1:
2: \documentclass[12pt]{amsart}
3: \usepackage{epsfig}
4: \textwidth=155mm
5: \textheight=205mm
6: \topmargin=-10mm
7: \parskip 8pt
8: \begin{document}
9: \topmargin=10mm
10: \oddsidemargin=0mm
11: \evensidemargin=0mm
12: \baselineskip=20pt
13: \parindent 20pt
14: \begin{center}
15: {\Large\bf Constructing periodic wave solutions of nonlinear
16: equations by Hirota bilinear method }\\[12pt]
17: Huihui Dai$^{a}$,\ \ Engui Fan$^{b}$, \ \ Xianguo Geng$^{c}$\\[8pt]
18: {\footnotesize a. Department of Mathematics, City University of Hong Kong, Hong
19: Kong SAR, China \\[6pt]
20: b. Institute of Mathematics, Fudan University, Shanghai 200433, P. R. China.
21: \\
22: c. Department of Mathematics, Zhengzhou University, Zhengzhou, P. R. China}\\[12pt]
23: \end{center}
24: {\bf Abstract.} Hirota bilinear method is proposed to directly
25: construct periodic wave solutions in terms of Riemann theta
26: functions for nonlinear equations. The asymptotic property of
27: periodic waves are analyzed in detail. It is shown that well-known
28: soliton solutions can be reduced from the periodic wave solutions.
29: \\[12pt]
30: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
31: {\bf 1. Introduction}
32:
33: The investigation of the exact solutions of
34: nonlinear equations plays an important role in the study of
35: nonlinear physical phenomena. For example, the wave phenomena
36: observed in fluid dynamics, plasma and elastic media are often
37: modelled by the bell shaped sech solutions and the kink shaped
38: tanh traveling wave solutions. The exact solution, if available,
39: of those nonlinear equations facilitates the verification of
40: numerical solvers and aids in the stability analysis of solutions.
41: In the past decades, there has been significant progression in
42: the development of these methods such as inverse scattering
43: method [1,2], Darboux transformation [3-6], Hirota bilinear
44: method [7-12], algebro-geometric method [13-18] and others. Among
45: them, the algebro-geometric method is an analogue of inverse
46: scattering transformation, which was first developed by Matveev,
47: Its, Novikov, and Dubrovin et al. The method can derive an
48: important class of exact solutions, which is called quasi-periodic
49: or algebro-geometric solution, to many soliton equations such as
50: KdV equation, sin-Gordon equation, and Schrodinger equation. In
51: recent years, such solutions of nonlinear equations have been
52: aroused much interest in the mathematical physics [19-24].
53: However, this method usually is applied in the integrable
54: nonlinear evolution equations admitting Lax pairs representation
55: and involves complicated algebraic geometry theory. These have
56: been used far less than their soliton counterparts. The main
57: reason for this is that they are far more complicated. Soliton
58: solutions are typically expressed in terms of rational or
59: hyperbolic functions, whereas qusi-periodic solutions require the
60: use of Riemann theta functions and calculus on Riemann surfaces.
61: Recently, Deconinck, Hoeij et al proposed an algorithm to compute
62: the Riemann theta function and Riemann constants [25,26].
63:
64: It is well known that the bilinear derivative method developed by Hirota is a
65: powerful and direct approach to construct exact solution of
66: nonlinear equations [7-9]. Once a nonlinear equation is written in
67: bilinear forms by a dependent variable transformation, then
68: multi-soliton solutions and rational solutions can be obtained.
69: In recent years, Hirota method also has been developed for
70: obtaining Wronskian and Pfaffian forms of N-soliton solution
71: [10,11]. In this paper, we propose Hirota method for directly
72: constructing periodic wave solutions in Riemann theta functions.
73: It is shown that the periodic solutions can be reduced to
74: classical soliton solutions under a certain limit. The appeal and
75: success of this method lies in the fact we circumvent complicated
76: algebro-geometric theory to directly get explicit periodic wave
77: solutions. Moreover, all parameters appearing in the solutions
78: are free variables, whereas usual quasi-periodic solutions
79: involve Riemann constants which are difficult to be determined
80: and need to make complicated Abel transformation on Riemann surface [13-18].
81: As illustrative examples, we then consider KdV equation and KP equation in this paper.
82: Other many equations also can be dealt with this way.\\[12pt]
83: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
84: {\bf 2. Periodic wave solutions of KdV equation and their
85: reduction}
86:
87: Consider KdV equation
88: $$u_t+6uu_x+u_{xxx}=0.\eqno(2.1)$$
89: Substituting transformation
90: $$u=u_0+2(\ln f)_{xx}\eqno(2.2)$$
91: into Eq.(2.1) and integrating once again, we then get the
92: following bilinear form
93: $$G(D_x,D_t)f\cdot f=(D_xD_t+6u_0D_x^2+D_x^4+c)f\cdot f=0\eqno(2.3)$$
94: where $c=c(t)$ is a integration constant and $u_0$ is a constant
95: solution of KdV equation.
96:
97: The Hirota bilinear differential operator is defined by
98: $$D_x^mD_t^n f(x,t)\cdot g(x,t)=(\partial_x-\partial_{x'})^m(\partial_t-\partial_{t'})^n
99: f(x,t) g(x,t)|_{x=x',t=t'}$$
100: The D-operator have good property when acting on exponential
101: functions
102: $$D_x^mD_t^n e^{\xi_1}\cdot
103: e^{\xi_2}=(k_1-k_2)^m(\omega_1-\omega_2)^n
104: e^{\xi_1+\xi_2}$$
105: where $\xi_j=k_jx+\omega_jt, j=1,2$. More general, we have
106: $$G(D_x,D_t)e^{\xi_1}\cdot
107: e^{\xi_2}=G(k_1-k_2,\omega_1-\omega_2)
108: e^{\xi_1+\xi_2}\eqno(2.4)$$
109:
110: We consider Riemann theta function solution of KdV equation
111: $$f=\sum_{n\in Z^N}e^{\pi i<\tau n,n>+2\pi i<\xi,n>}\eqno(2.5)$$
112: where $n=(n_1,\cdots,n_N), \xi =(\xi_1, \cdots, \xi_N)$, $\tau$ is
113: a symmetric matrix and ${\rm Im}\tau>0$, $\xi_j=k_jx+\omega_j t,
114: j=1,\cdots, N.$
115:
116: Consider the case when $N=1$, then (2.5) becomes
117: $$f=\sum_{n=-\infty}^{\infty}e^{2\pi in\xi+\pi in^2\tau}\eqno(2.6)$$
118:
119: Substituting (2.6) into (2.3) gives
120: \begin{eqnarray*}
121: &&Gf\cdot f=G(D_x,D_t)\sum_{n=-\infty}^{\infty}e^{2\pi in\xi+\pi
122: in^2\tau}\cdot\sum_{m=-\infty}^{\infty}
123: e^{2\pi i m\xi+\pi i m^2\tau}\\
124: &&=\sum_{n=-\infty}^{\infty}\sum_{m=-\infty}^{\infty}G(D_x,D_t)e^{2\pi
125: in\xi+\pi in^2\tau}\cdot e^{2\pi i m\xi+\pi i
126: m^2\tau}\\
127: &&=\sum_{n=-\infty}^{\infty}\sum_{m=-\infty}^{\infty}G(2\pi
128: i(n-m)k,2\pi i(n-m)\omega)e^{2\pi i(n+m)\xi+\pi i ( n^2+m^2)\tau}\\
129: &&\stackrel{n+m=m'}{=}\sum_{m'=-\infty}^{\infty}\left\{
130: \sum_{n=-\infty}^{\infty}G(2\pi i(2n-m')k,2\pi
131: i(2n-m')\omega)e^{\pi
132: i[(n^2+(n-m')^2]\tau}\right\}e^{2\pi i m'\xi}\\
133: &&\equiv\sum_{m'=-\infty}^{\infty}\bar{G}(m')e^{2\pi i m'\xi}=0.
134: \end{eqnarray*}
135:
136: Noticing that
137: \begin{eqnarray*}
138: &&\bar{G}(m')=\sum_{n=-\infty}^{\infty}G(2\pi i(2n-m')k,2\pi
139: i(2n-m')\omega)e^{\pi i[(n^2+(n-m')^2]\tau}\\
140: &&\stackrel{n=n'+1}{=}\sum_{n'=-\infty}^{\infty}G(2\pi
141: i[2n'-(m'-2)]k,2\pi i[2n'-(m'-2)]\omega)\\
142: &&\ \ \ \ \ \times \exp(\pi i[(n'^2+(n'-(m'-2))^2]\tau)
143: \exp(2\pi i(m'-1)\tau)\\
144: &&=\bar{G}(m'-2)e^{2\pi i (m'-1)\tau}\\
145: &&=\cdots=\left\{\begin{matrix}\bar{G}(0)e^{\pi im'(m'-1)},&m' \ \
146: {\rm is\ \ even}\cr\cr \bar{G}(1)e^{\pi i(m'+1)(m'-2)},&m' \ \
147: {\rm is\ \ odd}\end{matrix}\right.
148: \end{eqnarray*}
149: which implies that if $\bar{G}(0)=\bar{G}(1)=0$, then
150: $$\bar{G}(m')=0, \ \ m'\in Z$$
151:
152: In this way, we may let
153: $$\bar{G}(0)=\sum_{n=-\infty}^{\infty}(-16\pi^2n^2k\omega-96u_0\pi^2k^2n^2+256\pi^4n^4k^4+c)e^{2\pi
154: in^2\tau}=0,\eqno(2.7)$$
155: $$\bar{G}(1)=\sum_{n=-\infty}^{\infty}(-4\pi^2(2n-1)^2k\omega-24u_0\pi^2(2n-1)^2k^2
156: +16\pi^4(2n-1)^4k^4+c)e^{\pi i(2n^2-2n+1)\tau}=0,\eqno(2.8)$$
157:
158: Denote
159: $$\delta_1(n)=e^{2\pi in^2\tau},\ \ \delta_2(n)=e^{\pi i(2n^2-2n+1)\tau},$$
160: $$a_{11}=-\sum_{n=-\infty}^{\infty}16\pi^2n^2k\delta_1(n),\ \
161: a_{12}=\sum_{n=-\infty}^{\infty}\delta_1(n),$$
162: $$b_1=\sum_{n=-\infty}^{\infty}(256\pi^4n^4k^4-96u_0\pi^2k^2n^2)\delta_1(n),\ \
163: a_{21}=-\sum_{n=-\infty}^{\infty}4\pi^2(2n-1)^2k\delta_2(n),$$
164: $$a_{22}=\sum_{n=-\infty}^{\infty}\delta_2(n),\ \
165: b_2=\sum_{n=-\infty}^{\infty}[16\pi^4(2n-1)^4k^4-24u_0\pi^2(2n-1)^2k^2]\delta_2(n),$$
166: then (2.7) and (2.8) can be written as
167: $$a_{11}\omega+a_{12}c+b_1=0,$$
168: $$a_{21}\omega+a_{22}c+b_2=0.$$
169: Solving this system, we get
170: $$\omega=\frac{b_2a_{12}-b_1a_{22}}{a_{11}a_{22}-a_{12}a_{21}},\ \ \ \
171: c=\frac{b_1a_{21}-b_2a_{11}}{a_{11}a_{22}-a_{12}a_{21}}.\eqno(2.9)$$
172:
173: Finally we get periodic wave solution
174: $$u=u_0+2(\ln f)_{xx},\eqno(2.10)$$
175: where $f$ and $\omega$ are given by (2.6) and (2.9)
176: respectively. From fig.1-fig.3, we see that the parameter $\tau$
177: does affect the period and shape of wave, but parameter $k$ has
178: effect on the period and shape of wave.
179:
180: \begin{figure}[h]
181: $ \ \ \ \ $ {\footnotesize $(a) \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
182: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $}\
183: \ \ \ \ \ \ \ \ \
184: \ \ {\footnotesize $(b)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
185: \ \ \ \ $
186: \\[6pt]
187: \begin{minipage}[u]{0.48\linewidth}
188: \centering \epsfig{figure=kdvfig11.eps,width=2.8 in }
189: \end{minipage}
190: \begin{minipage}[v]{0.48\linewidth}
191: \centering \epsfig{file=kdvfig12.eps,width=2.8 in}
192: \end{minipage}
193: \begin{center}
194: \vspace{6mm}
195: {\footnotesize Figure 1. One-periodic wave
196: for KdV equation and the effect of parameters on wave shape:\\ (a)
197: along x-axis, (b) along t-axis, where $k=0.1, \tau=i$.}
198: \end{center}
199: \end{figure}
200:
201: \begin{figure}[h]
202: $ \ \ \ \ $ {\footnotesize $(a) \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
203: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $}\
204: \ \ \ \ \ \ \ \ \
205: \ \ {\footnotesize $(b)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
206: \ \ \ \ $
207: \\[6pt]
208: \begin{minipage}[u]{0.48\linewidth}
209: \centering \epsfig{figure=kdvfig13.eps,width=2.8 in }
210: \end{minipage}
211: \begin{minipage}[v]{0.48\linewidth}
212: \centering \epsfig{file=kdvfig14.eps,width=2.8 in}
213: \end{minipage}
214: \begin{center}
215: \vspace{6mm}
216: {\footnotesize Figure 2. One-periodic wave
217: for KdV equation and the effect of parameters on wave shape:\\ (a)
218: along x-axis, (b) along t-axis, where $k=0.1, \tau=3i$.}
219: \end{center}
220: \end{figure}
221:
222: \begin{figure}[h]
223: $ \ \ \ \ $ {\footnotesize $(a) \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
224: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $}\
225: \ \ \ \ \ \ \ \ \
226: \ \ {\footnotesize $(b)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
227: \ \ \ \ $
228: \\[6pt]
229: \begin{minipage}[u]{0.48\linewidth}
230: \centering \epsfig{figure=kdvfig15.eps,width=2.8 in }
231: \end{minipage}
232: \begin{minipage}[v]{0.48\linewidth}
233: \centering \epsfig{file=kdvfig16.eps,width=2.8 in}
234: \end{minipage}
235: \begin{center}
236: \vspace{6mm}
237: {\footnotesize Figure 3. One-periodic wave
238: for KdV equation and the effect of parameters on wave shape:\\ (a)
239: along x-axis, (b) along t-axis, where $k=0.15, \tau=i$.}
240: \end{center}
241: \end{figure}
242:
243: The well-known soliton solution of KdV equation can
244: be obtained as limit of the periodic solution (2.10). For this
245: purpose, we write $f$ as
246: $$ f=1+\alpha(e^{2\pi i\xi}+e^{-2\pi i\xi})+\alpha^4(e^{4\pi i\xi}+e^{-4\pi i\xi})
247: +\cdots,$$ where $\alpha=e^{i\pi \tau}$.
248:
249: Setting $u_0=0,\ \ \xi=\xi'-\tau/2,\ \ k'=2\pi i k,\ \ \omega'=2\pi i
250: \omega$,
251: we get
252: \begin{eqnarray*}
253: &&f=1+e^{k'x+\omega' t}+\alpha^2e^{-2\pi
254: i\xi'}+\alpha^2e^{4\pi i\xi'}+\alpha^6e^{-4\pi i\xi'}+\cdots\\
255: &&\longrightarrow 1+e^{k'x+\omega' t},\ \ {\rm as}\ \
256: \alpha\longrightarrow 0
257: \end{eqnarray*}
258:
259: So the periodic solution (2.10) convergence to well-known soliton
260: solution
261: $$u=2(\ln f)_{xx}, \ \ f=1+e^{k'x+\omega' t},$$
262: We only need to prove that
263: $$\omega'\longrightarrow -k'^3,\ \ {\rm as }\ \ \alpha\longrightarrow
264: 0.\eqno(2.11)$$
265:
266: In fact, it is easy to see that
267: \begin{eqnarray*}
268: &&a_{11}=-32k\pi^2(\alpha^2+4\alpha^8+\cdots),\\
269: &&a_{12}=1+2\alpha^2+\cdots,\\
270: &&a_{22}=2\alpha+\alpha^5+\cdots,\\
271: &&a_{21}=-8k\pi^2(\alpha+9\alpha^5+\cdots),\\
272: &&b_1=256k^4\pi^4(\alpha^2+16\alpha^8+\cdots),\\
273: &&b_2=16k^4\pi^4(2\alpha+81\alpha^5),
274: \end{eqnarray*}
275: which lead to
276: $$b_2a_{12}-b_1a_{22}=32k^4\pi^4\alpha+o(\alpha),\ \ \
277: a_{11}a_{22}-a_{12}a_{21}=8k\pi^2\alpha+o(\alpha)$$
278: Therefore we have
279: $$\omega\longrightarrow 4k^3\pi^2,\ \ \ {\rm as }\ \ \alpha\longrightarrow
280: 0,$$ which implies (2.11).
281:
282:
283: Now we consider two-periodic wave solution of KdV equation
284: ($N=2$). Substituting (2.5) into (2.3), we have
285: \begin{eqnarray*}
286: &&Gf\cdot f=\sum_{m,n\in Z^2} G(D_x,D_t)e^{2\pi i<\xi,n>+\pi
287: i<\tau n,n>}\cdot e^{2\pi i<\xi,m>+\pi
288: i<\tau m,m>}\\
289: &&=\sum_{m,n\in Z^2}G(2\pi
290: i<n-m,k>,2\pi i<n-m,\omega>)e^{2\pi i<\xi,n+m>+\pi i (<\tau m,m>+<\tau n, n>)}\\
291: &&\stackrel{n+m=m'}{=}\sum_{m'\in
292: Z^2}\sum_{n_1,n_2=-\infty}^{\infty}G(2\pi i<2n-m',k>,2\pi
293: i<2n-m',\omega>)\\
294: &&\ \ \ \ \ \ \ \ \ \ \ \ \ \times \exp(\pi i(<\tau (n-m'),n-m'>+<\tau n, n>))
295: \exp(2\pi i<\xi,m'>)\\
296: &&\equiv\sum_{m'\in Z^2}\bar{G}(m_1', m_2') e^{2\pi i <\xi,m'>}=0.
297: \end{eqnarray*}
298:
299: It is easy to calculate that
300: \begin{eqnarray*}
301: &&\bar{G}(m_1', m_2')\\
302: &&=\sum_{n_1,n_2=-\infty}^{\infty}G(2\pi i<2n-m',k>,2\pi
303: i<2n-m',\omega>)e^{\pi i(<\tau (n-m'),n-m'>+<\tau n, n>)}\\
304: &&\stackrel{n_j=n_j'+\delta_{jl},
305: l=1,2}{=}\sum_{n_1,n_2=-\infty}^{\infty}G\left(2\pi
306: i\sum_{j=1}^{2}(2n_j'-(m_j'-2\delta_{jl}))k_j,2\pi
307: i\sum_{j=1}^{2}(2n_j'-(m_j'-2\delta_{jl}))\omega_j\right)\\
308: &&\times \exp\left\{\pi
309: i\sum_{j,k=1}^{2}[(n_j'+\delta_{jl})\tau_{jk}(n_k'+\delta_{kl})+((m_j'-2\delta_{jl}-n_j')+\delta_{jl})
310: \tau_{jk}(m_k'-2\delta_{kl}-n_k')+\delta_{kl})]\right\}, \\
311: &&=\left\{\begin{matrix}\bar{G}(m_1'-2,m_2')e^{2\pi
312: i(m_1'-1)\tau_{11}+2\pi im_2'\tau_{12}},\ \ \ l=1\cr\cr
313: \bar{G}(m_1',m_2'-2) e^{2\pi i(m_2'-1)\tau_{22}+2\pi
314: im_1'\tau_{12}},\ \ \ l=2
315: \end{matrix}\right.
316: \end{eqnarray*}
317: which implies that if
318: $$\bar{G}(0,0)=\bar{G}(0,1)=\bar{G}(1,0)=\bar{G}(1,1)=0,$$
319: then $\bar{G}(m_1',m_2')=0$ and (2.2) is an exact solution
320: of KdV equation.
321:
322: Denote
323: \begin{eqnarray*}
324: &&a_{j1}=-\sum_{n_1,n_2=-\infty}^{\infty}
325: 4\pi^2<2n-m^j,k>(2n_1-m_1^j)\delta_j(n),\\
326: &&a_{j2}=-\sum_{n_1,n_2=-\infty}^{\infty}4\pi^2
327: <2n-m^j,k>(2n_2-m_2^j)\delta_j(n)\\
328: &&a_{j3}=-\sum_{n_1,n_2=-\infty}^{\infty}24\pi^2
329: <2n-m^j,k>^2\delta_j(n),\\
330: &&a_{j4}=\sum_{n_1,n_2=-\infty}^{\infty}\delta_j(n)\\
331: && b_j=-\sum_{n_1,n_2=-\infty}^{\infty}16\pi^4
332: <2n-m^j,k>^4\delta_j(n),\\
333: &&\delta_j(n)=e^{\pi i<\tau(n-m^j),n-m^j>+\pi i<\tau n,n>},\\
334: &&j=1, 2, 3, 4\ \ m^1=(0,0),\ \ m^2=(1,0),\ \ m^3=(0,1),\ \ m^4=(1,1)\\
335: &&A=(a_{kj})_{4\times 4},\ \ b=(b_1, b_2, b_3, b_4)^T,
336: \end{eqnarray*}
337: then we have $$A\left(\begin{matrix}\omega_1\cr\omega_2\cr u_0\cr
338: c\end{matrix}\right)=b,$$
339: from which we obtain
340: $$\omega_1=\frac{\Delta_1}{\Delta},\ \ \omega_2=\frac{\Delta_2}{\Delta},\ \
341: u_0=\frac{\Delta_3}{\Delta},\eqno(2.12)$$
342: where $\Delta=|A|$ and $\Delta_1, \Delta_2, \Delta_3$ are produced
343: from $\Delta$ by replacing its $1th, 2th, 3th$ column with $b$
344: respectively. Finally we get two-periodic wave solution (See
345: fig.2)
346: $$u=u_0+2(\ln f)_{xx},\eqno(2.13)$$
347: where $f$ and $\omega_1, \omega_2$ are given by (2.5) and (2.12)
348: respectively. The properties of the solution are shown in
349: fig.4-fig.7. Fig.4 and fig.5 shown effect of parameters $k_1, k_2$
350: on period and shape of wave. Fig.7 gives effect of parameter
351: $\tau_{11}, \tau_{12}, \tau_{22}$ on period and shape of wave. The
352: fig.7 and fig.8 show that two-periodic wave can degenerate to
353: one-periodic wave when $k_1$ is sufficient small.
354:
355: \begin{figure}[h]
356: $ \ \ \ \ $ {\footnotesize $(a) \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
357: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $}\
358: \ \ \ \ \ \ \ \ \
359: \ \ {\footnotesize $(b)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
360: \ \ \ \ $
361: \\[6pt]
362: \begin{minipage}[u]{0.48\linewidth}
363: \centering \epsfig{figure=kdvfig21.eps,width=2.8 in }
364: \end{minipage}
365: \begin{minipage}[v]{0.48\linewidth}
366: \centering \epsfig{file=kdvfig22.eps,width=2.8 in}
367: \end{minipage}
368: \begin{center}
369: \vspace{6mm}
370: {\footnotesize Figure 4. Two-periodic wave
371: for KdV equation and the effect of parameters on wave shape:\\
372: (a) along x-axis, (b) along t-axis, where $\tau_{11}=0.1i,
373: \tau_{12}=0.2i, \tau_{22}=3i, k_1=0.2, k_2=-0.3$.}
374: \end{center}
375: \end{figure}
376:
377:
378:
379: \begin{figure}[h]
380: $ \ \ \ \ $ {\footnotesize $(a) \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
381: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $}\
382: \ \ \ \ \ \ \ \ \
383: \ \ {\footnotesize $(b)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
384: \ \ \ \ $
385: \\[6pt]
386: \begin{minipage}[u]{0.48\linewidth}
387: \centering \epsfig{figure=kdvfig23.eps,width=2.8 in }
388: \end{minipage}
389: \begin{minipage}[v]{0.48\linewidth}
390: \centering \epsfig{file=kdvfig24.eps,width=2.8 in}
391: \end{minipage}
392: \begin{center}
393: \vspace{6mm}
394: {\footnotesize Figure 5. Two-periodic wave
395: for KdV equation and the effect of parameters on wave shape:\\ (a)
396: along x-axis, (b) along t-axis, where $\tau_{11}=0.1i,
397: \tau_{12}=0.2i, \tau_{22}=3i, k_1=0.1, k_2=-0.3$.}
398: \end{center}
399: \end{figure}
400:
401:
402: \begin{figure}[h]
403: $ \ \ \ \ $ {\footnotesize $(a) \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
404: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $}\
405: \ \ \ \ \ \ \ \ \
406: \ \ {\footnotesize $(b)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
407: \ \ \ \ $
408: \\[6pt]
409: \begin{minipage}[u]{0.48\linewidth}
410: \centering \epsfig{figure=kdvfig27.eps,width=2.8 in }
411: \end{minipage}
412: \begin{minipage}[v]{0.48\linewidth}
413: \centering \epsfig{file=kdvfig28.eps,width=2.8 in}
414: \end{minipage}
415: \begin{center}
416: \vspace{6mm}
417: {\footnotesize Figure 6. Two-periodic wave
418: for KdV equation and the effect of parameters on wave shape:\\ (a)
419: along x-axis, (b) along t-axis, where $\tau_{11}=0.1,
420: \tau_{12}=0.3i, \tau_{22}=2i, k_1=0.1, k_2=-0.3$.}
421: \end{center}
422: \end{figure}
423:
424:
425: \begin{figure}[h]
426: $ \ \ \ \ $ {\footnotesize $(a) \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
427: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $}\
428: \ \ \ \ \ \ \ \ \
429: \ \ {\footnotesize $(b)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
430: \ \ \ \ $
431: \\[6pt]
432: \begin{minipage}[u]{0.48\linewidth}
433: \centering \epsfig{figure=kdvfig25.eps,width=2.8 in }
434: \end{minipage}
435: \begin{minipage}[v]{0.48\linewidth}
436: \centering \epsfig{file=kdvfig26.eps,width=2.8 in}
437: \end{minipage}
438: \begin{center}
439: \vspace{6mm}
440: {\footnotesize Figure 7. Two-periodic wave
441: for KdV equation and the effect of parameters on wave shape:\\ (a)
442: along x-axis, (b) along t-axis, where $\tau_{11}=0.01,
443: \tau_{12}=0.2i, \tau_{22}=3i, k_1=0.1, k_2=-0.2$.}
444: \end{center}
445: \end{figure}
446:
447: \begin{figure}[h]
448: $ \ \ \ \ $ {\footnotesize $(a) \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
449: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $}\
450: \ \ \ \ \ \ \ \ \
451: \ \ {\footnotesize $(b)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
452: \ \ \ \ $
453: \\[6pt]
454: \begin{minipage}[u]{0.48\linewidth}
455: \centering \epsfig{figure=kdvfig29.eps,width=2.8 in }
456: \end{minipage}
457: \begin{minipage}[v]{0.48\linewidth}
458: \centering \epsfig{file=kdvfig210.eps,width=2.8 in}
459: \end{minipage}
460: \begin{center}
461: \vspace{6mm}
462: {\footnotesize Figure 8. Two-periodic wave
463: for KdV equation and the effect of parameters on wave shape:\\ (a)
464: along x-axis, (b) along t-axis, where $\tau_{11}=0.001,
465: \tau_{12}=0.2i, \tau_{22}=3i, k_1=0.1, k_2=-0.2$.}
466: \end{center}
467: \end{figure}
468:
469: The two-soliton solution of KdV equation can be obtained as limit
470: of the periodic solution (2.10). We write $f$ as
471: \begin{eqnarray*}
472: && f=1+(e^{2\pi i\xi_1}+e^{-2\pi i\xi_1})e^{\pi i\tau_{11}}
473: +(e^{2\pi i\xi_2}+e^{-2\pi i\xi_2})e^{\pi
474: i\tau_{22}}\\
475: &&+(e^{2\pi i(\xi_1+\xi_2)}+e^{-2\pi i(\xi_1+\xi_2)})e^{\pi
476: i(\tau_{11}+2\tau_{12}+\tau_{22})}+\cdots
477: \end{eqnarray*}
478:
479: Setting $ \xi_1'=2\pi i\xi_1-\pi i\tau_{11},\ \ \xi_2'=2\pi i\xi_2-\pi i\tau_{22},
480: \ \ \tau_{12}=i\widetilde{\tau}$ ($\widetilde{\tau}$ is a real),
481: we get
482: \begin{eqnarray*}
483: &&f=1+e^{\xi_1'}+e^{\xi_2'}+e^{\xi_1'+\xi_2'+2\pi i\tau_{12}
484: }+\alpha_1^2e^{-\xi_1'}
485: +\alpha_2^2e^{-\xi_2'}+\alpha_1^2\alpha_2^2e^{-\xi_1-\xi_2+2\pi i\tau_{11}}+\cdots\\
486: &&\longrightarrow 1+e^{\xi_1'}+e^{\xi_2'}+e^{\xi_1'+\xi_2'-2\pi
487: \widetilde{\tau}},\ \ {\rm as}\ \ \alpha_1, \alpha_2
488: \longrightarrow 0
489: \end{eqnarray*}
490: where
491: $$\alpha_1=e^{\pi i\tau_{11}},\ \ \alpha_2=e^{\pi
492: i\tau_{22}}, \ \ \xi_i'=k_i'x+\omega_i't,\ \ \ e^{-2\pi
493: \widetilde{\tau}}=\left(\frac{k_1'-k_2'}{\omega_1'-\omega_2'}\right)^2,$$
494: $$\omega_i'\rightarrow -k_i'^3,\ \ {\rm as}\ \ \alpha_1,
495: \alpha_2\rightarrow0.$$
496: \\[24pt]
497: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
498: {\bf 3. Periodic wave solutions of KP equation and their
499: reduction}
500:
501: Consider KP equation
502: $$u_t=(u_{xx}+3u^2)_x+3\partial^{-1}u_{yy}.\eqno(3.1)$$
503: Substituting transformation
504: $$u=u_0+2(\ln f)_{xx}\eqno(3.2)$$
505: into (3.1) and integrating once again, we then get the following
506: bilinear form
507: $$G(D_x,D_y,D_t)f\cdot f=(D_xD_t-6u_0D_x^2-D_x^4-3D_y^2+c_0)f\cdot f=0\eqno(3.3)$$
508: where $c_0$ is a constant of integration and $u_0$ is a constant
509: solution of KP equation. Let us consider Riemann theta function
510: solution of KP equation
511: $$f=\sum_{n\in Z^N}e^{\pi i<\tau n,n>+2\pi i<\xi,n>}\eqno(3.4)$$
512: where $n=(n_1,\cdots,n_N), \xi =(\xi_1, \cdots, \xi_N)$, $\tau$ is
513: a symmetric matrix and ${\rm Im}\tau>0$,
514: $\xi_j=k_j(x+\rho_jy+\omega_j t)+\gamma_j, j=1,\cdots, N.$
515:
516: In the following, we consider some special cases. When $N=1$,
517: (3.4) is reduced to
518: $$f=\sum_{n=-\infty}^{\infty}e^{2\pi in\xi+\pi in^2\tau}\eqno(3.5)$$
519: Substituting (3.5) into (3.3) gives
520: {\footnotesize\begin{eqnarray*} &&Gf\cdot
521: f=G(D_x,D_y,D_t)\sum_{n=-\infty}^{\infty}e^{2\pi in\xi+\pi
522: in^2\tau}\cdot\sum_{m=-\infty}^{\infty}
523: e^{2\pi i m\xi+\pi i m^2\tau}\\
524: &&=\sum_{n=-\infty}^{\infty}\sum_{m=-\infty}^{\infty}G(D_x,D_y,D_t)e^{2\pi
525: in\xi+\pi in^2\tau}\cdot e^{2\pi i m\xi+\pi i
526: m^2\tau}\\
527: &&=\sum_{n=-\infty}^{\infty}\sum_{m=-\infty}^{\infty}G(2\pi
528: i(n-m)k,2\pi i(n-m)k\rho,2\pi i(n-m)k\omega)e^{2\pi i(n+m)\xi+\pi i ( n^2+m^2)\tau}\\
529: &&\stackrel{n+m=m'}{=}\sum_{m'=-\infty}^{\infty}\left\{
530: \sum_{n=-\infty}^{\infty}G(2\pi i(2n-m')k,2\pi i(2n-m')k\rho,2\pi
531: i(2n-m')k\omega)e^{\pi
532: i[(n^2+(n-m')^2]\tau}\right\}e^{2\pi i m'\xi}\\
533: &&\equiv\sum_{m'=-\infty}^{\infty}\bar{G}(m')e^{2\pi i m'\xi}=0.
534: \end{eqnarray*}}
535: Noticing that
536: \begin{eqnarray*}
537: &&\bar{G}(m')=\sum_{n=-\infty}^{\infty}G(2\pi i(2n-m')k,2\pi
538: i(2n-m')k\rho,2\pi
539: i(2n-m')k\omega)e^{\pi i[(n^2+(n-m')^2]\tau}\\
540: &&\stackrel{n=n'+1}{=}\sum_{n'=-\infty}^{\infty}G(2\pi
541: i[2n'-(m'-2)]k,2\pi i[2n'-(m'-2)]k\rho,2\pi i[2n'-(m'-2)]k\omega)\\
542: &&\ \ \ \ \ \times \exp(\pi i[(n'^2+(n'-(m'-2))^2]\tau)
543: \exp(2\pi i(m'-1)\tau)\\
544: &&=\bar{G}(m'-2)e^{2\pi i (m'-1)\tau}\\
545: &&=\cdots=\left\{\begin{matrix}\bar{G}(0)e^{\pi im'(m'-1)},&m' \ \
546: {\rm is\ \ even}\cr\cr \bar{G}(1)e^{\pi i(m'+1)(m'-2)},&m' \ \
547: {\rm is\ \ odd}\end{matrix}\right.
548: \end{eqnarray*}
549: which implies that if $\bar{G}(0)=\bar{G}(1)=0$, then
550: $$\bar{G}(m')=0, \ \ m'\in Z$$
551: In this way, we may let
552: $$\bar{G}(0)=\sum_{n=-\infty}^{\infty}(-16\pi^2n^2k^2\omega+96u_0\pi^2k^2n^2-256\pi^4n^4k^4
553: +48\pi^2k^2\rho^2n^2+c_0)e^{2\pi in^2\tau}=0,\eqno(3.6)$$
554: $$\bar{G}(1)=\sum_{n=-\infty}^{\infty}(-4\pi^2(2n-1)^2k^2\omega+24u_0\pi^2(2n-1)^2k^2
555: -16\pi^4(2n-1)^4k^4$$
556: $$+12\pi^2(2n-1)^2k^2\rho^2+c_0)e^{\pi
557: i(2n^2-2n+1)\tau}=0,\eqno(3.7)$$ Denote
558: $$\alpha=e^{\pi\tau i},\quad a_{11}=-16\pi^2k^2\sum_{n=-\infty}^{\infty}n^2\alpha^{2n^2},\ \
559: a_{12}=\sum_{n=-\infty}^{\infty}\alpha^{2n^2},$$ {\footnotesize
560: $$b_1=-\sum_{n=-\infty}^{\infty}(96u_0\pi^2n^2k^2-256\pi^4n^4k^4+48\pi^2k^2\rho^2n^2)\alpha^{2n^2},\
561: \
562: a_{21}=-4\pi^2k^2\sum_{n=-\infty}^{\infty}(2n-1)^2\alpha^{2n^2-2n+1},$$
563: $$a_{22}=\sum_{n=-\infty}^{\infty}\alpha^{2n^2-2n+1},\ \
564: b_2=\sum_{n=-\infty}^{\infty}[16\pi^4(2n-1)^4k^4-24u_0\pi^2(2n-1)^2k^2-12\pi^2(2n-1)^2k^2\rho^2
565: ]\alpha^{2n^2-2n+1},$$}
566: then (3.6) and (3.7) can be written as
567: $$a_{11}\omega+a_{12}c_0=b_1,$$
568: $$a_{21}\omega+a_{22}c_0=b_2.$$
569: Solving this system, we get
570: $$\omega=\frac{b_1a_{22}-b_2a_{12}}{a_{11}a_{22}-a_{12}a_{21}},\ \ \ \
571: c_0=\frac{b_2a_{11}-b_1a_{21}}{a_{11}a_{22}-a_{12}a_{21}}.\eqno(3.8)$$
572: Finally we get periodic wave solution of KP equation (3.1)
573: $$u=u_0+2(\ln f)_{xx},\eqno(3.9)$$
574: where $f$ and $\omega$ are given by (3.4) and (3.8)
575: respectively. From fig.9-fig.11, we see that the parameter $\tau$
576: does affect the period and shape of wave, but parameter $k$ has
577: effect on the period and shape of wave.
578:
579:
580: \begin{figure}[h]
581: $ \ \ \ \ $ {\footnotesize $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
582: \ $}\
583: {\footnotesize $(a)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
584: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
585: \ \ \ \ $ \\[6pt]
586: \begin{minipage}[u]{0.48\linewidth}
587: \centering \epsfig{figure=kpfig11.eps,width=2.8 in }
588: \end{minipage}\\[12pt]
589: $ \ \ \ \ $ {\footnotesize $(b) \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
590: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $}\
591: \ \ \ \ \ \ \ \ \
592: \ \ {\footnotesize $(c)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
593: \ \ \ \ $ \\[6pt]
594: \begin{minipage}[u]{0.48\linewidth}
595: \centering \epsfig{figure=kpfig12.eps,width=2.8 in }
596: \end{minipage}
597: \begin{minipage}[v]{0.48\linewidth}
598: \centering \epsfig{file=kpfig13.eps,width=2.8 in}
599: \end{minipage}
600: \begin{center}
601: \vspace{6mm}
602: {\footnotesize Figure 9. One-periodic wave
603: for KP equation and the effect of parameters on wave shape:\\
604: (a) along x-axis, (b) along y-axis, (b) along t-axis, where
605: $k=0.1, \tau=i, \rho=2$.}
606: \end{center}
607: \end{figure}
608:
609:
610: \begin{figure}[h]
611: $ \ \ \ \ $ {\footnotesize $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
612: \ $}\
613: {\footnotesize $(a)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
614: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
615: \ \ \ \ $ \\[6pt]
616: \begin{minipage}[u]{0.48\linewidth}
617: \centering \epsfig{figure=kpfig14.eps,width=2.8 in }
618: \end{minipage}\\[12pt]
619: $ \ \ \ \ $ {\footnotesize $(b) \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
620: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $}\
621: \ \ \ \ \ \ \ \ \
622: \ \ {\footnotesize $(c)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
623: \ \ \ \ $ \\[6pt]
624: \begin{minipage}[u]{0.48\linewidth}
625: \centering \epsfig{figure=kpfig15.eps,width=2.8 in }
626: \end{minipage}
627: \begin{minipage}[v]{0.48\linewidth}
628: \centering \epsfig{file=kpfig16.eps,width=2.8 in}
629: \end{minipage}
630: \begin{center}
631: \vspace{6mm}
632: {\footnotesize Figure 10. One-periodic wave
633: for KP equation and the effect of parameters on wave shape:\\
634: (a) along x-axis, (b) along y-axis, (b) along t-axis, where
635: $k=0.1, \tau=3i, \rho=1$.}
636: \end{center}
637: \end{figure}
638:
639:
640: \begin{figure}[h]
641: $ \ \ \ \ $ {\footnotesize $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
642: \ $}\
643: {\footnotesize $(a)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
644: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
645: \ \ \ \ $ \\[6pt]
646: \begin{minipage}[u]{0.48\linewidth}
647: \centering \epsfig{figure=kpfig17.eps,width=2.8 in }
648: \end{minipage}\\[12pt]
649: $ \ \ \ \ $ {\footnotesize $(b) \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
650: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $}\
651: \ \ \ \ \ \ \ \ \
652: \ \ {\footnotesize $(c)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
653: \ \ \ \ $ \\[6pt]
654: \begin{minipage}[u]{0.48\linewidth}
655: \centering \epsfig{figure=kpfig18.eps,width=2.8 in }
656: \end{minipage}
657: \begin{minipage}[v]{0.48\linewidth}
658: \centering \epsfig{file=kpfig19.eps,width=2.8 in}
659: \end{minipage}
660: \begin{center}
661: \vspace{6mm}
662: {\footnotesize Figure 11. One-periodic wave
663: for KP equation and the effect of parameters on wave shape:\\
664: (a) along x-axis, (b) along y-axis, (c) along t-axis, where
665: $k=0.2, \tau=i, \rho=1$.}
666: \end{center}
667: \end{figure}
668:
669: The well-known soliton solution of the KP equation can be obtained
670: as limit of the periodic solution (3.9). To this end, we write $f$
671: as
672: $$ f=1+\alpha(e^{2\pi i\xi}+e^{-2\pi i\xi})+\alpha^4(e^{4\pi i\xi}+e^{-4\pi i\xi})
673: +\cdots .$$ Setting $u_0=0,\ \gamma=\gamma_0-\tau/2,\ k'=2\pi i
674: k,\ \gamma'_0=2\pi i \gamma_0$,
675: we get
676: \begin{eqnarray*}
677: &&f=1+e^{k'(x+\rho y+\omega t)+\gamma'_0}+\alpha^2e^{-k'(x+\rho
678: y+\omega t)-\gamma'_0}+\alpha^2e^{2k'(x+\rho y+\omega
679: t)+2\gamma'_0}\\
680: &&\quad\quad\quad
681: +\alpha^6e^{-2k'(x+\rho y+\omega
682: t)-2\gamma'_0}+\cdots\\
683: &&\quad \longrightarrow 1+e^{k'(x+\rho y+\omega t)+\gamma'_0},\ \
684: {\rm as}\ \ \alpha\longrightarrow 0
685: \end{eqnarray*}
686: Then the periodic solution (3.9) convergence to well-known soliton
687: solution
688: $$u=2(\ln f)_{xx}, \ \ f=1+e^{k'(x+\rho y+\omega t)+\gamma'_0}.\eqno(3.10)$$
689: We only need to prove that
690: $$\omega\longrightarrow -4k^3\pi^2+3k\rho^2,\ \ \ {\rm as }\ \ \alpha\longrightarrow
691: 0.\eqno(3.11)$$
692:
693: In fact, it is easy to see that
694: \begin{eqnarray*}
695: &&a_{11}=-32k\pi^2(\alpha^2+4\alpha^8+\cdots),\\
696: &&a_{12}=1+2\alpha^2+\cdots,\\
697: &&a_{21}=-8\pi^2k(\alpha+9\alpha^5+\cdots),\\
698: &&a_{22}=2(\alpha+\alpha^5+\cdots),\\
699: &&b_1=512k^4\pi^4(\alpha^2+16\alpha^8+\cdots)-96k^2\pi^2\rho^2(\alpha^2+4\alpha^8+\cdots),\\
700: &&b_2=32k^4\pi^4(\alpha+81\alpha^5+\cdots)-24k^2\pi^2\rho^2(\alpha+9\alpha^5+\cdots).
701: \end{eqnarray*}
702: Therefore we have
703: $$b_1a_{22}-b_2a_{12}=-32k^4\pi^4\alpha+24\pi^2k^2\rho^2\alpha+o(\alpha),\ \ \
704: a_{11}a_{22}-a_{12}a_{21}=8k\pi^2\alpha+o(\alpha),$$
705: which lead to (3.11).
706:
707:
708: Now we consider two-periodic wave solution of the KP equation
709: ($N=2$). Let $\hat\rho=(k_1\rho_1, k_2\rho_2),\
710: \hat\omega=(k_1\omega_1, k_2\omega_2).$ Substituting (3.4) into
711: (3.3), we have {\footnotesize\begin{eqnarray*} &&Gf\cdot
712: f=\sum_{m,n\in Z^2} G(D_x,D_y,D_t)e^{2\pi i<\xi,n>+\pi i<\tau
713: n,n>}\cdot e^{2\pi i<\xi,m>+\pi
714: i<\tau m,m>}\\
715: &&=\sum_{m,n\in Z^2}G(2\pi
716: i<n-m,k>,2\pi i<n-m,\hat\rho>,2\pi i<n-m,\hat\omega>)e^{2\pi i<\xi,n+m>+\pi i (<\tau m,m>+<\tau n, n>)}\\
717: &&\stackrel{n+m=m'}{=}\sum_{m'\in
718: Z^2}\sum_{n_1,n_2=-\infty}^{\infty}G(2\pi i<2n-m',k>,2\pi
719: i<2n-m',\hat\rho>,2\pi
720: i<2n-m',\hat\omega>)\\
721: &&\ \ \ \ \ \ \ \ \ \ \ \ \ \times \exp(\pi i(<\tau
722: (n-m'),n-m'>+<\tau n, n>))
723: \exp(2\pi i<\xi,m'>)\\
724: &&\equiv\sum_{m'\in Z^2}\bar{G}(m_1', m_2') e^{2\pi i <\xi,m'>}=0.
725: \end{eqnarray*}}
726: It is easy to calculate that
727: {\footnotesize\begin{eqnarray*}
728: &&\bar{G}(m_1', m_2')\\
729: && =\sum_{n_1,n_2=-\infty}^{\infty}G(2\pi i<2n-m',k>,2\pi
730: i<2n-m',\hat\rho>,2\pi
731: i<2n-m',\hat\omega>)e^{\pi i(<\tau (n-m'),n-m'>+<\tau n, n>)}\\
732: &&{\small =\sum_{n_1,n_2=-\infty}^{\infty}G\left(2\pi
733: i\sum_{j=1}^{2}(2n_j'-(m_j'-2\delta_{jl}))k_j,2\pi
734: i\sum_{j=1}^{2}(2n_j'-(m_j'-2\delta_{jl}))k_j\rho_j,2\pi
735: i\sum_{j=1}^{2}(2n_j'-(m_j'-2\delta_{jl}))k_j\omega_j\right)}\\
736: &&\times \exp\left\{\pi
737: i\sum_{j,k=1}^{2}[(n_j'+\delta_{jl})\tau_{jk}(n_k'+\delta_{kl})+((m_j'-2\delta_{jl}-n_j')+\delta_{jl})
738: \tau_{jk}(m_k'-2\delta_{kl}-n_k')+\delta_{kl})]\right\}, \\
739: &&=\left\{\begin{matrix}\bar{G}(m_1'-2,m_2')e^{2\pi
740: i(m_1'-1)\tau_{11}+2\pi im_2'\tau_{12}},\ \ \ l=1\cr\cr
741: \bar{G}(m_1',m_2'-2) e^{2\pi i(m_2'-1)\tau_{22}+2\pi
742: im_1'\tau_{12}},\ \ \ l=2
743: \end{matrix}\right.
744: \end{eqnarray*}}
745: which implies that if
746: $$\bar{G}(0,0)=\bar{G}(0,1)=\bar{G}(1,0)=\bar{G}(1,1)=0,$$
747: then $\bar{G}(m_1',m_2')=0$ and (3.2) is an exact solution
748: of the KP equation. Denote
749: \begin{eqnarray*}
750: &&a_{j1}=-4\pi^2k_1\sum_{n_1,n_2=-\infty}^{\infty}
751: <2n-m^j,k>(2n_1-m_1^j)\delta_j(n),\\
752: &&a_{j2}=-4\pi^2k_2\sum_{n_1,n_2=-\infty}^{\infty}
753: <2n-m^j,k>(2n_2-m_2^j)\delta_j(n)\\
754: &&a_{j3}=24\pi^2\sum_{n_1,n_2=-\infty}^{\infty}
755: <2n-m^j,k>^2\delta_j(n),\\
756: &&a_{j4}=\sum_{n_1,n_2=-\infty}^{\infty}\delta_j(n)\\
757: && b_j=\sum_{n_1,n_2=-\infty}^{\infty}(16\pi^4
758: <2n-m^j,k>^4-12\pi^2<2n-m^j,\hat\rho>^2)\delta_j(n),\\
759: &&\delta_j(n)=e^{\pi i<\tau(n-m^j),n-m^j>+\pi i<\tau n,n>},\quad j=1, 2, 3, 4\\
760: &&m^1=(0,0),\ \ m^2=(1,0),\ \ m^3=(0,1),\ \ m^4=(1,1)\\
761: &&A=(a_{kj})_{4\times 4},\ \ b=(b_1, b_2, b_3, b_4)^T,
762: \end{eqnarray*}
763: then we have $$A\left(\begin{matrix}\omega_1\cr\omega_2\cr u_0\cr
764: c\end{matrix}\right)=b,\eqno(3.12)$$
765: from which we obtain
766: $$\omega_1=\frac{\Delta_1}{\Delta},\ \ \omega_2=\frac{\Delta_2}{\Delta},\ \
767: u_0=\frac{\Delta_3}{\Delta},\quad c_0=\frac{\Delta_4}{\Delta}\eqno(3.13)$$
768: where $\Delta=|A|$ and $\Delta_j$ are produced from $\Delta$ by
769: replacing its $j$th column with $b$ respectively. Therefore, we
770: arrive at two-periodic wave solution
771: $$u=u_0+2(\ln f)_{xx},\eqno(3.14)$$
772: where $f$ and $\omega_1, \omega_2$ are given by (3.4) and (3.13)
773: respectively. The properties of the solution are shown in
774: fig.12-fig.16. Fig.12 and fig.13 shown effect of parameters $k_1,
775: k_2$ on period and shape of wave. Fig.14 gives effect of parameter
776: $\tau_{11}, \tau_{12}, \tau_{22}$ on period and shape of wave. The
777: fig.15 and fig.16 show that two-periodic wave can degenerate to
778: one-periodic wave when $k_1$ is sufficient small.
779:
780: \begin{figure}[h]
781: $ \ \ \ \ $ {\footnotesize $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
782: \ $}\
783: {\footnotesize $(a)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
784: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
785: \ \ \ \ $ \\[6pt]
786: \begin{minipage}[u]{0.48\linewidth}
787: \centering \epsfig{figure=kpfig21.eps,width=2.8 in }
788: \end{minipage}\\[12pt]
789: $ \ \ \ \ $ {\footnotesize $(b) \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
790: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $}\
791: \ \ \ \ \ \ \ \ \
792: \ \ {\footnotesize $(c)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
793: \ \ \ \ $ \\[6pt]
794: \begin{minipage}[u]{0.48\linewidth}
795: \centering \epsfig{figure=kpfig22.eps,width=2.8 in }
796: \end{minipage}
797: \begin{minipage}[v]{0.48\linewidth}
798: \centering \epsfig{file=kpfig23.eps,width=2.8 in}
799: \end{minipage}
800: \begin{center}
801: \vspace{6mm}
802: {\footnotesize Figure 12. Two-periodic wave
803: for KP equation and the effect of parameters on wave shape: \\ (a)
804: along x-axis, (b) along y-axis, (c) along t-axis, where $k_1=0.2,
805: k_2=-0.3,$ $\tau_{11}=0.1i, \tau_{12}=0.2i, \tau_{22}=3i,
806: \rho_1=1, \rho_2=2$.}
807: \end{center}
808: \end{figure}
809:
810:
811: \begin{figure}[h]
812: $ \ \ \ \ $ {\footnotesize $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
813: \ $}\
814: {\footnotesize $(a)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
815: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
816: \ \ \ \ $ \\[6pt]
817: \begin{minipage}[u]{0.48\linewidth}
818: \centering \epsfig{figure=kpfig24.eps,width=2.8 in }
819: \end{minipage}\\[12pt]
820: $ \ \ \ \ $ {\footnotesize $(b) \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
821: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $}\
822: \ \ \ \ \ \ \ \ \
823: \ \ {\footnotesize $(c)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
824: \ \ \ \ $ \\[6pt]
825: \begin{minipage}[u]{0.48\linewidth}
826: \centering \epsfig{figure=kpfig25.eps,width=2.8 in }
827: \end{minipage}
828: \begin{minipage}[v]{0.48\linewidth}
829: \centering \epsfig{file=kpfig26.eps,width=2.8 in}
830: \end{minipage}
831: \begin{center}
832: \vspace{6mm}
833: {\footnotesize Figure 13. Two-periodic wave
834: for KP equation and the effect of parameters on wave shape: \\ (a)
835: along x-axis, (b) along y-axis, (c) along t-axis, where $k_1=0.1,
836: k_2=-0.3,$ $\tau_{11}=0.1i, \tau_{12}=0.2i, \tau_{22}=3i,
837: \rho_1=1, \rho_2=2$.}
838: \end{center}
839: \end{figure}
840:
841:
842:
843: \begin{figure}[h]
844: $ \ \ \ \ $ {\footnotesize $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
845: \ $}\
846: {\footnotesize $(a)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
847: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
848: \ \ \ \ $ \\[6pt]
849: \begin{minipage}[u]{0.48\linewidth}
850: \centering \epsfig{figure=kpfig27.eps,width=2.8 in }
851: \end{minipage}\\[12pt]
852: $ \ \ \ \ $ {\footnotesize $(b) \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
853: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $}\
854: \ \ \ \ \ \ \ \ \
855: \ \ {\footnotesize $(c)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
856: \ \ \ \ $ \\[6pt]
857: \begin{minipage}[u]{0.48\linewidth}
858: \centering \epsfig{figure=kpfig28.eps,width=2.8 in }
859: \end{minipage}
860: \begin{minipage}[v]{0.48\linewidth}
861: \centering \epsfig{file=kpfig29.eps,width=2.8 in}
862: \end{minipage}
863: \begin{center}
864: \vspace{6mm}
865: {\footnotesize Figure 14. Two-periodic wave
866: for KP equation and the effect of parameters on wave shape: \\ (a)
867: along x-axis, (b) along y-axis, (c) along t-axis, where $k_1=0.2,
868: k_2=-0.3, $ $\tau_{11}=0.1i, \tau_{12}=0.2i, \tau_{22}=i,
869: \rho_1=0.2, \rho_2=0.3$.}
870: \end{center}
871: \end{figure}
872:
873:
874: \begin{figure}[h]
875: $ \ \ \ \ $ {\footnotesize $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
876: \ $}\
877: {\footnotesize $(a)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
878: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
879: \ \ \ \ $ \\[6pt]
880: \begin{minipage}[u]{0.48\linewidth}
881: \centering \epsfig{figure=kpfig210.eps,width=2.8 in }
882: \end{minipage}\\[12pt]
883: $ \ \ \ \ $ {\footnotesize $(b) \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
884: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $}\
885: \ \ \ \ \ \ \ \ \
886: \ \ {\footnotesize $(c)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
887: \ \ \ \ $ \\[6pt]
888: \begin{minipage}[u]{0.48\linewidth}
889: \centering \epsfig{figure=kpfig211.eps,width=2.8 in }
890: \end{minipage}
891: \begin{minipage}[v]{0.48\linewidth}
892: \centering \epsfig{file=kpfig212.eps,width=2.8 in}
893: \end{minipage}
894: \begin{center}
895: \vspace{6mm}
896: {\footnotesize Figure 15. Two-periodic wave
897: for KP equation and the effect of parameters on wave shape: \\ (a)
898: along x-axis, (b) along y-axis, (c) along t-axis, where
899: $k_1=0.01, k_2=-0.3, $ $\tau_{11}=0.1i, \tau_{12}=0.2i,
900: \tau_{22}=i, \rho_1=0.2, \rho_2=0.3$.}
901: \end{center}
902: \end{figure}
903:
904:
905: \begin{figure}[h]
906: $ \ \ \ \ $ {\footnotesize $ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
907: \ $}\
908: {\footnotesize $(a)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
909: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
910: \ \ \ \ $ \\[6pt]
911: \begin{minipage}[u]{0.48\linewidth}
912: \centering \epsfig{figure=kpfig213.eps,width=2.8 in }
913: \end{minipage}\\[12pt]
914: $ \ \ \ \ $ {\footnotesize $(b) \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
915: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $}\
916: \ \ \ \ \ \ \ \ \
917: \ \ {\footnotesize $(c)$}$ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
918: \ \ \ \ $ \\[6pt]
919: \begin{minipage}[u]{0.48\linewidth}
920: \centering \epsfig{figure=kpfig214.eps,width=2.8 in }
921: \end{minipage}
922: \begin{minipage}[v]{0.48\linewidth}
923: \centering \epsfig{file=kpfig215.eps,width=2.8 in}
924: \end{minipage}
925: \begin{center}
926: \vspace{6mm}
927: {\footnotesize Figure 16. Two-periodic wave
928: for KP equation and the effect of parameters on wave shape: \\ (a)
929: along x-axis, (b) along y-axis, (c) along t-axis, where
930: $k_1=0.001, k_2=-0.3, $ $\tau_{11}=0.1i, \tau_{12}=0.2i,
931: \tau_{22}=i, \rho_1=0.2, \rho_2=0.3$.}
932: \end{center}
933: \end{figure}
934:
935: Two-soliton solution of the KP equation can be obtained as limit
936: of the periodic solution (3.14). We write $f$ as
937: \begin{eqnarray*}
938: && f=1+(e^{2\pi i\xi_1}+e^{-2\pi i\xi_1})e^{\pi i\tau_{11}}
939: +(e^{2\pi i\xi_2}+e^{-2\pi i\xi_2})e^{\pi
940: i\tau_{22}}\\
941: &&+(e^{2\pi i(\xi_1+\xi_2)}+e^{-2\pi i(\xi_1+\xi_2)})e^{\pi
942: i(\tau_{11}+2\tau_{12}+\tau_{22})}+\cdots
943: \end{eqnarray*}
944: Setting $\gamma'_1=\gamma_1+\frac{1}{2}\tau_{11},\
945: \gamma'_2=\gamma_2+\frac{1}{2}\tau_{22},\ \xi_1'=2\pi i\xi_1-\pi
946: i\tau_{11},\ \ \xi_2'=2\pi i\xi_2-\pi i\tau_{22},
947: \ \ \tau_{12}=i{\tau}'$ (${\tau}'$ is a real),
948: we get
949: \begin{eqnarray*}
950: &&f=1+e^{\xi_1'}+e^{\xi_2'}+e^{\xi_1'+\xi_2'+2\pi i\tau_{12}
951: }+\alpha_1^2e^{-\xi_1'}
952: +\alpha_2^2e^{-\xi_2'}+\alpha_1^2\alpha_2^2e^{-\xi_1-\xi_2+2\pi i\tau_{11}}+\cdots\\
953: &&\longrightarrow 1+e^{\xi_1'}+e^{\xi_2'}+e^{\xi_1'+\xi_2'-2\pi
954: {\tau'}},\ \ {\rm as}\ \ \alpha_1, \alpha_2
955: \longrightarrow 0
956: \end{eqnarray*}
957: where
958: $$\alpha_1=e^{\pi i\tau_{11}},\ \ \alpha_2=e^{\pi
959: i\tau_{22}}, \ \ \xi_j'=2\pi
960: i[k_j(x+\rho_j+\omega_jt)+\gamma'_j],$$
961: $$ e^{-2\pi
962: {\tau'}}=\frac{(k_1-k_2)^2-(\rho_1-\rho_2)^2}{(k_1+k_2)^2-(\rho_1-\rho_2)^2},\
963: \ \ \omega_i'\rightarrow -4k_j^3\pi^2+3k\rho_j^2,\ \ \ \ {\rm
964: as}\ \ \ \ \alpha_1, \alpha_2\rightarrow0.$$
965: \\[12pt]
966: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
967: {\leftline{\normalsize {\bf Acknowledgments}}}
968:
969: The work described in this paper was partially supported by a
970: grant from City University of Hong Kong (Project No. 9040967).
971: \newpage
972:
973: \begin{thebibliography}{99}
974: \footnotesize
975: \baselineskip=18pt
976: \bibitem{ab} M.J. Ablowitz and P.A. Clarkson, Solitons, Nonlinear Evolution
977: Equations and Inverse Scattering, Cambridge University Press, 1991.
978:
979: \bibitem{be} R. Beals and R.R. Coifman, Comm. pure. Appl. Math. 37 (1984) 39.
980:
981: \bibitem{} V.B. Matveev and M.A. Salle, Darboux transformation and solitons,
982: Springer, Berlin, 1991.
983:
984: \bibitem{} C.H. Gu, H.S. Hu and Z.X. Zhou, Darboux Transformations in Soliton
985: Theory and its Geometric Applications,Shanghai Science Technology
986: Publisher, Shanghai, 1999.
987:
988: \bibitem{} S.B. Leble and N.V. Ustinov, J. Phys. A 26 (1993) 5007.
989:
990: \bibitem{} P.G. Esteevez, J. Math. Phys. 40 (1999) 1406.
991:
992: \bibitem{} R. Hirota and J. Satsuma, Prog. Theor. Phys. {\bf 57} (1977) 797.
993: \bibitem{} R. Hirota, Direct methods in soliton theory, Springer, 2004.
994: \bibitem{} X. B. Hu, P. A. Clarkson, J. Phys. A,{\bf 28}(1995), 5009.
995: \bibitem{} X. B. Hu, C X Li, J.J. C. Nimmo, G. F. Yu, J. Phys A,
996: {\bf 38}(2005), 195.
997: \bibitem{} R Hirota, Y. Ohta, J. Phys Soc. Jpn {\bf 60}(1991) 798
998: \bibitem{} J X Zhao, C X Li, X. B Hu, J. Phys Soc. Jpn {\bf 73}(2004) 1159
999:
1000: \bibitem{} B. A. Dubrovin, Funct. Anal. Appl. {\bf 9} (1975) 41.
1001: \bibitem{} A. Its and V. Matveev, Funct. Anal. Appl. {\bf 9} (1975) 69.
1002: \bibitem{} E. Belokolos, A. Bobenko, V. Enol'skij, A, Its and V. Matveev,
1003: {\it Algebro-Geometrical Approach to Nonlinear Integrable
1004: Equations}, Springer, Berlin, 1994.
1005: \bibitem{} S.P. Novikov, and S.V. Manakov, L.P. Pitaeskii and V.E. Zakharov,
1006: {\it Theory of Solitons, The Snverse Scattering Methods}, Nauka, Moscow, 1980.
1007: \bibitem{} P. L. Christiansen, J. C. Eilbeck, V. Z. Enolskii and N. A.
1008: Kostov, Proc. R. Soc. London A Math. {\bf 451} (1995) 685
1009: \bibitem{} R. G. Zhou, J. Math. Phys. {\bf 38} (1997) 2535.
1010: \bibitem{} C. W. Cao, Y. T. Wu and X. G. Geng, J. Math. Phys. {\bf 40} 1999, 3948.
1011: \bibitem{} X. G. Geng and Y. T. Wu, J. Math. Phys. {\bf 40} (1999), 2971.
1012: \bibitem{} X. G. Geng, C. W. Cao and H. H. Dai, J. Phys. A, {\bf 34} 2001, 989.
1013: \bibitem{} Z. J. Qiao, Reviews in Math. Phys. {\bf 13} (2001), 545.
1014: \bibitem{} C. W. Cao, X. G. Geng and H. Y. Wang, J. Math. Phys. {\bf 43} 2002, 621.
1015: \bibitem{} H. H. Dai and X. G. Geng, Chaos, Solitons and Fractrals, {\bf 18}(2003), 1031
1016: \bibitem{} B. Deconinck, M. V. Hoeij, Physica D, {\bf 152-153}(2001) 28
1017: \bibitem{} B Docninck, M Meil, A. Bobenko, M. V. Hoeij and M. Schmies,
1018: Math. of Comput. 1417.
1019: \end{thebibliography}
1020: \end{document}
1021: