1}\frac{v_ip}{p^2-p^2_i}, B_n\bigg\}+ \bigg\{B_n, -\sum^N_{i=1}\frac{v_ip}{p^2-p^2_i}\bigg\}_{-}+\{B_k, (L^n)_{-}\}_{+}\\ &=&\bigg\{B_k-\sum^N_{i=1}\frac{v_ip}{p^2-p^2_i}, B_n\bigg\}+ \bigg\{B_n, -\sum^N_{i=1}\frac{v_ip}{p^2-p^2_i}\bigg\}_{-}+B_{k, T_n}\\ &=&\bigg\{B_k-\sum^N_{i=1}\frac{v_ip}{p^2-p^2_i}, B_n\bigg\}+ \bigg(B_k-\sum^N_{i=1}\frac{v_ip}{p^2-p^2_i}\bigg)_{T_n}. \end{eqnarray*} $

即证.

注4.2  在这里我们通过给出推广的dBKP方程族 (4.2) 的零曲率方程及其相应的守恒方程, 证明了推广的dBKP方程族 (4.2) 是可积的.此外, 我们还发现方程族 (4.6b) 本质上就是带自相容源的无色散BKP方程族, 而此方程族也是首次得到的.

接下来, 我们将从 (4.6) 式中导出一些重要的方程.当$n=3, k=5$时, 方程族 (4.6) 为第一型带自相容源的无色散BKP方程方程族 (dBKPESCS)

$ 3U_{\tau_5}-\frac53\partial_X^{-1}U_{T_3T_3}-5UU_{T_3} -5U_X\partial^{-1}_XU_{T_3}+15U^2U_X-3\sum^N_{i=1}v_{i, X}=0, $ (4.8a)
$ p_{i, T_3}=(p^3_i+3Up_i)_X, v_{i, T_3}=[v_i(3p^2_i+3U)]_X, i=1, \cdots, N. $ (4.8b)

第一型dBKPESCS (4.8) 的守恒方程为

$ \begin{array}{l} p_{T_3}=(p^3+3Up)_X, \\[2mm] p_{\tau_5}=\bigg[p^5+5Up^3+5(U_3+2U)p-\sum^N_{i=1}\frac{v_ip}{p^2-p^2_i}\bigg]_X, \end{array} $ (4.9)

其中$U_1=U, U_3=\frac13\partial_X^{-1}U_{T_3}-U^2$.

$n=5, k=3$时, (4.6) 式化为第二型dBKPESCS

$ 3U_{T_5}-\frac53\partial_X^{-1}U_{\tau_3\tau_3}-5UU_{\tau_3}-5U_X\partial_X^{-1}U_{\tau_3}+15U^2U_X-\frac53\sum\limits_{i=1}^Nv_{i, \tau_3} $
$ -70U\sum\limits_{i=1}^Nv_{i, X}-30U_X\sum^N_{i=1}v_i-5\sum^N_{i=1}v_{i, X}p_i^2-50\sum^N_{i=1}v_ip_ip_{i, X}=0, $ (4.10a)
$ \begin{array}{l} p_{i, T_5}= \bigg[p^5_i+5Up_i^3+5 \bigg(\frac13\partial^{-1}_XU_{\tau_3}+U^2+\frac13\sum^N_{i=1}v_i\bigg)p_i\bigg]_X, \\ v_{i, T_5}= \bigg(v_i\bigg[5p^4_i+15Up^2_i+5\bigg(\frac13\partial_X^{-1}U_{\tau_3}+U^2 +\frac13\sum^N_{i=1}v_i\bigg)\bigg]\bigg)_X, i=1, \cdots, N. \end{array} $ (4.10b)

第二型dBKPESCS的守恒方程为

$ \begin{array}{l} p_{T_5}=\bigg[p^5+5Up^5+5 \bigg(\frac13\partial^{-1}_XU_{T_3}+U^2+\frac13\sum^N_{i=1}v_i\bigg)p\bigg]_X, \\ p_{\tau_3}=\bigg(p^3+3Up-\sum^N_{i=1}\frac{v_ip}{p^2-p^2_i}\bigg)_X. \end{array} $ (4.11)
5 第一型dBKPESCS的速端解

在本节中, 将利用约化及速端变换的方法来求解第一型dBKPESCS (4.8), 并考虑在$M=1$$M=2$情况下的解.受文献[4]启发, 我们需对方程 (4.9) 进行$M$-约化, 使得动量函数$p, p_i, v_i, i=1, 2, \cdots, N$仅依赖于一组函数$W=(W_1, W_2, \cdots, W_M)$, 其中$W_1=U$, 且$(W_1, W_2, \cdots, W_M)$满足交换流

$ \begin{equation} \frac{\partial W}{\partial T_n}=A_n(W)\frac{\partial W}{\partial X}, n\geq2, \end{equation} $ (5.1)

其中$A_n$$M\times M$的矩阵且是关于$(W_1, W_2, \cdots, W_M)$的函数.

(1) $M=1$.

$M=1, p=p(U), v_i=v_i(U), p_i=p_i(U)$时, 方程 (5.1) 化为

$ \begin{equation} U_Y=A(U)U_X, U_T=B(U)U_X, \end{equation} $ (5.2)

其中$\tau_5=T, T_3=Y$.

由方程 (4.8b) 和方程 (4.9) 中, 我们将得到以下关于$p, p_i, v_i$的方程, $i=1, 2, \cdots, N, $

$ \begin{eqnarray} (A-3p^2_i-3U)\frac{dp_i}{dU}&=&3p_i, \\ (A-3p^2-3U)\frac{dp}{dU}&=&3p, \\ (A-3p^2_i-3U)\frac{dv_i}{dU}&=&3v_i(2p_i\frac{dp_i}{dU}+1), \\ B\frac{dp}{dU}&=&5p^4\frac{dp}{dU}+5p^3+15Up^2\frac{dp}{dU}+5p\frac{dU_3}{dU}\\ &&+20pU+(5U_3+10U^2)\frac{dp}{dU} \\ &&-\sum^N_{i=1}\frac{p\frac{dv_i}{dU}+v_i\frac{dp}{dU}}{p^2-p^2_i} +\sum^N_{i=1}\frac{2pv_i(p\frac{dp}{dU}-p_i\frac{dp_i}{dU})}{(p^2-p^2_i)^2}. \end{eqnarray} $ (5.3)

由方程 (5.3) 导出

$ B=\frac{5A^2}9+\frac53AU+5U_3+\sum^N_{i=1}\frac{dv_i}{dU}, \quad A=3\frac{dU_3}{dU}+2U. $ (5.4)

此时很容易验证在 (5.3) 式及 (5.4) 式下, (5.2) 式是相容的.在速端变换下将变量转换为$(X, Y, T)\rightarrow(U, Y, T)$$X=X(U, Y, T)$, 我们将得到下列关于$X$的速端方程

$ \frac{\partial X}{\partial Y}=-A, \quad \frac{\partial X}{\partial T}=-B=-\bigg(\frac{5A^2}9+\frac53AU+5U_3+\sum^N_{i=1}\frac{dv_i}{dU}\bigg). $ (5.5)

我们注意到方程 (5.5) 容易被积分, 从而获得 (5.2) 式的显式解, 进而获得方程 (4.8) 的解

$ X+A(U)Y+\bigg(\frac{5A^2}9+\frac53AU+5U_3+\sum^N_{i=1}\frac{dv_i}{dU}\bigg)T=F(U), $ (5.6)

这里$F(U)$是一个关于$U$的任意函数.

如果我们选择$A(U)=C$, $C$为常数, $F(U)=\alpha U$, 从 (5.4) 式和 (5.6) 式, 我们将得到第一型dBKPESCS的一个隐式解

$ X+CY+ \bigg[\frac{10}3CU+\frac59C^2-\frac53U^2+\sum^N_{i=1}\frac{3c_i(C+3p_i^2-3U)}{(C-3p_i^2-3U)^3} \bigg]T=\alpha U, $ (5.7a)
$ v_i=\frac{c_i}{C-3p^2_i-3U}, $ (5.7b)
$ p_i=\sqrt[3]{-\frac{d_i}2+\sqrt{\frac{d_i^2}4+\frac{(3U-C)^3}{27}}}+\sqrt[3]{-\frac{d_i}2-\sqrt{\frac{d_i^2}4+\frac{(3U-C)^3}{27}}}, $ (5.7c)

其中$c_i, d_i, i=1, \cdots, N$是常数.

如果我们取$d_1=d_2=\cdots=d_N, \sum\limits^N_{i=1}c_i=0$, 等式 (5.7a) 将退化为dBKP方程[11]的速端解.

(2) $M=2$.

在这种情况下, 我们令$W_1=U, W_2=W, v_i=v_i(U, W), p_i=p_i(U, W)$$p=p(U, W)$. $(U, W)$满足如下交换流

$ \left(\begin{array}{c} U\\ W\end{array}\right)_Y=A \left(\begin{array}{c} {U_X}\\ {W_X} \end{array}\right), \quad \left(\begin{array}{c} U\\ W\end{array}\right)_T =B\left(\begin{array}{c} {U_X}\\ {W_X} \end{array}\right), $ (5.8)

这里$A=(A_{ij})$$B=(B_{ij})$分别是关于函数$U$和函数$W$$2\times2$矩阵.若$U_X$$W_X$是相互独立的, 则由方程 (4.8b) 和方程 (4.9) 得到如下关于$v_i(U, W), p_i(U, W)$$p(U, W)$的方程

$ \begin{eqnarray} \bigg(\frac{\partial p}{\partial U}, \frac{\partial p}{\partial W}\bigg)A&=&3p^2\bigg(\frac{\partial p}{\partial U}, \frac{\partial p}{\partial W}\bigg)+3U\bigg(\frac{\partial p}{\partial U}, \frac{\partial p}{\partial W}\bigg)+3p(1, 0), \\ \bigg(\frac{\partial p_i}{\partial U}, \frac{\partial p_i}{\partial W}\bigg)A&=&3p^2_i\bigg(\frac{\partial p_i}{\partial U}, \frac{\partial p_i}{\partial W}\bigg)+3U\bigg(\frac{\partial p_i}{\partial U}, \frac{\partial p_i}{\partial W}\bigg)+3p_i(1, 0), \\ \bigg(\frac{\partial v_i}{\partial U}, \frac{\partial v_i}{\partial W}\bigg)A&=&(3p^2_i+3U)\bigg(\frac{\partial v_i}{\partial U}, \frac{\partial v_i}{\partial W}\bigg)+6p_iv_i\bigg(\frac{\partial p_i}{\partial U}, \frac{\partial p_i}{\partial W}\bigg)+3v_i(1, 0), \\ \bigg(\frac{\partial p}{\partial U}, \frac{\partial p}{\partial W}\bigg)B&=&[5p^4+15Up^2+5(U_3+2U^2)] \bigg(\frac{\partial p}{\partial U}, \frac{\partial p}{\partial W}\bigg)+5p\bigg(\frac{\partial U_3}{\partial U}, \frac{\partial U_3}{\partial W}\bigg)\\ &&+ (20Up+5p^3)(1, 0)+\sum^N_{i=1}\frac{v_ip}{(p^2-p^2_i)^2} \bigg(\frac{\partial (p^2-p^2_i)}{\partial U}, \frac{\partial (p^2-p^2_i)}{\partial W}\bigg)\\ &&-\sum^N_{i=1}\frac1{(p^2-p^2_i)}\bigg(\frac{\partial (v_ip)}{\partial U}, \frac{\partial (v_ip)}{\partial W}\bigg). \end{eqnarray} $ (5.9)

我们从方程 (5.9) 中可以很容易地发现, $A(U, W)$$B(U, W)$必须满足

$ B=\frac{5A^2}9+\frac53AU+5U_3I+\sum^N_{i=1}\frac{\partial v_i}{\partial U}I- \left(\begin{array}{cc} 0 ~~& \frac{\partial v_i}{\partial W} \\ \frac{A_{21}}{A_{12}}\frac{\partial v_i}{\partial W} ~~& \frac {A_{22}-A_{11}}{A_{12}}\frac{\partial v_i}{\partial W} \end{array}\right), $ (5.10)

这里$I$$2\times2$的恒等矩阵, $A_{11}=6U+3\frac{\partial U_3}{\partial U}$$A_{12}=3\frac{\partial U_3}{\partial W}$.

为简单起见, 我们假设

$ \frac{\partial v_i}{\partial W}=0, ~~ i=1, \cdots, N, $

通过应用公式

$ A^2=({\rm tr}A)A-(\det A)I, $ (5.11)

我们得到

$ B=\bigg(\frac59{\rm tr}A+\frac53U\bigg)A+ \bigg(5U_3-\frac59\det A+\sum^N_{i=1}\frac{\partial v_i}{\partial U}\bigg)I, $ (5.12)

其中$\det A=A_{11}A_{22}-A_{12}A_{21}, {\rm tr}A=A_{11}+A_{22}.$

由 (5.8) 式和 (5.12) 式的兼容性条件, 要求$A$满足

$ \left(\begin{array}{c} { -\frac{\partial(5U_3-\frac59\det A)}{\partial W}}\\ { \frac{\partial(5U_3-\frac59\det A+\sum\limits^N_{i=1}\frac{\partial v_i}{\partial U})}{\partial U}} \end{array}\right) =A\left(\begin{array}{c} { \frac{\frac59\partial(trA)}{\partial W}}\\ { -\frac59\frac{\partial(trA)}{\partial U}-\frac53} \end{array}\right) . $ (5.13)

为了求解 (5.8) 式, 我们作速端变换把变量$(X, Y, T)$变为$(U, W, T)$, 其中

$ X=X(U, W, T), ~~Y=Y(U, W, T). $

在新的变换下, (5.8) 式变为

$ \left(\begin{array}{c} {-X_W}\\ {X_U}\end{array}\right) =A \left(\begin{array}{c} {Y_W}\\ {-Y_U}\end{array}\right), \quad \left(\begin{array}{c} { \frac{\partial(X, Y)}{\partial(W, T)}}\\ { -\frac{\partial(X, Y)}{\partial(U, T)}}\end{array}\right) =B \left(\begin{array}{c} {Y_W}\\ {-Y_U} \end{array}\right) $ (5.14)

其中$\frac{\partial(X, Y)}{\partial(W, T)}=X_WY_T-X_TY_W$.

可以很容易地发现, (5.14) 式存在如下解的形式

$ \begin{array}{l} X+\bigg(5U_3-\frac59\det A+\sum^N_{i=1}\frac{\partial v_i}{\partial U}\bigg)T=F(U, W), \\ Y+\bigg(\frac59{\rm tr}A+\frac53U\bigg)T=G(U, W), \end{array} $ (5.15)

其中, $Y_U$$Y_W$是相互独立的, $F, G$满足线性方程

$ \left(\begin{array}{c} {-F_W}\\ {F_U} \end{array}\right) =A\left(\begin{array}{c} {G_W}\\ {-G_U}\end{array}\right). $ (5.16)

在这里我们给出一个简单的列子

$ A= \left[\begin{array}{cc} 3W^2+6U~~&6WU \\ 3W~~&3W^2+3U \end{array}\right], $ (5.17)

并且

$ p_i=W, v_i=c_iU, U_3=W^2U, i=1, \cdots, N, $ (5.18)

其中$c_i, i=1, \cdots, N$是常数.

由此, (5.15) 式变化为

$ X-\bigg(5W^4+10U^2-\sum^N_{i=1}c_i\bigg)T=F(U, W), $ (5.19a)
$ Y+\bigg(\frac {10}3W^2+\frac{20}3U\bigg)T=G(U, W). $ (5.19b)

结合方程 (5.16) 和$F_{WU}=F_{UW}, G$需要满足

$ 4WG_U+2WUG_{UU}-WG_{WW}-3G_W=0. $

如果我们选择$G=2U+W^2$, 再由方程 (5.16) 得到

$ F=-\frac23W^4-3U^2, $

通过计算我们获得第一型dBKPESCS (4.8) 的一个显式解

$ U=\frac {-(3\xi+5T)\pm\sqrt{(15T-\frac92)\xi^2+(10T-9)\eta+30\xi T+25T^2}}{9-10T}, $
$ p_i=\sqrt{-2U-\xi}=\sqrt{\frac{3\xi-10T\xi-10T\pm2\sqrt{(15T-\frac92)\xi^2+(10T-9)\eta+30\xi T+25T^2}}{10T-9}}, $ (5.20)
$ v_i=c_iU=c_i[\frac{-(3\xi+5T)\pm\sqrt{(15T-\frac92)\xi^2+(10T-9)\eta+30\xi T+25T^2}}{9-10T}, $

其中$\xi=\frac{Y}{\frac{10}3T-1}, \eta=X+5\xi T+\Big(\sum\limits^N_{i=1}c_i\Big)T, i=1, \cdots, N.$

6 结论和探讨

本文研究的是dBKP方程族的另一种新的推广.为了推广dBKP方程族, 本文通过对BKP方程族基于特征函数及共轭特征函数表示的对称约束取无色散极限, 得到无色散BKP (dBKP) 方程族的对称约束; 其次, 基于dBKP方程族的对称约束, 构造了新的推广的dBKP方程族.通过计算零曲率方程, 我们导出第一、二类型的带自相容源的dBKP方程 (dBKPESCS), 及其相应的守恒方程.最后, 利用速端变换及约化的方法求解了第一型dBKPESCS.本文的结果不但是前人研究的一个重要补充, 而且也提供了dBKP方程族一个更完整的刻画.当然我们也注意到一个问题, 那就是如何求解第二型的带自相容源的无色散BKP方程, 始终没有解决.这个问题, 将在我们未来的研究工作中得到进一步讨论.

参考文献
[1] Mineev-Weinstein M, Wiegmann P B, Zabrodin A. Integrable structure of interface dynamics. Phys Rev Lett, 2000, 84(22): 5106–5109. DOI:10.1103/PhysRevLett.84.5106
[2] Gibbons J, Tsarev S P. Reductions of the Benney equations. Phys Lett A, 1996, 211(1): 19–24. DOI:10.1016/0375-9601(95)00954-X
[3] Gibbons J, Tsarev S P. Conformal maps and reductions of the Benney equation. Phys Lett A, 1999, 258(4-6): 263–271. DOI:10.1016/S0375-9601(99)00389-8
[4] Kodama Y, Gibbons J. A method for solving the dispersionless KP equation and its exact solutions. Phys Lett A, 1988, 129: 223–226. DOI:10.1016/0375-9601(88)90354-4
[5] Krichever I M. The dispersionless Lax equations and topological minimal models. Commun Math Phys, 1992, 143(2): 415–429. DOI:10.1007/BF02099016
[6] Aoyama S, Kodama Y. Topological Landau-Ginzburg theory with a rational potential and the dispersionless KP equation. Commun Math Phys, 1996, 182(1): 185–219. DOI:10.1007/BF02506390
[7] Lax P D, Levermore C D. The zero dispersion limit for the KdV equation Ⅰ, Ⅱ, Ⅲ. Commun Pure Appl Math, 1983, 36: 253–290. DOI:10.1002/(ISSN)1097-0312
[8] Takasaki T, Takebe T. Integrable hierarchies and dispersionless limit. Rev Math Phys, 1995, 47: 743–808.
[9] Takasaki T, Takebe T. Quasi-classical limit of KP hierarchy, W-symmetries and free fermions. J Math Sci, 1999, 94(4): 1635–1641. DOI:10.1007/BF02365211
[10] Konopelchenko B, Martinez Alonso L, Ragnisco O. The∂-approach to the dispersionless KP hierarchy. J Phys A, 2001, 34(47): 10209–10217. DOI:10.1088/0305-4470/34/47/322
[11] Tasakaki K. Quasi-classical limit of BKP hierarchy and W-infinity symmetries. Letters in Mathematical Physics, 1993, 28: 177–185. DOI:10.1007/BF00745149
[12] Dasgupta N, Roy A. A new integrable equation in the 2+1-dimension-dispersionless limit of the BKP equation. J Phys A:Math Gen, 1992, 25: 1033–1038. DOI:10.1088/0305-4470/25/17/003
[13] Chen Y T, Tu M H. A note on the dispersionless BKP hierarchy. J Math A:Math Gen, 2006, 39: 7641–7655. DOI:10.1088/0305-4470/39/24/003
[14] Bogdanov L V, Konopelchenko B G. Symmetry constraints for dispersionless integrable equations and systems of hydrodynamic type. Physics Letters A, 2004, 6: 448–459.
[15] VandeLeur J W. Schlesinger-Bäcklund transformation for N-component KP. J Math Phys, 1998, 39: 2833–2847. DOI:10.1063/1.532423
[16] Kacand V G, VandeLeur J W. The N-component KP hierarchy and representation theory. J Math Phys, 2003, 44: 3245–3293. DOI:10.1063/1.1590055
[17] Mel'nikov V K. On equation for wave interaction. Lett Math Phys, 1983, 7: 129–136. DOI:10.1007/BF00419931
[18] Hu X B, Wang H Y. Construction of dKP and BKP equations with self-consistent sources. Inverse Probl, 2006, 22: 1903–1920. DOI:10.1088/0266-5611/22/5/022
[19] Xiao T, Zeng Y B. The quasiclassical limit of the symmetry constraint of the KP hierarchy and the dispersionless KP hierarchy with self-consistent soueces. J Nonlinear Math Phys, 2006, 13: 193–204. DOI:10.2991/jnmp.2006.13.2.4
[20] Liu X J, Zeng Y B, Lin R L. A new extended KP hierarchy. Physics Letters A, 2008, 372: 3819–3823. DOI:10.1016/j.physleta.2008.02.070
[21] Xiao T, Zeng Y B. The constrained dispersionless mKP hierarchy and the dispersionless mKP hierarchy with self-consistent sources. Phys Lett A, 2006, 349: 128–134. DOI:10.1016/j.physleta.2005.08.078
[22] Cheng J P, He J S. The integral type gauge transformation and the additional symmetry for the constrained KP hierarchy. Acta Mathematica Scientia, 2015, 35B(5): 1111–1121.
[23] Date E, Jimbo M, Kashiwara M, et al. Transformation groups for soliton equations//Nonlinear Integrable Systems-Classical Theory and Quantum Theory. Koyo Singapore:World Scientific, 1983: 39–119.
[24] Loris I, Willox R. Symmetry reductions of the BKP hierarchy. J Math Phys, 1999, 40: 1420–1431. DOI:10.1063/1.532812