Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Research Article
The Semidirect Sum of Lie Algebras and Its Applications
to C-KdV Hierarchy
Copyright © 2014 Xia Dong et al. This is an open access article distributed under the Creative Commons Attribution License, which
permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
̃ integrable coupling of C-KdV hierarchy and its bi-Hamiltonian structures are obtained by Tu scheme
By use of the loop algebra 𝐺,
and the quadratic-form identity. The method can be used to produce the integrable coupling and its Hamiltonian structures to the
other integrable systems.
𝐼 satisfies [𝐼, 𝐴] ⊆ 𝐼 for arbitrary Lie algebra 𝐴; then 𝐼 is determine the C-KdV hierarchy of Soliton equations
called Lie ideal.
Lie algebra 𝐴 is called simple Lie algebra if 𝐴 has 𝐴 and 𝑞 0 −𝜕 𝑎 𝛿𝐻
0 as Lie ideal and without other Lie ideal. Semisimple Lie 𝑢𝑡 = ( ) = ( ) ( 𝑛+1 ) = 𝐽 𝑛 , (13)
𝑟 𝑡 −𝜕 0 −𝑐𝑛+1 𝛿𝑢
algebra 𝑅 can be written as
𝑅 = ⨁𝐴 𝑖 , (6) where
𝑖
𝑎𝑛+1
where 𝐴 𝑖 is simple Lie algebra. We have already known 𝐻𝑛 = , (𝑛 ≥ 1) ,
𝑛
that 𝐴 𝑛 , 𝐵𝑛 , 𝐶𝑛 , 𝐷𝑛 , 𝐸6,7,8 , 𝐹4 , and 𝐺2 are all semisimple Lie (14)
algebras which has been studied by Cartan long ago [5]. We 1
also know that Lie algebra 𝑅 can be written as 𝐻1 = 𝑟, 𝐻2 = (−𝑟𝑥 + 2𝑞𝑟) .
2
𝑅 = 𝑅1 z 𝑅2 , (7)
4. A New Integrable Coupling of
where 𝑅1 is semisimple Lie algebras and 𝑅2 is solvable Lie the C-KdV Hierarchy
algebras [3, 6, 7] and z denote the semidirect sum. So we
can apply the above basic principle to integrable coupling In what follows, we expand Lie algebra 𝐺 into a bigger one as
systems. the following Lie algebra 𝐺:
3. C-KdV Hierarchy 𝑒 0 𝑒 0 𝑒 0
ℎ1 = ( 1 ) , ℎ2 = ( 2 ) , ℎ3 = ( 3 ) ,
0 𝑒1 0 𝑒2 0 𝑒3
Firstly, let us recall the construction of the C-KdV hierarchy
[8, 9]. Consider the basis of 𝐺 0 𝑒 0 𝑒 0 𝑒
ℎ4 = ( 1 ) , ℎ5 = ( 2 ) , ℎ6 = ( 3 ) .
0 0 0 0 0 0
1 0 0 1 0 0
𝑒1 = ( ), 𝑒2 = ( ), 𝑒3 = ( ). (8) (15)
0 −1 0 0 1 0
̃ is presented as 𝑒𝑖 (𝑛) = 𝑒𝑖 𝜆𝑛 .
The loop algebra 𝐺 We do this along with the following commutative relations:
The C-KdV spectral problem reads as
[ℎ1 , ℎ2 ] = 2ℎ2 , [ℎ1 , ℎ3 ] = −2ℎ3 , [ℎ1 , ℎ5 ] = 2ℎ5 ,
𝜓𝑥 = 𝑈𝜓, 𝜆𝑡 = 0
𝑞−𝜆 [ℎ1 , ℎ6 ] = −2ℎ6 , [ℎ2 , ℎ3 ] = ℎ1 , [ℎ2 , ℎ4 ] = −2ℎ5 ,
−𝑟
2 𝑞 (9) [ℎ2 , ℎ6 ] = ℎ4 , [ℎ3 , ℎ4 ] = 2ℎ6 , [ℎ3 , ℎ5 ] = −ℎ4 ,
𝑈= ( −𝑞 + 𝜆 ) , 𝑢 = ( ).
1 𝑟
2 [ℎ1 , ℎ4 ] = [ℎ2 , ℎ5 ] = [ℎ3 , ℎ6 ] = [ℎ4 , ℎ5 ]
= [ℎ4 , ℎ6 ] = [ℎ5 , ℎ6 ] = 0.
Upon setting 𝑉 = Σ𝑚≥0 (𝑎𝑚 𝑒1 (−𝑚) + 𝑏𝑚 𝑒2 (−𝑚) + 𝑐𝑚 𝑒3 (−𝑚)),
(16)
solving the stationary zero curvature equation,
𝑉𝑥 = [𝑈, 𝑉] , (10) Taking 𝐺1 = span{ℎ1 , ℎ2 , ℎ3 } and 𝐺2 = span{ℎ4 , ℎ5 , ℎ6 }, it is
easy to verify that
engenders
The compatibility conditions of the spectral problems [ℎ1 (𝑖, 𝑚) , ℎ2 (𝑗, 𝑛)] = 2ℎ2 (𝛿𝑖𝑗 , 𝑚 + 𝑛 + 𝜌𝑖𝑗 )
(𝑛)
𝜓𝑥 = 𝑈𝜓, 𝜓𝑡 = 𝑉 𝜓; [ℎ1 (𝑖, 𝑚) , ℎ3 (𝑗, 𝑛)] = −2ℎ3 (𝛿𝑖𝑗 , 𝑚 + 𝑛 + 𝜌𝑖𝑗 )
1 (12)
𝑉(𝑛) = (𝜆𝑛 𝑉)+ + 𝑐𝑛+1 𝑒1 (0) , 𝑛 ≥ 0, [ℎ1 (𝑖, 𝑚) , ℎ5 (𝑗, 𝑛)] = 2ℎ5 (𝛿𝑖𝑗 , 𝑚 + 𝑛 + 𝜌𝑖𝑗 )
2
Abstract and Applied Analysis 3
[ℎ1 (𝑖, 𝑚) , ℎ6 (𝑗, 𝑛)] = −2ℎ6 (𝛿𝑖𝑗 , 𝑚 + 𝑛 + 𝜌𝑖𝑗 ) 𝑐𝑥 (1, 𝑚) = 𝑐 (0, 𝑚) − 𝑞𝑐 (1, 𝑚) + 2𝑎 (1, 𝑚) ,
[ℎ2 (𝑖, 𝑚) , ℎ3 (𝑗, 𝑛)] = ℎ1 (𝛿𝑖𝑗 , 𝑚 + 𝑛 + 𝜌𝑖𝑗 ) 𝑑𝑥 (0, 𝑚) = − 𝑟𝑓 (0, 𝑚) − 𝑒 (0, 𝑚) + 𝑢2 𝑐 (0, 𝑚) ,
[ℎ2 (𝑖, 𝑚) , ℎ4 (𝑗, 𝑛)] = −2ℎ5 (𝛿𝑖𝑗 , 𝑚 + 𝑛 + 𝜌𝑖𝑗 ) 𝑑𝑥 (1, 𝑚) = − 𝑟𝑓 (1, 𝑚) − 𝑒 (1, 𝑚) + 𝑢2 𝑐 (1, 𝑚) ,
[ℎ2 (𝑖, 𝑚) , ℎ4 (𝑗, 𝑛)] = −2ℎ5 (𝛿𝑖𝑗 , 𝑚 + 𝑛 + 𝜌𝑖𝑗 ) 𝑒𝑥 (0, 𝑚) = −𝑒 (1, 𝑚 + 1) + 𝑞𝑒 (0, 𝑚)
[ℎ2 (𝑖, 𝑚) , ℎ4 (𝑗, 𝑛)] = −2ℎ5 (𝛿𝑖𝑗 , 𝑚 + 𝑛 + 𝜌𝑖𝑗 ) + 2𝑟𝑑 (0, 𝑚) + 2𝑢1 𝑏 (0, 𝑚)
[ℎ3 (𝑖, 𝑚) , ℎ4 (𝑗, 𝑛)] = 2ℎ6 (𝛿𝑖𝑗 , 𝑚 + 𝑛 + 𝜌𝑖𝑗 ) 𝑒𝑥 (1, 𝑚) = −𝑒 (0, 𝑚) + 𝑞𝑒 (1, 𝑚) + 2𝑟𝑑 (1, 𝑚)
[ℎ3 (𝑖, 𝑚) , ℎ5 (𝑗, 𝑛)] = −ℎ4 (𝛿𝑖𝑗 , 𝑚 + 𝑛 + 𝜌𝑖𝑗 ) + 2𝑢1 𝑏 (1, 𝑚) − 2𝑢2 𝑎 (1, 𝑚) ,
[ℎ1 (𝑖, 𝑚) , ℎ4 (𝑗, 𝑛)] = [ℎ2 (𝑖, 𝑚) , ℎ5 (𝑗, 𝑛)] = 0 𝑓𝑥 (0, 𝑚) = 𝑓 (1, 𝑚 + 1) − 𝑞𝑓 (0, 𝑚)
[ℎ3 (𝑖, 𝑚) , ℎ6 (𝑗, 𝑛)] = [ℎ4 (𝑖, 𝑚) , ℎ5 (𝑗, 𝑛)] = 0 + 2𝑑 (0, 𝑚) − 2𝑢1 𝑐 (0, 𝑚) ,
[ℎ4 (𝑖, 𝑚) , ℎ6 (𝑗, 𝑛)] = [ℎ5 (𝑖, 𝑚) , ℎ6 (𝑗, 𝑛)] = 0. 𝑓𝑥 (1, 𝑚) = 𝑓 (0, 𝑚) − 𝑞𝑓 (1, 𝑚)
(18)
+ 2𝑑 (1, 𝑚) − 2𝑢1 𝑐 (1, 𝑚) ,
(21)
𝛿𝑖𝑗 = 𝑖 + 𝑗, 𝜌𝑖𝑗 = 0, when 𝑖 + 𝑗 < 2, and 𝛿𝑖𝑗 = 0, 𝜌𝑖𝑗 = 1, when
𝑎 (0, 0) = V1 , 𝑑 (0, 0) = V2 ,
𝑖 + 𝑗 = 2. With the help of above equations, we consider an
isospectral problem:
𝑏 (0, 0) = 𝑐 (0, 0) = 𝑒 (0, 0) = 𝑓 (0, 0) = 0,
A direct calculation reads Take 𝑉(𝑛) = 𝑉+(𝑛) +1/2𝑐(1, 𝑛+1)ℎ1 (0, 0)+1/2𝑓(1, 𝑛+1)ℎ4 (0, 0);
(𝑛)
then the zero curvature equation
− 𝑉+𝑥 + [𝑈, 𝑉+(𝑛) ]
= 𝑏 (1, 𝑛 + 1) ℎ2 (0, 0) − 𝑐 (1, 𝑛 + 1) ℎ3 (0, 0) (24) 𝑈𝑡 − 𝑉𝑥(𝑛) + [𝑈, 𝑉(𝑛) ] = 0 (25)
0 0 0 𝜕
𝑞 1 𝑎 (1, 𝑛 + 1) + 𝑑 (1, 𝑛 + 1)
0 0 − 𝜕 0
𝑟 2 −𝑐 (1, 𝑛 + 1) − 𝑓 (1, 𝑛 + 1)
𝑢𝑡 = ( ) = ( 1 1 )⋅( )
𝑢1 0 − 𝜕 0 − 𝜕 2𝑎 (1, 𝑛 + 1)
𝑢2 𝑡 2 2 𝑐 (1, 𝑛 + 1)
1
𝜕 0 − 𝜕 0
( 2 )
𝑎 (1, 𝑛 + 1) + 𝑑 (1, 𝑛 + 1)
−𝑐 (1, 𝑛 + 1) − 𝑓 (1, 𝑛 + 1)
= 𝐽1 ( )
2𝑎 (1, 𝑛 + 1)
𝑐 (1, 𝑛 + 1)
(26)
0 0 −𝜕 𝜕2 + 𝜕𝑞
1 2
0 0 (𝜕 − 𝑞𝜕) 𝑟𝜕 + 𝜕𝑟
( 2 )
= 1 1 1
−𝜕 − (𝜕2 + 𝜕𝑞) 𝜕 − (𝜕2 + 𝜕𝑞) + 𝜕𝑢1
2 2 2
1 2
−𝜕2 + 𝑞𝜕 𝑟𝜕 + 𝜕𝑟 (𝜕 − 𝑞𝜕) + 𝑢1 𝜕 𝑟𝜕 + 𝜕𝑟 + 𝑢2 𝜕 + 𝜕𝑢2
( 2 )
𝑎 (0, 𝑛) + 𝑑 (0, 𝑛) 𝑎 (0, 𝑛) + 𝑑 (0, 𝑛)
−𝑐 (0, 𝑛) − 𝑓 (0, 𝑛) −𝑐 (0, 𝑛) − 𝑓 (0, 𝑛)
⋅( ) = 𝐽2 ( ),
2𝑎 (0, 𝑛) 2𝑎 (0, 𝑛)
𝑐 (0, 𝑛) 𝑐 (0, 𝑛)
where 𝐽1 and 𝐽2 are Hamiltonian operators. Furthermore, in the following part we will point out that there
From (22), a recurrence operator 𝐿 is obtained, which exist bi-Hamiltonian structures from constructing of Lie loop
satisfies algebras.
𝑎 (1, 𝑛 + 1) + 𝑑 (1, 𝑛 + 1) 𝑎 (0, 𝑛) + 𝑑 (0, 𝑛)
−𝑐 (1, 𝑛 + 1) − 𝑓 (1, 𝑛 + 1) −𝑐 (0, 𝑛) − 𝑓 (0, 𝑛) 5. The Bi-Hamiltonian Structures of
( ) = 𝐿( ),
2𝑎 (1, 𝑛 + 1) 2𝑎 (0, 𝑛) the Hierarchy (26)
𝑐 (1, 𝑛 + 1) 𝑐 (0, 𝑛)
(27) Let
where 6 6
𝑎 = ∑𝑎𝑖 ℎ𝑖 , 𝑏 = ∑𝑏𝑖 ℎ𝑖 . (30)
𝐿 𝑖=1 𝑖=1
−𝜕 + 𝜕−1 𝑞𝜕 𝜕−1 𝑟𝜕 + 𝑟 𝜕−1 𝑢1 𝜕 𝜕−1 𝑢2 𝜕 + 𝑢2
2 𝑞+𝜕 0 −2𝑢1 We have [𝑎, 𝑏] = (𝑎2 𝑏3 −𝑎3 𝑏2 , 2𝑎1 𝑏2 −2𝑎2 𝑏1 , 2𝑎3 𝑏1 −2𝑎1 𝑏3 , 𝑎2 𝑏6 −
=( ).
0 0 −𝜕 − 𝜕−1 𝑞𝜕 −2𝜕−1 𝑟𝜕 − 2𝑟 𝑎6 𝑏2 + 𝑎5 𝑏3 − 𝑎3 𝑏5 , 2𝑎1 𝑏5 − 2𝑎5 𝑏1 + 2𝑎4 𝑏2 − 2𝑎2 𝑏4 , 2𝑎3 𝑏4 − 2𝑎1 𝑏6 +
0 0 −1 𝜕+𝑞 2𝑎6 𝑏1 − 2𝑎4 𝑏3 )𝑇 .
(28) In what follows, from [𝑎, 𝑏]𝑇 = 𝑎𝑇 𝑅(𝑏), we get
It is easy to verify that
0 2𝑏2 −2𝑏3 0 2𝑏5 −2𝑏6
𝐽1 𝐿 = 𝐿∗ 𝐽1 = 𝐽2 . (29) 𝑏3 −2𝑏1 0 𝑏6 − 2𝑏4 0
(−𝑏2 0 2𝑏1 −𝑏5 0 2𝑏4 )
Therefore, the hierarchy (26) is Liouville integrable. Taking 𝑅 (𝑏) = (
( 0 0 0 0 2𝑏2 −2𝑏3 )
) . (31)
𝑞 = 𝑟 = 0, 𝑢1 = 𝑞/2, and 𝑢2 = −𝑟, (26) reduces 0 0 0 𝑏3 −2𝑏1 0
to (13). According to the integrable theory, the hierarchy 0 0 0 −𝑏2 0 2𝑏1
(26) is the integrable coupling of the C-KdV hierarchy. ( )
Abstract and Applied Analysis 5
Solving the matrix equation 𝑅(𝑏)𝐹 = −(𝑅(𝑏)𝐹)𝑇 for 𝐹 gives Comparison of coefficients of 𝜆−2𝑛−1 of both sides of the above
rise to equations gives
2 0 0 2 0 0
𝛿
0 0 1 0 0 1 (−𝑎 (1, 𝑛 + 1) − 𝑑 (1, 𝑛 + 1))
𝛿𝑢
(0 1 0 0 1 0)
𝐹=(
(2 0 0 0 0 0)
). (32) 𝑎 (0, 𝑛) + 𝑑 (0, 𝑛) (37)
0 0 1 0 0 0 −𝑐 (0, 𝑛) − 𝑓 (0, 𝑛)
0 1 0 0 0 0 = (−2𝑛 + 𝛾) ( ).
2𝑎 (0, 𝑛)
( ) 𝑐 (0, 𝑛)
𝑇
So we have {𝑎, 𝑏} = 𝑎 𝐹𝑏 = (2𝑎1 + 2𝑎4 )𝑏1 + (𝑎3 + 𝑎6 )𝑏2 + (𝑎2 +
In this situation, we have 𝛾 = 0.
𝑎5 )𝑏3 + 2𝑎1 𝑏4 + 𝑎3 𝑏5 + 𝑎2 𝑏6 .
So
A direct calculation reads
𝜕𝑈 𝑎 (0, 𝑛) + 𝑑 (0, 𝑛)
{𝑉, } = − 𝑎 (0) − 𝑎 (1) − 𝑑 (0) − 𝑑 (1) −𝑐 (0, 𝑛) − 𝑓 (0, 𝑛) 𝛿𝐻 (2, 𝑛)
𝜕𝜆 ( )= , (38)
2𝑎 (0, 𝑛) 𝛿𝑢
𝜕𝑈 𝑐 (0, 𝑛)
{𝑉, } = 𝑎 (0) + 𝑎 (1) + 𝑑 (0) + 𝑑 (1)
𝜕𝑞
𝑎 (1, 𝑛 + 1) + 𝑑 (1, 𝑛 + 1)
𝜕𝑈 𝐻 (2, 𝑛) = . (39)
{𝑉, } = − 𝑐 (0) − 𝑐 (1) − 𝑓 (0) − 𝑓 (1) (33) 2𝑛
𝜕𝑟
Thus the bi-Hamiltonian structures of the system (26) are
𝜕𝑈 given by
{𝑉, } = 2𝑎 (0) + 2𝑎 (1)
𝜕𝑢1
𝛿𝐻 (1, 𝑛) 𝛿𝐻 (2, 𝑛)
𝜕𝑈 𝑢𝑡 = 𝐽1 = 𝐽2 . (40)
{𝑉, } = 𝑐 (0) + 𝑐 (1) , 𝛿𝑢 𝛿𝑢
𝜕𝑢2
From the system (26), we easily give the following equations:
where 𝑎(0) = ∑𝑚≥0 𝑎(0, 𝑚)𝜆−2𝑚 and 𝑎(1) = ∑𝑚≥0 𝑎(1, 𝑚)
𝜆−2𝑚+1 ⋅ ⋅ ⋅ . 𝑞𝑡 = −2V1 𝑞𝑥𝑥 − 4V1 𝑞𝑞𝑥 − 2V3 𝑞𝑥 − 4V1 𝑟𝑥 ,
Substituting the above formulas into the quadratic-form (41)
identity yields 𝑟𝑡 = −4V1 (𝑞𝑟)𝑥𝑥 − 2V3 𝑟𝑥 − 2V1 𝑟𝑥𝑥 .
𝛿
(−𝑎 (0) − 𝑎 (1) − 𝑑 (0) − 𝑑 (1)) 6. Conclusion
𝛿𝑢
𝑎 (0) + 𝑎 (1) + 𝑑 (0) + 𝑑 (1) (34) On the one hand, we obtain a new integrable coupling of
𝜕 −𝑐 (0) − 𝑐 (1) − 𝑓 (0) − 𝑓 (1) C-KdV hierarchy by expanding a bigger Lie algebra. On the
= 𝜆−𝛾 𝜆𝛾 ( ).
𝜕𝜆 2𝑎 (0) + 2𝑎 (1) other hand, the bi-Hamiltonian structures of the integrable
𝑐 (0) + 𝑐 (1) coupling of C-KdV hierarchy are observed by use of the
quadratic-form identity.
Comparison of coefficients of 𝜆−2𝑛−2 of both sides of the above
equations leads to
Conflict of Interests
𝛿
(−𝑎 (0, 𝑛 + 1) − 𝑑 (0, 𝑛 + 1))
𝛿𝑢 The authors declare that there is no conflict of interests
regarding the publication of this paper.
𝑎 (1, 𝑛 + 1) + 𝑑 (1, 𝑛 + 1) (35)
−𝑐 (1, 𝑛 + 1) − 𝑓 (1, 𝑛 + 1)
= (−2𝑛 − 1 + 𝛾) ( ). Acknowledgments
2𝑎 (1, 𝑛 + 1)
𝑐 (1, 𝑛 + 1)
The project is in part supported by the Natural Science
To fix the 𝛾 we take 𝑛 = 0 into the above equation and find Foundation of China (Grant no. 11271008), the First-class
𝛾 = 1. Discipline of University in Shanghai, and the Shanghai
So University.
𝑎 (1, 𝑛 + 1) + 𝑑 (1, 𝑛 + 1)
−𝑐 (1, 𝑛 + 1) − 𝑓 (1, 𝑛 + 1) 𝛿𝐻 (1, 𝑛) References
( )= ,
2𝑎 (1, 𝑛 + 1) 𝛿𝑢
𝑐 (1, 𝑛 + 1) (36) [1] F. K. Guo and Y. F. Zhang, “The quadratic-form identity for
constructing the Hamiltonian structure of integrable systems,”
𝑎 (0, 𝑛 + 1) + 𝑑 (0, 𝑛 + 1) Journal of Physics A: Mathematical and General, vol. 38, no. 40,
𝐻 (1, 𝑛) = . pp. 8537–8548, 2005.
2𝑛
6 Abstract and Applied Analysis
International
Journal of Journal of
Mathematics and
Mathematical
Discrete Mathematics
Sciences