Whitham's Method and Dubrovin-Novikov Bracket in Single-Phase and Multiphase Cases
In this paper we examine in detail the procedure of averaging of the local field-theoretic Poisson brackets proposed by B.A. Dubrovin and S.P. Novikov for the method of Whitham. The main attention is paid to the questions of justification and the conditions of applicability of the Dubrovin-Novikov p...
Saved in:
| Published in: | Symmetry, Integrability and Geometry: Methods and Applications |
|---|---|
| Date: | 2012 |
| Main Author: | |
| Format: | Article |
| Language: | English |
| Published: |
Інститут математики НАН України
2012
|
| Online Access: | https://nasplib.isofts.kiev.ua/handle/123456789/149185 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Journal Title: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Cite this: | Whitham's Method and Dubrovin-Novikov Bracket in Single-Phase and Multiphase Cases / A.Y. Maltsev // Symmetry, Integrability and Geometry: Methods and Applications. — 2012. — Т. 8. — Бібліогр.: 50 назв. — англ. |
Institution
Digital Library of Periodicals of National Academy of Sciences of Ukraine| _version_ | 1859619794518540288 |
|---|---|
| author | Maltsev, A.Y. |
| author_facet | Maltsev, A.Y. |
| citation_txt | Whitham's Method and Dubrovin-Novikov Bracket in Single-Phase and Multiphase Cases / A.Y. Maltsev // Symmetry, Integrability and Geometry: Methods and Applications. — 2012. — Т. 8. — Бібліогр.: 50 назв. — англ. |
| collection | DSpace DC |
| container_title | Symmetry, Integrability and Geometry: Methods and Applications |
| description | In this paper we examine in detail the procedure of averaging of the local field-theoretic Poisson brackets proposed by B.A. Dubrovin and S.P. Novikov for the method of Whitham. The main attention is paid to the questions of justification and the conditions of applicability of the Dubrovin-Novikov procedure. Separate consideration is given to special features of single-phase and multiphase cases. In particular, one of the main results is the insensitivity of the procedure of bracket averaging to the appearance of ''resonances'' which can arise in the multi-phase situation.
|
| first_indexed | 2025-11-29T01:49:00Z |
| format | Article |
| fulltext |
Symmetry, Integrability and Geometry: Methods and Applications SIGMA 8 (2012), 103, 54 pages
Whitham’s Method and Dubrovin–Novikov Bracket
in Single-Phase and Multiphase Cases?
Andrei Ya. MALTSEV
L.D. Landau Institute for Theoretical Physics, 1A Ak. Semenova Ave.,
Chernogolovka, Moscow reg., 142432, Russia
E-mail: maltsev@itp.ac.ru
Received April 23, 2012, in final form December 11, 2012; Published online December 24, 2012
http://dx.doi.org/10.3842/SIGMA.2012.103
Abstract. In this paper we examine in detail the procedure of averaging of the local field-
theoretic Poisson brackets proposed by B.A. Dubrovin and S.P. Novikov for the method of
Whitham. The main attention is paid to the questions of justification and the conditions of
applicability of the Dubrovin–Novikov procedure. Separate consideration is given to special
features of single-phase and multiphase cases. In particular, one of the main results is the
insensitivity of the procedure of bracket averaging to the appearance of “resonances” which
can arise in the multi-phase situation.
Key words: quasiperiodic solutions; slow modulations; Hamiltonian structures
2010 Mathematics Subject Classification: 37K05; 35B10; 35B15; 35B34; 35L65
1 Introduction
As is well-known, the Whitham method [48–50] is associated with slow modulations of periodic
or quasiperiodic m-phase solutions of nonlinear systems
F i(ϕ,ϕt,ϕx, . . . ) = 0, i = 1, . . . , n, ϕ =
(
ϕ1, . . . , ϕn
)
, (1.1)
which are usually represented in the form
ϕi(x, t) = Φi
(
k(U)x+ ω(U)t+ θ0,U
)
. (1.2)
Let us note that we will consider here systems with one spatial variable x and one time
variable t. In these notations the functions k(U) and ω(U) play the role of the “wave numbers”
and “frequencies” of m-phase solutions, while the parameters θ0 represent the “initial phase
shifts”. The parameters U = (U1, . . . , UN ) can be chosen in an arbitrary way, we just assume
that they do not change under shifts of the initial phases of solutions θ0.
The functions Φi(θ) satisfy the system
F i
(
Φ, ωαΦθα , k
βΦθβ , . . .
)
≡ 0, i = 1, . . . , n, (1.3)
and we have to choose for each value of U some function Φ(θ,U) as having “zero initial phase
shift”. The corresponding set of m-phase solutions of (1.1) can be then represented in the
form (1.2). For m-phase solutions of (1.1) we have in this case k(U) = (k1(U), . . . , km(U)),
ω(U) = (ω1(U), . . . , ωm(U)), θ0 = (θ1
0, . . . , θ
m
0 ), where U = (U1, . . . , UN ) are the parameters
of a solution. We will require also that all the functions Φi(θ,U) are 2π-periodic with respect
to each θα, α = 1, . . . ,m.
?This paper is a contribution to the Special Issue “Geometrical Methods in Mathematical Physics”. The full
collection is available at http://www.emis.de/journals/SIGMA/GMMP2012.html
mailto:maltsev@itp.ac.ru
http://dx.doi.org/10.3842/SIGMA.2012.103
http://www.emis.de/journals/SIGMA/GMMP2012.html
2 A.Ya. Maltsev
Consider a set Λ of functions Φ(θ + θ0,U), depending smoothly on the parameters U and
satisfying system (1.3) for all U.
In the Whitham approach the parameters U and θ0 become slowly varying functions of x
and t, U = U(X,T ), θ0 = θ0(X,T ), where X = εx, T = εt (ε→ 0).
For the construction of the corresponding asymptotic solution the functions U(X,T ) must sa-
tisfy some system of differential equations (the Whitham system). In the simplest case (see [31]),
we try to find asymptotic solutions
ϕi(θ, X, T ) =
∑
k≥0
Ψi
(k)
(
S(X,T )
ε
+ θ, X, T
)
εk (1.4)
with 2π-periodic in θ functions Ψ(k) satisfying system (1.1), i.e.
F i (ϕ, εϕT , εϕX , . . . ) = 0, i = 1, . . . , n.
The function S(X,T ) = (S1(X,T ), . . . , Sm(X,T )) is called the “modulated phase” of solu-
tion (1.4).
Assume now that the function Ψ(0)(θ, X, T ) belongs to the family Λ of m-phase solutions
of (1.1) for all X and T . We have then
Ψ(0)(θ, X, T ) = Φ
(
θ + θ0(X,T ),U(X,T )
)
, (1.5)
and
SαT (X,T ) = ωα(U(X,T )), SαX(X,T ) = kα(U(X,T )),
as follows after the substitution of (1.4) into system (1.1).
In the simplest case the functions Ψ(k)(θ, X, T ) are determined from the linear systems
L̂ij[U,θ0](X,T )Ψj
(k)(θ, X, T ) = f i(k)(θ, X, T ), (1.6)
where L̂ij[U,θ0](X,T ) is a linear operator defined by the linearization of system (1.3) on the so-
lution (1.5). The resolvability conditions of systems (1.6) in the space of periodic functions can
be written as the conditions of orthogonality of the functions f(k)(θ, X, T ) to all the “left eigen-
vectors” (the eigenvectors of the adjoint operator) of the operator L̂ij[U,θ0](X,T ) corresponding
to zero eigenvalue.
We should say, however, that the resolvability conditions of systems (1.6) can actually be quite
complicated in general multi-phase case, since the eigenspaces of the operators L̂[U,θ0] and L̂†[U,θ0]
on the space of 2π-periodic functions can be rather nontrivial in the multi-phase situation. Thus,
even the dimension of the kernels of L̂[U,θ0] and L̂†[U,θ0] can depend in a highly nontrivial way
on the values of U. In general, the picture arising in the U-space can be rather complicated. As
a result, the determination of the next corrections from systems (1.6) is impossible in general
multiphase situation and the corrections to the main approximation (1.5) have more complicated
and rather nontrivial form [4–6].
These difficulties do not arise commonly in the single-phase situation (m = 1) where the
behavior of eigenvectors of L̂[U,θ0] and L̂†[U,θ0], as a rule, is quite regular. The resolvability
conditions of system (1.6) for k = 1
L̂ij[U,θ0](X,T )Ψj
(1)(θ, X, T ) = f i(1)(θ, X, T ) (1.7)
with relations kT = ωX define in this case the Whitham system for the single-phase solutions
of (1.1) which plays the central role in considering the slow modulations.
Whitham’s Method and Dubrovin–Novikov Bracket 3
For the multi-phase solutions the Whitham system is usually given by the orthogonality
conditions of the right-hand part of (1.7) to the maximal set of “regular” left eigenvectors
corresponding to zero eigenvalues which are defined for all values of U and depend smoothly
on U. As a rule, this system is equivalent to the conditions obtained by the averaging of some
complete set of conservation laws of system (1.1) having the local form
P νt (ϕ,ϕt,ϕx, . . . ) = Qνx(ϕ,ϕt,ϕx, . . . ), ν = 1, . . . , N.
The Whitham system is written usually as a system of hydrodynamic type
UνT = V ν
µ (U)UµX (1.8)
and gives the main approximation for the connection between spatial and time derivatives of the
parameters Uν(X,T ). The variables T and X represent the “slow” variables T = εt, X = εx,
connected with the variables t and x by a small parameter ε. Thus, the Whitham system (1.8)
represents a homogeneous quasi-linear system of hydrodynamic type connecting the derivatives
of slow modulated parameters.
As mentioned above, the construction of the asymptotic series (1.4) in the multi-phase case
is impossible in general situation (see [4–6]). Nevertheless, the Whitham system (1.8) and the
leading term of the expansion (1.4) play the major role in consideration of modulated solutions
also in this case, representing the main approximation for corresponding modulated solutions.
The corrections to the main term have in general more nontrivial form than (1.4), but they also
tend to zero in the limit ε→ 0 [4–6].
Let us give here just some incomplete list of the classical papers devoted to the foundations of
the Whitham method [1,4–8,14–16,22–25,28,29,31,39,40,42,48–50]. We will be interested here
only in Hamiltonian aspects of the Whitham method. In the remaining part of the Introduction
we will give the definition of the “regular” Whitham system for the complete regular family of
m-phase solutions which will be used everywhere below.
Let us use for simplicity the notation Λ both for the family of the functions Φ(θ+θ0,U) and
the corresponding family of m-phase solutions of system (1.1), such that we will denote by Λ both
the families of the functions Φ(θ+θ0,U) in the space of 2π-periodic in all θα functions ϕ(θ) and
ϕ[U,θ0](x) = Φ(k(U)x+ θ0,U). We will assume everywhere below that the family Λ represents
a smooth family of m-phase solutions of system (1.1) in the sense discussed above.
It is generally assumed that the parameters kα, ωα are independent on the family Λ, such that
the full family of the m-phase solutions of (1.1) depends on N = 2m+ s (s ≥ 0) parameters Uν
and m initials phase shifts θα0 . In this case it is convenient to represent the parameters U in the
form U = (k,ω,n), where k represents the wave numbers, ω are the frequencies of the m-phase
solutions and n = (n1, . . . , ns) are some additional parameters (if any).
It is easy to see that the functions Φθα(θ + θ0,k,ω,n), α = 1, . . . ,m, Φnl(θ + θ0,k,ω,n),
l = 1, . . . , s, belong to the kernel of the operator L̂ij[k,ω,n,θ0]. In the regular case it is natural
to assume that the set of the functions (Φθα , Φnl) represents the maximal linearly independent
set of the kernel vectors of the operator L̂ regularly depending on the parameters (k,ω,n). For
the construction of the “regular” Whitham system we have to require the following property of
regularity and completeness of the family of m-phase solutions of system (1.1).
Definition 1.1. We call a family Λ a complete regular family of m-phase solutions of sys-
tem (1.1) if:
1) the values k = (k1, . . . , km), ω = (ω1, . . . , ωm) represent independent parameters on the
family Λ, such that the total set of parameters of the m-phase solutions can be represented
in the form (U,θ0) = (k,ω,n,θ0);
4 A.Ya. Maltsev
2) the functions Φθα(θ + θ0,k,ω,n), Φnl(θ + θ0,k,ω,n) are linearly independent and give
the maximal linearly independent set among the kernel vectors of the operator L̂ij[k,ω,n,θ0],
smoothly depending on the parameters (k,ω,n) on the whole set of parameters;
3) the operator L̂ij[k,ω,n,θ0] has exactly m+ s linearly independent left eigenvectors with zero
eigenvalue
κ
(q)
[U](θ + θ0) = κ
(q)
[k,ω,n](θ + θ0), q = 1, . . . ,m+ s,
among the vectors smoothly depending on the parameters (k,ω,n) on the whole set of
parameters.
By definition, we will call the regular Whitham system for a complete regular family of
m-phase solutions of (1.1) the conditions of orthogonality of the discrepancy f(1)(θ, X, T ) to the
functions κ
(q)
[U(X,T )](θ + θ0(X,T ))
∫ 2π
0
· · ·
∫ 2π
0
κ
(q)
[U(X,T )]i(θ + θ0(X,T ))f i(1)(θ, X, T )
dmθ
(2π)m
= 0, q = 1, . . . ,m+ s, (1.9)
with the compatibility conditions
kαT = ωαX . (1.10)
System (1.9), (1.10) gives m + (m + s) = 2m + s = N conditions at every X and T for the
parameters of the zero approximation Ψ(0)(θ, X, T ).
It is well known that the Whitham system does not include the parameters θα0 (X,T ) and
provides restrictions only to the parameters Uν(X,T ) of the zero approximation. Let us prove
here a simple lemma which confirms this property under the conditions formulated above1.
Lemma 1.1. Under the regularity conditions formulated above the orthogonality conditions (1.9)
do not contain the functions θα0 (X,T ) and give constraints only to the functions Uν(X,T ), having
the form
C(q)
ν (U)UνT −D(q)
ν (U)UνX = 0, q = 1, . . . ,m+ s,
with some functions C
(q)
ν (U), D
(q)
ν (U).
Proof. Let us write down the part f ′(1) of the function f(1), which contains the derivatives
θβ0T (X,T ) and θβ0X(X,T ). We have
f ′i(1)(θ, X, T ) = −∂F
i
∂ϕjt
(
Ψ(0), . . .
)
Ψj
(0)θβ
θβ0T −
∂F i
∂ϕjx
(
Ψ(0), . . .
)
Ψj
(0)θβ
θβ0X
− ∂F i
∂ϕjtt
(
Ψ(0), . . .
)
2ωα(X,T )Ψj
(0)θαθβ
θβ0T
− ∂F i
∂ϕjxx
(
Ψ(0), . . .
)
2kα(X,T )Ψj
(0)θαθβ
θβ0X − · · · .
1This simple fact was present in the Whitham approach from the very beginning (see [31, 48–50]). In fact,
under various assumptions it can be also shown that the additional phase shifts θα0 (X,T ) can be always absorbed
by the functions Sα(X,T ) after a suitable correction of initial data (see, e.g., [9, 23, 24, 35]). It should be noted,
however, that the corresponding phase shift can play rather important role in the weakly nonlocal case [39] (see
also [9, 33]).
Whitham’s Method and Dubrovin–Novikov Bracket 5
Let us choose the parameters U in the form
U = (k1, . . . , km, ω1, . . . , ωm, n1, . . . , ns).
We can write then
f ′i(1)(θ, X, T ) =
[
− ∂
∂ωβ
F i (Φ(θ + θ0,U), . . . ) + L̂ij
∂
∂ωβ
Φj(θ + θ0,U)
]
θβ0T
+
[
− ∂
∂kβ
F i (Φ(θ + θ0,U), . . . ) + L̂ij
∂
∂kβ
Φj(θ + θ0,U)
]
θβ0X .
The total derivatives ∂F i/∂ωβ and ∂F i/∂kβ are identically equal to zero on Λ according
to (1.3). We have then∫ 2π
0
· · ·
∫ 2π
0
κ
(q)
[U(X,T )]i(θ + θ0(X,T ))f ′i(1)(θ, X, T )
dmθ
(2π)m
≡ 0
since κ
(q)
[U(X,T )](θ + θ0(X,T )) are left eigenvectors of L̂ with the zero eigenvalues.
It is not difficult to see also that all the θ0(X,T ) in the arguments of Φ and κ(q) disappear
after the integration with respect to θ, so we get the statement of the lemma. �
We can claim then that the regular Whitham system has the following general form
∂kα
∂Uν
UνT =
∂ωα
∂Uν
UνX , α = 1, . . . ,m,
C(q)
ν (U)UνT = D(q)
ν (U)UνX , q = 1, . . . ,m+ s. (1.11)
Let us note that according to our assumptions we have here rank ||∂kα/∂Uν || = m. In generic
case the derivatives UνT can be expressed in terms of UµX and the Whitham system can be written
in the form (1.8).
2 Lagrangian and Hamiltonian formulations
of the Whitham method
Together with the formulation of Whitham’s method the Lagrangian structure of the equations
of slow modulations was proposed [48–50]. The method of averaging of Lagrangian function
introduced by Whitham can be formulated in the following way. We assume that the original
system (1.1) is lagrangian with the local action of the form
S =
∫∫
L (ϕ,ϕt,ϕx,ϕtt,ϕxt,ϕxx, . . . ) dxdt,
such that the functions F i have the form
F i(ϕ,ϕt,ϕx, . . . ) =
δS
δϕi(x, t)
=
∂L
∂ϕi
− ∂
∂t
∂L
∂ϕit
− ∂
∂x
∂L
∂ϕix
+ · · · .
Let us assume here for simplicity that the parameters (k,ω) = (k1, . . . , km, ω1, . . . , ωm) give
the complete set of independent parameters on the family of m-phase solutions (excluding the
initial phase shifts), such that the number of parameters Uν is equal to 2m.
The linearized operator L̂ij[k,ω,θ0](X,T ) in (1.6) is given now by the distribution
Lij[k,ω,θ0](θ,θ
′) =
δ2S
δΦi(θ)δΦj(θ′)
,
6 A.Ya. Maltsev
where
S =
∫ 2π
0
· · ·
∫ 2π
0
L
(
Φ, ωαΦθα , k
βΦθβ , . . .
) dmθ
(2π)m
is a self-adjoint operator.
Throughout the paper we will always understand the integration with respect to θ as the
averaging procedure. For this reason, all the integrals over dmθ will be defined with the factor
1/(2π)m. In particular, we will also assume that the variational derivatives of the type δS/δϕi(θ)
are defined as
δS ≡
∫ 2π
0
· · ·
∫ 2π
0
δS
δϕi(θ)
δϕi(θ)
dmθ
(2π)m
on the space of 2π-periodic in θ functions.
We also define here the delta function δ(θ − θ′) and its higher derivatives δθα1 ...θαs (θ − θ′)
on the space of 2π-periodic functions by the formula∫ 2π
0
· · ·
∫ 2π
0
δθα1 ···θαs (θ − θ′)ψ(θ′)
dmθ′
(2π)m
≡ ψθα1 ···θαs (θ).
The functions Φθα , α = 1, . . . ,m, represent both the left and the right eigenfunctions of the
operator L̂ij[k,ω,θ0](X,T ), corresponding to the zero eigenvalue.
Under the assumption that the family of the m-phase solutions Λ is a complete regular
family of m-phase solutions of (1.1) we assume that the functions Φθα(θ+ θ0,k,ω) are linearly
independent and give the maximal linearly independent set among the kernel vectors of the
operator L̂ij[k,ω,θ0] smoothly depending on the parameters (k,ω). The regular Whitham system
is given then by the conditions kαT = ωαX and m conditions of orthogonality of the function
f(1)(θ, X, T ) to the functions Φθα(θ + θ0,k,ω).
According to the Whitham procedure the Whitham system on the parameters (k,ω) is ob-
tained from the condition of extremality of the action
Σ(0) [S] =
∫∫∫ 2π
0
· · ·
∫ 2π
0
L
(
Φ, SαTΦθα , S
β
XΦθβ , . . .
) dmθ
(2π)m
dXdT (2.1)
under the conditions kα = SαX , ωα = SαT .
The conditions kαT = ωαX and δΣ/δSα(X,T ) = 0 give a system of 2m equations on the
parameters (k,ω).
It is not difficult to see that the system given by the variation of the “averaged” action
coincides with the conditions of orthogonality of the function f(1)(θ, X, T ) to the functions
Φθα(θ + θ0,k,ω). Indeed, let us consider the action
Σ [S,ϕ, ε] =
∫
L
(
ϕ
(
S(X,T )
ε
+ θ, X, T
)
, ε
∂
∂T
ϕ
(
S(X,T )
ε
+ θ, X, T
)
, . . .
)
dmθ
(2π)m
dXdT
= Σ(0) [S,ϕ] + εΣ(1) [S,ϕ] + ε2Σ(2) [S,ϕ] + · · ·
defined on the functions ϕ(θ, X, T ), 2π-periodic in each θα. Taking into account the relation
δΣ
δSα(X,T )
= ε−1
∫ 2π
0
· · ·
∫ 2π
0
ϕiθα (θ, X, T )
δΣ
δϕi(θ, X, T )
dmθ
(2π)m
and the invariance of the action with respect to the shifts
S(X,T )→ S(X,T ) + ∆S,
Whitham’s Method and Dubrovin–Novikov Bracket 7
it is easy to see that∫ 2π
0
· · ·
∫ 2π
0
ϕiθα (θ, X, T )
δΣ(0)
δϕi(θ, X, T )
dmθ
(2π)m
≡ 0,
δΣ(0)
δSα(X,T )
=
∫ 2π
0
· · ·
∫ 2π
0
ϕiθα (θ, X, T )
δΣ(1)
δϕi(θ, X, T )
dmθ
(2π)m
(2.2)
etc.
Substituting the functions ϕ(θ, X, T ) in the form ϕ(θ, X, T ) = Φ(θ + θ0,SX ,ST ) in the
relations above, we can see that we have to include now the additional dependence of the
functions ϕ(θ, X, T ) on SX and ST in relation (2.2). However, due to the relation
δΣ(0)
δϕi(θ, X, T )
≡ 0
on the family Λ, the relation (2.2) will not change in this situation. Taking also into account
the equality
f i(1) (θ, X, T ) =
δΣ(1)
δϕi(θ, X, T )
we get the required statement.
Under the assumption of the completeness and regularity of the family Λ we can see then
that the averaged action (2.1) defines a lagrangian structure of the regular Whitham system
in general multiphase case. We should say also that the cases with additional parameters n,
as a rule, can be also included into the scheme described above with the aid of the Whitham
“pseudo-phases” [49]. Let us note also that different questions connected with the justification
of the averaging of Lagrangian functions in different orders can be found in [7].
Another approach to the construction of the regular Whitham system is connected with the
method of averaging of conservation laws. According to further consideration of the Hamiltonian
structure of the Whitham equations we will assume now that system (1.1) is written in an
evolutionary form
ϕit = F i (ϕ,ϕx,ϕxx, . . . ) . (2.3)
The families of the m-phase solutions of (2.3) are defined then by solutions of the system
ωαϕiθα = F i
(
ϕ, kβϕθβ , . . .
)
(2.4)
on the space of 2π-periodic in each θα functions ϕ(θ).
We will assume that the conservation laws of system (2.3) have the form
P νt (ϕ,ϕx,ϕxx, . . . ) = Qνx (ϕ,ϕx,ϕxx, . . . ) ,
such that the values
Iν =
∫ +∞
−∞
P ν (ϕ,ϕx,ϕxx, . . . ) dx
represent translationally invariant conservative quantities for the system (2.3) in the case of
the rapidly decreasing at infinity functions ϕ(x). We can also define the conservation laws for
system (2.3) in the periodic case with a fixed period K
Iν =
1
K
∫ K
0
P ν (ϕ,ϕx,ϕxx, . . . ) dx,
8 A.Ya. Maltsev
or in the quasiperiodic case
Iν = lim
K→∞
1
2K
∫ K
−K
P ν (ϕ,ϕx,ϕxx, . . . ) dx.
It is natural also to define the variational derivatives of the functionals Iν with respect to the
variations of ϕ(x) having the same periodic or quasiperiodic properties as the original functions.
Easy to see then that the standard Euler–Lagrange expressions for the variational derivatives
can be used in this case.
Let us write the functionals Iν in the general form
Iν =
∫
P ν (ϕ,ϕx,ϕxx, . . . ) dx (2.5)
assuming the appropriate definition in the corresponding situations.
Let us define a quasiperiodic function ϕ(x) with fixed quasiperiods (k1, . . . , km) as a smooth
periodic function ϕ(θ) on the torus Tm which is restricted on the corresponding straight-line
winding
ϕ(kx+ θ0)→ ϕ(x).
Let us define the functionals
Jν =
∫ 2π
0
· · ·
∫ 2π
0
P ν
(
ϕ, kβϕθβ , . . .
) dmθ
(2π)m
(2.6)
on the space of 2π-periodic in θ functions.
It’s not difficult to see that the functions
ζ
(ν)
i[U](θ + θ0) =
δJν
δϕi(θ)
∣∣∣∣
ϕ(θ)=Φ(θ+θ0,U)
(2.7)
represent left eigenvectors of the operator L̂ij[U,θ0] with zero eigenvalues regularly depending on
parameters U on a fixed smooth family Λ.
Indeed, the operator L̂ij[U,θ0] is defined in this case by the distribution
Lij[U,θ0](θ,θ
′) = δijω
αδθα(θ − θ′)− δF i(ϕ, kβϕθβ , . . . )
δϕj(θ′)
∣∣∣∣
ϕ(θ)=Φ(θ+θ0,U)
.
We have∫ 2π
0
· · ·
∫ 2π
0
δJν
δϕi(θ)
(
ωαϕiθα − F i(ϕ, kβϕθβ , . . . )
) dmθ
(2π)m
≡ 0
for any translationally invariant integral of (2.3). Taking the variational derivative of this
relation with respect to ϕj(θ′) on Λ we get the required statement.
Thus, we can write
ζ
(ν)
i[U](θ) =
∑
q
cνq (U)κ
(q)
i[U](θ) (2.8)
with some smooth functions cνq (U) on a complete regular family Λ.
For the construction of the regular Whitham system on a complete regular family of m-phase
solutions of (2.3) we need a sufficient number of the first integrals (2.5) such that the values of
the functionals Jν on Λ represent the full set of parameters Uν = Jν |Λ. Besides that, we should
Whitham’s Method and Dubrovin–Novikov Bracket 9
require that the maximal linearly independent subset of the functions (2.7) give a complete set
of linearly independent left eigenvectors of the operator L̂ij[U,θ0] with zero eigenvalues among
the vectors regularly depending on the parameters U on the family Λ.
Coming back to the definition of a complete regular family of m-phase solutions of sys-
tem (2.3) we can see that in the case of a complete regular family Λ the number of linearly
independent vectors (2.7) on Λ is always finite. More precisely, if N = 2m+ s is the number of
parameters of m-phase solutions of (2.3) (excluding the initial phase shifts) then for a complete
regular family of m-phase solutions we require the presence of exactly m+s = N −m left eigen-
vectors κ
(q)
[U](θ + θ0) with zero eigenvalues, regularly depending on parameters, in accordance
with the number of the vectors Φθα , Φnl . Thus, according to Definition 1.1, we assume here
that the number of linearly independent vectors defined by formula (2.7) is exactly equal to
m+ s = N −m for a complete regular family Λ.
We should note that the conditions on the variational derivatives of Jν formulated above do
not contradict to the condition that the values Jν (ν = 1, . . . , N) can be chosen as parameters Uν
on the family of m-phase solutions. Indeed, the definition of Jν (2.6) explicitly includes the
additional m functions kα, which provide the necessary functional independence of the values
of Jν on Λ. In other words, we can use the Euler–Lagrange expressions for the variational
derivatives of Iν only on subspaces with fixed quasiperiods (k1, . . . , km). The variation of the
quasiperiods gives linearly growing variations which do not allow to use the Euler–Lagrange
expressions.
Moreover, under the assumptions formulated above, we can show that the condition of the
completeness of the variational derivatives (2.7) of the functionals Jν in the space of regular left
eigenvectors of the operator L̂ij[U,θ0] with zero eigenvalues follows in fact from the condition that
the values Uν = Jν |Λ can be chosen as the full set of parameters (excluding the initial phase
shifts) on the family Λ.
Let us make the agreement that we will always assume here that the Jacobian of the coordi-
nate transformation
(k,ω,n)→
(
U1, . . . , UN
)
is different from zero on Λ whenever we say that the values Uν(k,ω,n) represent a complete
set of parameters on Λ (excluding the initial phase shifts).
Under the conditions formulated above let us prove here the following proposition.
Proposition 2.1. Let Λ be a complete regular family of m-phase solutions of system (2.3). Let
the values (U1, . . . , UN ) of the functionals (J1, . . . , JN ) (2.6) give a complete set of parameters
on Λ excluding the initial phase shifts. Then:
1) the set of the vectors{
Φωα(θ + θ0,k,ω,n), Φnl(θ + θ0,k,ω,n), α = 1, . . . ,m, l = 1, . . . , s
}
is linearly independent on Λ;
2) the variation derivatives ζ
(ν)
i[U](θ+θ0), given by (2.7), generate the full space of the regular
left eigenvectors of the operator L̂ij[U,θ0] with zero eigenvalues on the family Λ.
Proof. Indeed, we require that the rows given by the derivatives(
∂U1
∂ωα
, . . . ,
∂UN
∂ωα
)
,
(
∂U1
∂nl
, . . . ,
∂UN
∂nl
)
10 A.Ya. Maltsev
are linearly independent on Λ. Using the expressions
∂Uν
∂ωα
=
∫ 2π
0
· · ·
∫ 2π
0
ζ
(ν)
i[U](θ) Φi
ωα(θ,U)
dmθ
(2π)m
, α = 1, . . . ,m,
∂Uν
∂nl
=
∫ 2π
0
· · ·
∫ 2π
0
ζ
(ν)
i[U](θ) Φi
nl(θ,U)
dmθ
(2π)m
, l = 1, . . . , s,
on Λ, we get that the set {Φωα ,Φnl} is linearly independent on Λ and the number of linearly
independent variation derivatives (2.7) is not less than m+ s.
We obtain then that the variation derivatives (2.7) generate in this case a space of regular
left eigenvectors of the operator L̂ij[U,θ0] with zero eigenvalues of dimension (m+ s). �
As a remark, let us note here some general fact associated with multiphase solutions of partial
differential equations. As is well known, the presence of families of multiphase quasiperiodic
solutions, as a rule, is connected with integrability of system (1.1) by the inverse scattering
methods.
The first m-phase solutions for the KdV equation given by the Novikov potentials were
introduced exactly as the functional families represented by the extremals of linear combinations
of some set of higher integrals of the system, i.e. the families where the variational derivatives
of integrals of the set become linearly dependent.
If we have a natural hierarchy of the first integrals and commuting flows of an integrable sys-
tem, some of the first integrals of the hierarchy (I1, . . . , Iq) are usually used for the construction
of m-phase solutions. So, the functions given by the conditions
c1δI
1 + · · ·+ cqδI
q = 0
for all possible (c1, . . . , cq) form a complete family of m-phase solutions of the integrable sys-
tem [41].
Thus, for Novikov potentials we have q = m+2 while the dimensions of the families ofm-phase
solutions for KdV (excluding initial phase shifts) are equal to 2m+1. The first 2m+1 of integrals
of KdV (I1, . . . , I2m+1) can be used for the construction of parameters (U1, . . . , U2m+1) on the
families of m-phase solutions of KdV. The number of linearly independent variational derivatives
of these functionals on the family of m-phase solutions is exactly equal to m+1. It’s not difficult
to show also that the variational derivatives of all the higher integrals of the KdV hierarchy are
given by linear combinations of the variational derivatives of the set (I1, . . . , Iq) on the families
of m-phase solutions.
The construction proposed in [41] in fact is used without substantial changes for many sys-
tems that are integrable by the inverse scattering methods, and represents the basic scheme
for constructing of m-phase solutions of integrable systems. This circumstance gives therefore
a convenient method of checking the above relations for most specific examples.
Let us prove here the following lemma, which we will need in further considerations.
Lemma 2.1. Let the values Uν of the functionals Jν on a complete regular family of m-phase
solutions Λ be functionally independent and give a complete set of parameters (excluding initial
phase shifts) on Λ, such that we have kα = kα(U1, . . . , UN ). Then the functionals kα(J1, . . . , JN )
have zero variational derivatives on Λ.
Proof. As we have seen, the conditions of the lemma imply the existence of m independent
relations
N∑
ν=1
λαν (U)
δJν
δϕi(θ)
∣∣∣∣
ϕ(θ)=Φ(θ+θ0,U)
≡ 0, α = 1, . . . ,m, (2.9)
on Λ.
Whitham’s Method and Dubrovin–Novikov Bracket 11
For the corresponding coordinates Uν on Λ this implies the relations
N∑
ν=1
λαν (U)dUν =
m∑
β=1
µ
(α)
β (U)dkβ(U)
for some matrix µ
(α)
β (U).
Since Uν provide coordinates on Λ the matrix µ
(α)
β (U) has the full rank, and therefore in-
vertible. We can then write
dkβ =
m∑
α=1
(
µ̂−1
)β
(α)
(U)
N∑
ν=1
λ(α)
ν (U)dUν .
The assertion of the lemma follows then from (2.9). �
The regular Whitham system in the described approach can be written as
〈P ν〉T = 〈Qν〉X , ν = 1, . . . , N, (2.10)
where 〈. . . 〉 denotes the averaging operation on Λ defined by the formula
〈f(ϕ,ϕx, . . . )〉 ≡
∫ 2π
0
· · ·
∫ 2π
0
f
(
Φ, kβΦθβ , . . .
) dmθ
(2π)m
.
Let us prove here the following lemma about the connection between the systems (2.10)
and (1.11).
Lemma 2.2. Let the values Uν of the functionals Jν on a complete regular family of m-phase
solutions Λ be functionally independent and give a complete set of parameters on Λ excluding
the initial phase shifts. Then the system (2.10) is equivalent to (1.11).
Proof. Let us introduce the functions
Πν
i(l)(ϕ,ϕx, . . . ) ≡
∂P ν(ϕ,ϕx, . . . )
∂ϕilx
for l ≥ 0.
Using the expression for the evolution of the densities P ν(ϕ, εϕX , . . . ) we can write the
following identities
P νt (ϕ, εϕX , . . . ) =
∑
l≥0
εlΠν
i(l)(ϕ, εϕX , . . . )
(
F i(ϕ, εϕX , . . . )
)
lX
≡ εQνX(ϕ, εϕX , . . . ). (2.11)
To calculate the values ε〈Qν〉X let us put now
ϕi(θ, X) = Φi
(
S(X)
ε
+ θ,U(X)
)
, (2.12)
where SαX = kα(U(X)).
The operator ε∂/∂X acting on the functions (2.12) can be naturally represented as a sum of
kα∂/∂θα and the terms proportional to ε. So, any expression f(ϕ, εϕX , . . . ) on the submani-
fold (2.12) can be naturally represented in the form
f(ϕ, εϕX , . . . ) =
∑
l≥0
εlf[l] [Φ,U] ,
12 A.Ya. Maltsev
where f[l][Φ,U] are smooth functions of (Φ,Φθα ,ΦUν , . . . ) and (U,UX ,UXX , . . . ), polynomial
in the derivatives (UX ,UXX , . . . ), and having degree l in terms of the total number of derivations
of U w.r.t. X. Note also that the functions Φ appear in f[l] with the phase shift S(X)/ε according
to (2.12). The common phase shift is not important for the integration with respect to θ, so
let us assume below that the phase shift S(X)/ε is omitted after taking all the differentiations
with respect to X.
According to (2.11) and (2.4) we can write
ε〈Qν〉X = ε
∫ 2π
0
· · ·
∫ 2π
0
QνX[1]
dmθ
(2π)m
= ε
∫ 2π
0
· · ·
∫ 2π
0
∑
l≥0
(
Πν
i(l)[0]F
i
lX[1] + Πν
i(l)[1]F
i
lX[0]
) dmθ
(2π)m
= ε
∫ 2π
0
· · ·
∫ 2π
0
∑
l≥0
(
Πν
i(l)[0]k
γ1 · · · kγlF i[1]θγ1 ...θγl
+ Πν
i(l)[0]lk
γ1 · · · kγl−1
(
ωβΦi
θβθγ1 ···θγl−1
)
X[1]
+ Πν
i(l)[0]
l(l − 1)
2
kγ1X k
γ2 · · · kγl−1ωβΦi
θβθγ1 ···θγl−1
+ Πν
i(l)[1]k
γ1 · · · kγlωβΦi
θβθγ1 ···θγl
)
dmθ
(2π)m
.
It is not difficult to see also that for arbitrary dependence of parameters U of T , the derivative
of the average 〈P ν〉 w.r.t. T can be written as
〈P ν〉T =
∫ 2π
0
· · ·
∫ 2π
0
∑
l≥0
Πν
i(l)[0]
(
kγ1 · · · kγlΦi
θγ1 ···θγl
)
T
dmθ
(2π)m
.
Now, we can write the relations 〈P ν〉T = 〈Qν〉X as∫ 2π
0
· · ·
∫ 2π
0
∑
l≥0
(
Πν
i(l)[0]k
γ1 · · · kγlΦi
θγ1 ···θγlT + Πν
i(l)[0]lk
γ1 · · · kγl−1kγlT Φi
θγ1 ···θγl
) dmθ
(2π)m
=
∫ 2π
0
· · ·
∫ 2π
0
∑
l≥0
(
Πν
i(l)[0]k
γ1 · · · kγlF i[1]θγ1 ···θγl
+ Πν
i(l)[0]lk
γ1 · · · kγl−1ωβXΦi
θβθγ1 ···θγl−1 + Πν
i(l)[0]lk
γ1 · · · kγl−1ωβΦi
θβθγ1 ···θγl−1X[1]
+ Πν
i(l)[0]
l(l − 1)
2
kγ1X k
γ2 · · · kγl−1ωβΦi
θβθγ1 ···θγl−1 + Πν
i(l)[1]k
γ1 · · · kγlωβΦi
θβθγ1 ···θγl
)
dmθ
(2π)m
.
The last three terms in the right-hand part represent the integral of the value∑
l≥0
ωβ
(
Πν
i(l)[0]Φ
i
θβ ,lX[1] + Πν
i(l)[1]Φ
i
θβ ,lX[0]
)
= ωβ∂P ν[1]/∂θ
β
and are equal to zero. The remaining terms after integration by parts can be written in the form∫ 2π
0
· · ·
∫ 2π
0
(
ζ
(ν)
i[U(X)](θ)
[
Φi
T (θ,U(X))− F i[1](θ, X)
]
+
(
kβT − ω
β
X
)∑
l≥0
Πν
i(l)[0]lk
γ1 · · · kγl−1Φi
θβθγ1 ···θγl−1
)
dmθ
(2π)m
= 0,
where the values ζ
(ν)
i[U(X)](θ) are given by (2.7).
Whitham’s Method and Dubrovin–Novikov Bracket 13
Consider the convolution (in ν) of the above expression with the values ∂kα/∂Uν . The
expressions
∂kα
∂Uν
(U(X))ζ
(ν)
i[U(X)](θ)
are identically equal to zero according to Lemma 2.1.
From the other hand we have
∂kα
∂Uν
∫ 2π
0
· · ·
∫ 2π
0
∑
l≥1
lkβ1 · · · kβl−1Φi
θβθβ1 ···θβl−1
Πν
i(l) (Φ, kγΦθγ , . . . )
dmθ
(2π)m
=
(
∂kα
∂Uν
∂
∂kβ
Jν [ϕ,k]
)∣∣∣∣
ϕ(θ)=Φ(θ,U)
= δαβ , (2.13)
since the variations of the functions Φ are insignificant for the values of kα according to
Lemma 2.1.
We get then that conditions (2.10) imply the relations kαT = ωαX , which are the first part of
system (1.11).
Now the conditions∫ 2π
0
· · ·
∫ 2π
0
ζ
(ν)
i[U(X)](θ)
[
Φi
T (θ,U(X))− F i[1](θ, X)
] dmθ
(2π)m
= 0
express the conditions of orthogonality of the vectors (2.7) to the function −ΦT + F[1], which
coincides exactly with the right-hand part of the equation (1.7) in our case. Since the linear
span of the vectors (2.7) coincides with the linear span of the complete set of the regular left
eigenvectors of the operator L̂ij[U,θ0](X,T ) with zero eigenvalues, we get that system (2.10) is
equivalent to system (1.11). �
Let us note here that it follows from Lemma 2.2 that systems (2.10), obtained from different
sets of conservation laws are equivalent to each other. In other words, if system (2.3) has
additional conservation laws then their averaging gives relations following from system (2.10).
Let us note also that the justification questions discussed above, as a rule, can be considered
in a simpler way under additional assumptions about the next corrections to the main approxi-
mation (1.5) (see, e.g. [7]). We note again that here we don’t make any additional assumptions
of this kind and consider the regular Whitham system as an independent object that is associa-
ted only with description of the main approximation (1.5). As we have said, we will follow this
approach everywhere in the paper.
The Hamiltonian properties of systems (2.10) and more general systems (1.8) play very im-
portant role in their consideration. The general theory of systems (1.8), which are Hamilto-
nian with respect to local Poisson brackets of hydrodynamic type (Dubrovin–Novikov brackets)
was constructed by B.A. Dubrovin and S.P. Novikov. Let us give here a brief description of
the Dubrovin–Novikov Hamiltonian structures and of the properties of the corresponding sys-
tems (1.8).
The Dubrovin–Novikov bracket on the space of fields (U1(X), . . . , UN (X)) has the form
{Uν(X), Uµ(Y )} = gνµ(U)δ′(X − Y ) + bνµγ (U)UγXδ(X − Y ), ν, µ = 1, . . . , N. (2.14)
The Hamiltonian operator corresponding to (2.14) can be written in the form
Ĵνµ = gνµ(U)
d
dX
+ bνµγ (U)UγX .
As was shown by B.A. Dubrovin and S.P. Novikov [14–16], expression (2.14) with non-
degenerate tensor gνµ(U) defines a Poisson bracket on the space of fields U(X) if and only if:
14 A.Ya. Maltsev
1) tensor gνµ(U) gives a symmetric flat pseudo-Riemannian metric with upper indexes on
the space of parameters (U1, . . . , UN );
2) the values
Γνµγ = −gµλbλνγ ,
where gνλ(U)gλµ(U) = δνµ, represent the Christoffel symbols for the corresponding met-
ric gνµ(U).
As follows from the statements above, every Dubrovin–Novikov bracket with non-degenerate
tensor gνµ(U) can be written in the canonical form [14–16]
{nν(X), nµ(Y )} = ενδνµδ′(X − Y ), εν = ±1,
after the transition to the flat coordinates nν = nν(U) for the metric gνµ(U).
The functionals
Nν =
∫ +∞
−∞
nν(X)dX
represent the annihilators of the Dubrovin–Novikov bracket while the functional
P =
∫ +∞
−∞
1
2
N∑
ν=1
εν(nν)2(X)dX
represents the momentum functional for the bracket (2.14).
The Hamiltonian functions in the theory of brackets (2.14) are represented by the functionals
of hydrodynamic type, i.e.
H =
∫ +∞
−∞
h(U)dX.
The bracket (2.14) has also two other important forms on the space of U(X). One of them
is the “Liouville” form [14–16] having the form
{Uν(X), Uµ(Y )} = (γνµ(U) + γµν(U))δ′(X − Y ) +
∂γνµ
∂Uλ
UλXδ(X − Y )
for some functions γνµ(U).
The “Liouville” form of the Dubrovin–Novikov bracket is called also the physical form and
corresponds to the case when the integrals of coordinates Uν
Iν =
∫ +∞
−∞
Uν(X)dX
commute with each other.
Another important form of the Dubrovin–Novikov bracket is the diagonal form. It corre-
sponds to the case when the coordinates Uν represent the diagonal coordinates for the met-
ric gνµ(U) and the tensor gνµ(U) in (2.14) has a diagonal form. This form of the Dubrovin–
Novikov bracket is closely connected with the integration theory of systems of hydrodynamic
type
UνT = V ν
µ (U)UµX
Whitham’s Method and Dubrovin–Novikov Bracket 15
which can be written in the diagonal form
UνT = V ν(U)UνX (2.15)
(no summation) and are Hamiltonian with respect to bracket (2.14).
It was conjectured by S.P. Novikov that all the systems of hydrodynamic type having the
form (2.15) and Hamiltonian with respect to any bracket (2.14) are integrable. This conjecture
was proved by S.P. Tsarev in [46] where the method (the “generalized hodograph method”) of
integration of these systems was suggested. However, the method of Tsarev proved to be appli-
cable to a wider class of diagonalizable systems of hydrodynamic type which was called by Tsarev
“semi-Hamiltonian”. As it turned out later in the class of “semi-Hamiltonian systems” fall also
the systems Hamiltonian with respect to generalizations of the Dubrovin–Novikov bracket — the
weakly nonlocal Mokhov–Ferapontov bracket [38] and the Ferapontov bracket [18, 21]. Various
aspects of the weakly nonlocal brackets of hydrodynamic type are discussed in [18–21,36,38,43].
Let us note also that the generalization of the Dubrovin–Novikov procedure for weakly nonlocal
case was proposed in [34].
Let us describe now the procedure for constructing the Dubrovin–Novikov bracket for the
Whitham system in the case when the original system (2.3) is Hamiltonian with respect to
a local field-theoretic bracket
{ϕi(x), ϕj(y)} =
∑
k≥0
Bij
(k)(ϕ,ϕx, . . . )δ
(k)(x− y) (2.16)
with the local Hamiltonian of the form
H =
∫
PH(ϕ,ϕx, . . . )dx, (2.17)
which was suggested by B.A. Dubrovin and S.P. Novikov [14–16].
Method of B.A. Dubrovin and S.P. Novikov is based on the existence of N (equal to the
number of parameters Uν of the family of m-phase solutions of (2.3)) local integrals
Iν =
∫
P ν(ϕ,ϕx, . . . )dx, (2.18)
which commute with the Hamiltonian (2.17) and with each other
{Iν , H} = 0, {Iν , Iµ} = 0 (2.19)
and can be described as follows.
We calculate the pairwise Poisson brackets of the densities P ν having the form
{P ν(x), Pµ(y)} =
∑
k≥0
Aνµk (ϕ,ϕx, . . . )δ
(k)(x− y), (2.20)
where
Aνµ0 (ϕ,ϕx, . . . ) ≡ ∂xQνµ(ϕ,ϕx, . . . )
according to (2.19).
The corresponding Dubrovin–Novikov bracket on the space of functions U(X) has the form
{Uν(X), Uµ(Y )} = 〈Aνµ1 〉(U)δ′(X − Y ) +
∂〈Qνµ〉
∂Uγ
UγXδ(X − Y ). (2.21)
16 A.Ya. Maltsev
Let us remind that we assume that the parameters Uν coincide with the values of the func-
tionals Iν defined on the corresponding quasiperiodic solutions of the family Λ
Uν = 〈P ν(x)〉.
The Whitham system (2.10) is Hamiltonian with respect to the Dubrovin–Novikov brac-
ket (2.21) with the Hamiltonian
Hav =
∫ +∞
−∞
〈PH〉 (U(X)) dX. (2.22)
The proof of the Jacobi identity for the bracket (2.21) was suggested in [32] under certain
assumptions about the family of m-phase solutions of (2.3). Besides that, it was shown in [37]
that the Dubrovin–Novikov procedure is compatible with the procedure of averaging of local
Lagrangian functions when carrying out of both the procedures is possible.
Let us note also that in [44] all the local brackets (2.14) for the Whitham equations for KdV,
NLS, and SG equations were found. Besides that, in [2] the hierarchies of the weakly nonlocal
Hamiltonian structures for the Whitham systems for KdV were represented.
In this paper we give the most detailed discussion and justification of the Dubrovin–Novikov
procedure separately for the single-phase and the multiphase cases. In the multiphase case we
will show also that the justification of the Dubrovin–Novikov procedure can be done under
weaker assumptions than in previous reviews. In particular, we will show that the justification
of the procedure is in fact insensitive to the appearance of “resonances” which can arise in
the multi-phase case, which is the basis for its widespread use in the multiphase situation. In
all cases we will consider the regular Whitham system defined above without any assumptions
about the form of higher corrections to the main approximation (1.5) of a slowly modulated
solution.
In this chapter we consider general constructions connected with the Dubrovin–Novikov pro-
cedure and prove some technical lemmas needed for its justification. At the end of the chapter,
we give a detailed justification of the procedure in the single-phase case. In the next chapter we
will consider multiphase case, where the justification of the Dubrovin–Novikov procedure will
have a more complicated form. In conclusion, we consider a simple but typical example of the
justification of the Dubrovin–Novikov procedure using the Gardner–Zakharov–Faddeev bracket
for KdV.
As we said above, we will now consider the general structures connected with the construction
of the Hamiltonian structure for the regular Whitham system.
Let us note that according to (2.19) the flows
ϕitν = Siν(ϕ,ϕx, . . . ) = {ϕi(x), Iν} (2.23)
generated by the functionals Iν according to the bracket (2.16) commute with the initial
flow (2.3). The flows (2.23) leave invariant the full families of m-phase solutions of (2.3) as
well as the values Uν = Iν of the functionals Iν on them. For a complete regular family
of m-phase solutions with independent parameters (k1, . . . , km) it’s not difficult to show that
the flows (2.23) generate linear (in time) shifts of the phases θα0 with some constant frequen-
cies ωαν(U), such that
Siν
(
Φ, kβΦθβ , . . .
)
= ωαν(U)Φi
θα(θ,U). (2.24)
According to Lemma 2.1 we have also that the functionals kα(I) should generate the zero
flows on the corresponding family of m-phase solutions of (2.3). For Uν coinciding with the
Whitham’s Method and Dubrovin–Novikov Bracket 17
values of Iν on Λ we get then the relations
∂kα(U)
∂Uν
ωβν(U) ≡ 0, α, β = 1, . . . ,m. (2.25)
Let us give here the definition of a regular Hamiltonian family of m-phase solutions of sys-
tem (2.3) and of a complete Hamiltonian set of the functionals (2.18).
Definition 2.1. We call family Λ of m-phase solutions of system (2.3) a regular Hamiltonian
family if:
1) it represents a complete regular family of m-phase solutions of (2.3) in the sense of
Definition 1.1;
2) the corresponding bracket (2.16) has on Λ constant number of annihilators N1, . . . , N s with
linearly independent variational derivatives δN l/δϕi(x) which coincides with the number
of independent annihilators in the neighborhood of Λ.
Let us say that according to the generalized Darboux theorem we can identify the num-
ber of the variational derivatives δN l/δϕi(x) on Λ with the number of linearly independent
quasiperiodic solutions v
(l)
i (x) of the equation∑
k≥0
Bij
(k)(ϕ,ϕx, . . . )
∣∣∣
Λ
v
(l)
j,kx = 0,
where v
(l)
i (x) have the same quasiperiods as the corresponding functions ϕ(x) on Λ.
Definition 2.2. We call a set (I1, . . . , IN ) of commuting functionals (2.18) a complete Hamil-
tonian set on a regular Hamiltonian family Λ of m-phase solutions of system (2.3) if:
1) the restriction of the functionals (I1, . . . , IN ) on the quasiperiodic solutions of the family
Λ gives a complete set of parameters (U1, . . . , UN ) on this family;
2) the Hamiltonian flows generated by (I1, . . . , IN ) generate on Λ linear phase shifts of θ0
with frequencies ων(U), such that
rank ||ωαν(U)|| = m;
3) the linear space generated by the variational derivatives δIν/δϕi(x) on Λ contains the
variational derivatives of all the annihilators N q of the bracket (2.16), such that
δN l
δϕi(x)
∣∣∣∣
Λ
=
∑
ν
γlν(U)
δIν
δϕi(x)
∣∣∣∣
Λ
for some smooth functions γlν(U) on the family Λ.
Let us note that it follows from Definition 2.2 that if a complete Hamiltonian set of integrals
(I1, . . . , IN ) exists for a regular Hamiltonian family Λ then the number of the additional para-
meters (n1, . . . , ns) discussed above is equal to the number of annihilators of the bracket (2.16).
Indeed, according to Definitions 2.1 and 2.2, the number of the functionals Iν having linearly
independent variational derivatives on Λ exactly equals to m + s, where s is the number of
annihilators of the bracket (2.16). The total number of independent parameters Uν on Λ is then
equal to 2m+ s due to the wave vectors kα, α = 1, . . . ,m, which implies the above assertion.
As follows from the condition (2) of Definition 2.2 and from the invariance of the functio-
nals N l and Iν with respect to the flows (2.23), the values γlν(U) can always be chosen indepen-
dent on the initial phase shifts on the family Λ. The values δIν/δϕi(x)|Λ are linearly dependent
18 A.Ya. Maltsev
on Λ, so it’s natural to choose a complete linearly independent subsystem. Remembering that
the variational derivatives of Jν (2.7) are linear combinations of the regular left eigenvectors
κ
(q)
i[U](θ + θ0), we can write
δN l
δϕi(x)
∣∣∣∣
ϕ(x)=Φ(k(U)x+θ0,U)
=
∑
q
nlq(U)κ
(q)
i[U] (k(U)x+ θ0) (2.26)
for some smooth functions nlq(U). The functions nlq(U) are then uniquely determined on Λ and
we have rank||nlq(U)|| = s by Definition 2.1.
We will need now the construction of the Dirac restriction of a Poisson bracket on a subma-
nifold. We describe here briefly this construction. Using the terminology of finite-dimensional
spaces we can say that the Dirac restriction of a Poisson bracket on a submanifold N k ⊂Mn is
associated with a special choice of coordinates in the neighborhood of the submanifold N k. The
coordinates in the neighborhood of the submanifold N k are divided into the “coordinates on the
submanifold” (U1, . . . , Uk) and the constraints (g1, . . . , gn−k) which define the submanifold N k.
It is assumed that the submanifold N k is defined by the conditions
gi(x) = 0, i = 1, . . . , n− k,
while the functions U1(x), . . . , Uk(x) on Mn play the role of a coordinate system on N k after
the restriction to this submanifold.
If the Hamiltonian flows generated by the functions U j(x) leave the submanifoldN k invariant,
i.e. we have
{U j(x), gi(x)} = 0 for g(x) = 0,
then the pairwise Poisson brackets of the functions U j(x) define the Poisson tensor in the
coordinates (U1, . . . , Uk) after restriction on N k which is called the Dirac restriction of the
Poisson bracket {·, ·} on the submanifold N k ⊂Mn.
In general, according to the procedure of Dirac, if we have some constraints gi(x) which
define the submanifold N k and some functions U j(x) which give a coordinate system on N k we
must find k linear combinations βjs(U)gs(x) at every point of N k, such that for the functions
Ũ j(x) = U j(x) + βjs(U)gs(x), j = 1, . . . , k,
we have the relations{
Ũ j(x), gi(x)
}
= 0
for g(x) = 0.
The functions Ũ j(x) take the same values that U j(x) at the points of N k and we can define
the Dirac bracket {·, ·}D on N k by the formula
{U i, U j}D =
{
Ũ i(x), Ũ j(x)
}∣∣
N k(U).
The functions βjs(U) are determined from the linear systems{
gi(x), gs(x)
}∣∣
N kβ
j
s(U) +
{
gi(x), U j(x)
}∣∣
N k = 0, i = 1, . . . , n− k,
and we can also write{
U i, U j
}
D
=
{
U i(x), U j(x)
}∣∣
N k − β
i
s(U)
{
gs(x), gq(x)
}∣∣
N kβ
j
q(U)
for the Dirac bracket on N k.
Whitham’s Method and Dubrovin–Novikov Bracket 19
We now describe the procedure of the Dirac restriction, which will be needed in our situation.
We consider now system (2.3) which is Hamiltonian with respect to some local bracket (2.16)
with a local Hamiltonian function of the form (2.17). We first introduce the extended space of
fields
ϕ(x)→ ϕ(θ, x),
where the functions ϕ(θ, x) are 2π-periodic with respect to each θα, and define the extended
Poisson bracket{
ϕi(θ, x), ϕj(θ′, y)
}
=
∑
k≥0
Bij
(k)(ϕ,ϕx, . . . )δ
(k)(x− y)δ(θ − θ′).
We also make the replacement x→ X = εx and define the Poisson bracket{
ϕi(θ, X), ϕj(θ′, Y )
}
=
∑
k≥0
εkBij
(k)(ϕ, εϕX , . . . )δ
(k)(X − Y )δ(θ − θ′) (2.27)
on the space of fields ϕ(θ, X).
Let us define again the submanifold K by the formula (2.12) in the extended space of fields.
We thus assume that the functions ϕ(θ, X) represent functions from the family Λ of the m-phase
solutions of (2.3) with some parameters U(X) at every X.
It is convenient to choose the boundary conditions in the form U(X)→ U0, X → ±∞, such
that kα(U0) = 0 and define the functions Sα(X) by the formula
Sα(X) =
1
2
∫ +∞
−∞
sgn(X − Y )kα(Y )dY.
We can write then
ϕi(θ, X) = Φi
(
S[U](X)
ε
+ θ,U(X)
)
, (2.28)
where Sα[U] is a single-valued functional of coordinates U(X) on the submanifold K.2
Let us introduce the functionals Jν(X) on the functions ϕ(θ, X) by the formula
Jν(X) =
∫ 2π
0
· · ·
∫ 2π
0
P ν
(
ϕ, εϕX , ε
2ϕXX , . . .
) dmθ
(2π)m
(2.29)
and consider their values on the functions of the family K.
We can write on K,
Jν(X) = Uν(X) +
∑
l≥1
εlJν(l)(X), (2.30)
where Jν(l)(X) are polynomials in the derivatives UX ,UXX , . . . with coefficients depending on U
and have grading degree l in terms of total number of derivations with respect to X.
The transformation (2.30) can be inverted as a formal series in ε, such that we can write
Uν(X) = Jν(X) +
∑
l≥1
εlUν(l)(X) (2.31)
2The case when the values kα = 0 are absent on the space of the parameters U requires just a simple
modification of the definition of S(X) which we do not consider here.
20 A.Ya. Maltsev
on the functions of the submanifold K. In the formula (2.31) the functions Uν(l) are functions of
J,JX ,JXX , . . . , polynomial in the derivatives JX ,JXX , . . . , and having degree l in terms of the
number of derivations w.r.t. X.
The values of the functionals Jν(X) or Uν(X) on the functions from K can thus be chosen as
coordinates on K. We can define also the functionals U(X) on the whole functional space using
the definition of the functionals J(X) and relations (2.31). Let us introduce also the constraints
gi(θ, X) defining the submanifold K by the conditions gi(θ, X) = 0 and numbered by the values
of θ and X
gi(θ, X) = ϕi(θ, X)− Φi
(
S[U[J]](X)
ε
+ θ,U[J](X)
)
. (2.32)
The constraints gi(θ, X) are functionals on the whole extended space of fields ϕi(θ, X) by
virtue of the corresponding definition of the functionals Jν(X).
The constraints (2.32) are not independent since the following relations hold identically for
the “gradients” δgi(θ, X)/δϕj(θ′, Y ) on the submanifold K∫ +∞
−∞
∫ 2π
0
· · ·
∫ 2π
0
δJν(Z)
δϕi(θ, X)
∣∣∣∣
K
δgi(θ, X)
δϕj(θ′, Y )
∣∣∣∣
K
dmθ
(2π)m
dX ≡ 0. (2.33)
Nevertheless, it will be convenient here not to choose an independent subsystem from (2.32)
and keep the system of constraints in the form (2.32) keeping in mind the existence of identi-
ties (2.33).
Thus, we consider now the values of the functionals [U(X),g(θ, X)] as a coordinate system
in the neighborhood of K with the relations (2.33). The values of the functionals U(X) will be
considered as a coordinate system on K.
Using the explicit expression for the quantities δJν(Z)/δϕi(θ, X) on K
δJν(Z)
δϕi(θ, X)
∣∣∣∣
K
=
∑
l≥0
Πν
i(l)
(
Φ
(
S(Z)
ε
+ θ,U(Z)
)
, ε
∂
∂Z
Φ
(
S(Z)
ε
+ θ,U(Z)
)
, . . .
)
εlδ(l)(Z −X),
we can write the convolution (2.33) in the form of action of the operator∫ 2π
0
· · ·
∫ 2π
0
dmθ
(2π)m
∑
l≥0
Πν
i(l)
(
Φ
(
S(Z)
ε
+ θ,U(Z)
)
, ε
∂
∂Z
Φ
(
S(Z)
ε
+ θ,U(Z)
)
, . . .
)
εl
dl
dZ l
(2.34)
on the distributions δgi(θ, Z)/δϕj(θ′, Y )|K.
Pairwise Poisson brackets of the functionals Jν(X), Jµ(Y ) have the form
{Jν(X), Jµ(Y )} =
∑
k≥0
∫ 2π
0
· · ·
∫ 2π
0
Aνµk
(
ϕ(θ, X), εϕX(θ, X), . . .
) dmθ
(2π)m
εkδ(k)(X − Y ),
where
Aνµ0 (ϕ, εϕX , . . . ) ≡ ε∂XQνµ(ϕ, εϕX , . . . ). (2.35)
Substituting the functions ϕi(θ, X) in the form (2.28) it is easy to see that the Dubrovin–
Novikov bracket is the main part (in ε) of the pairwise Poisson brackets {Jν(X), Jµ(Y )} on the
submanifold K in the coordinates [U(Z)].
Whitham’s Method and Dubrovin–Novikov Bracket 21
The Poisson brackets of the fields ϕi(θ, X) with the functionals Jµ(Y ) can be written as
{ϕi(θ, X), Jµ(Y )} =
∑
l≥0
εlCiµ(l) (ϕ(θ, X), εϕX(θ, X), . . . ) δ(l)(X − Y ) (2.36)
with some smooth functions Ciµ(l)(ϕ, εϕX , . . . ).
We also have in this case
Ciµ(0) (ϕ(θ, X), εϕX(θ, X), . . . ) ≡ Siµ (ϕ(θ, X), εϕX(θ, X), . . . ) (2.37)
by virtue of (2.23).
For any function of the slow variable q(Y ) we can write{
ϕi(θ, X),
∫
q(Y )Jµ(Y )dY
}
=
∑
l≥0
εlCiµ(l) (ϕ(θ, X), εϕX(θ, X), . . . ) qlX . (2.38)
The leading term in the expression (2.38) on K has the form{
ϕi(θ, X),
∫
q(Y )Jµ(Y )dY
}∣∣∣∣
K[0]
= Ciµ(0)
(
Φ
(
S(X)
ε
+ θ, X
)
, kβΦθβ
(
S(X)
ε
+ θ, X
)
, . . .
)
q(X)
= q(X)Siµ
(
Φ
(
S(X)
ε
+ θ, X
)
, kβΦθβ
(
S(X)
ε
+ θ, X
)
, . . .
)
,
where Siµ(ϕ,ϕx, . . . ) is the flow (2.23) generated by the functional Iµ. According to (2.24) we
can write then{
ϕi(θ, X),
∫
q(Y )Jµ(Y )dY
}∣∣∣∣
K
= ωαµ(X)Φi
θα
(
S(X)
ε
+ θ,U(X)
)
q(X) +O(ε). (2.39)
Similarly, for any smooth function Q(θ, X), 2π-periodic in each θα, we can write on the basis
of (2.36)∫ 2π
0
· · ·
∫ 2π
0
Q
(
S(X)
ε
+ θ, X
) {
ϕi(θ, X), Jµ(Y )
}∣∣
K
dmθ
(2π)m
= ωαµ(X)
∫ 2π
0
· · ·
∫ 2π
0
Q(θ, X)Φi
θα(θ,U(X))
dmθ
(2π)m
δ(X − Y ) +O(ε). (2.40)
In view of relations (2.30), (2.31) for the functionals Jµ(Y ), Uµ(Y ) we can also write{
ϕi(θ, X),
∫
q(Y )Uµ(Y )dY
}∣∣∣∣
K
= ωαµ(X)Φi
θα
(
S(X)
ε
+ θ,U(X)
)
q(X) +O(ε), (2.41)∫ 2π
0
· · ·
∫ 2π
0
Q
(
S(X)
ε
+ θ, X
) {
ϕi(θ, X), Uµ(Y )
}∣∣
K
dmθ
(2π)m
= ωαµ(X)
∫ 2π
0
· · ·
∫ 2π
0
Q(θ, X)Φi
θα(θ,U(X))
dmθ
(2π)m
δ(X − Y ) +O(ε). (2.42)
In this case, by virtue of (2.25) we have{
ϕi(θ, X),
∫
q(Y )kα(U(Y ))dY
}∣∣∣∣
K
= O(ε), (2.43)
22 A.Ya. Maltsev∫ 2π
0
· · ·
∫ 2π
0
Q
(
S(X)
ε
+ θ, X
) {
ϕi(θ, X), kα(U(Y ))
}∣∣
K
dmθ
(2π)m
= O(ε) (2.44)
for fixed values of coordinates [U(Z)].
Let us note here that since we consider the change of coordinates, which depends explicitly
on ε, we shall assume now, that all relations are written in the coordinate system [U(X),g(θ, X)]
near K. In particular, all the values on K will be functionals of U(X), represented usually in
the form of graded expansions in ε with coefficients depending on U,UX ,UXX , . . . .
Let us prove here some lemmas about the structure of Poisson brackets on the submanifold K
which we will need in the further consideration.
Lemma 2.3. Let the values Uν of the functionals Iν on a complete regular family Λ of m-phase
solutions of (2.3) be functionally independent and give a complete set of parameters on Λ, ex-
cluding the initial phases. Then for the Poisson brackets of the functionals kα(U(X)) and Jµ(Y )
on K we have the following relation{
kα(U(X)), Jµ(Y )
}∣∣
K = ε
[
ωαµ(U(X))δ(X − Y )
]
X
+O
(
ε2
)
. (2.45)
Proof. The conditions of Lemma 2.3 coincide with the conditions of Lemmas 2.1 and 2.2.
Consider the Hamiltonian flow generated by the functional
∫
q(Y )Jµ(Y )dY according to the
bracket (2.27) with a compactly supported function q(Y ) of the slow variable Y = εy.
According to (2.38) we can write at the “points” of the submanifold K
ϕit = q(X)ωβµ(U(X))Φi
θβ
(
S(X)
ε
+ θ,U(X)
)
+ εηi[q]
(
S(X)
ε
+ θ, X
)
+O
(
ε2
)
(2.46)
with some (2π-periodic in each θα) functions ηi[q](θ, X).
Let us introduce the functionals kα(J(X)) = kα(J1(X), . . . , JN (X)) using the same func-
tions kα(U). According to (2.30), (2.31) we can write{
kα(U(X)), Jµ(Y )
}∣∣
K =
{
kα(J(X)), Jµ(Y )
}∣∣
K +O
(
ε2
)
.
Consider the evolution of the functionals kα(J(X)) according to the flow generated by∫
q(Y )Jµ(Y )dY on the submanifold K. Using relations (2.46) we can write
kαt (J(X)) =
∂kα
∂Uν
(J(X)) Jνt (X) =
∂kα
∂Uν
(J(X)) q(X)ωβµ(U(X))
×
∫ 2π
0
· · ·
∫ 2π
0
∑
l≥0
Πν
i(l)
(
Φ
(
S(X)
ε
+ θ,U(X)
)
, ε
∂
∂X
Φ
(
S(X)
ε
+ θ,U(X)
)
, . . .
)
× εl ∂
l
∂X l
Φi
θβ
(
S(X)
ε
+ θ,U(X)
)
dmθ
(2π)m
+ ε
∂kα
∂Uν
(U(X))
(
q(X)ωβµ(U(X))
)
X
×
∫ 2π
0
· · ·
∫ 2π
0
∑
l≥1
lΠν
i(l) (Φ (θ,U(X)) , kγΦθγ (θ,U(X)) , . . . )
× kα1(X) · · · kαl−1(X)Φi
θβθα1 ···θαl−1 (θ,U(X))
dmθ
(2π)m
+ ε
∂kα
∂Uν
(U(X))
∫ 2π
0
· · ·
∫ 2π
0
∑
l≥0
Πν
i(l) (Φ (θ,U(X)) , kγΦθγ (θ,U(X)) , . . . )
× kα1(X) · · · kαl(X)ηi[q]θα1 ···θαl (θ, X)
dmθ
(2π)m
+O
(
ε2
)
.
Whitham’s Method and Dubrovin–Novikov Bracket 23
It’s not difficult to see that the first part of the above expression contains the integrals over θ
of the expressions P ν
θβ
(θ, X) and is equal to zero. It is easy to see also after integration by parts
that the third part of the above expression represents the value
ε
∫ 2π
0
· · ·
∫ 2π
0
∂kα
∂Uν
(U(X)) ζ
(ν)
i[U(X)](θ)ηi[q](θ, X)
dmθ
(2π)m
and is equal to zero according to Lemma 2.1.
Thus, we can write in the main order the expressions for the evolution of the functionals
kα(J(X)) on the submanifold K in the form
kαt (J(X)) = ε
(
q(X)ωβµ(U(X))
)
X
× ∂kα
∂Uν
∫ 2π
0
· · ·
∫ 2π
0
∑
l≥1
lkβ1 · · · kβl−1Φi
θβθβ1 ···θβl−1
Πν
i(l) (Φ, kγΦθγ , . . . )
dmθ
(2π)m
+O
(
ε2
)
.
Using relations (2.13) we get then
kαt (J(X)) = ε [q(X)ωαµ(U(X))]X +O
(
ε2
)
,
i.e.
{kα(J(X)), Jµ(Y )}|K = εωαµ(U(X))δ′(X − Y ) + εωαµX δ(X − Y ) +O
(
ε2
)
,
which implies (2.45). �
Easy to see that according to Lemma 2.3 and relations (2.30), (2.31) we can write also
{kα(U(X)), Uµ(Y )}|K = ε [ωαµ(U(X))δ(X − Y )]X +O
(
ε2
)
(2.47)
on the submanifold K.
Corollary 2.1. Under the conditions of Lemma 2.3, we have also{
kα(X), kβ(Y )
}∣∣
K = O
(
ε2
)
(2.48)
for the functionals k(U(X)).
Indeed, by virtue of (2.25) we have
{
kα(X), kβ(Y )
}∣∣
K =
{
kα(X), Uµ(Y )
}∣∣
K
∂kβ
∂Uµ
(U(Y ))
= εωαµ(X)kβUµ(X)δ′(X − Y ) + εωαµ(X)
(
kβUµ(X)
)
X
δ(X − Y )
+ ε (ωαµ(X))X k
β
Uµ(X)δ(X − Y ) +O
(
ε2
)
= O
(
ε2
)
.
Let us note now that in the proof of Lemma 2.3, we have not used the fact that the function-
als Jµ(Y ) belong to our special set of functionals (2.29) and used only the fact that the flow
generated by the functional Iµ leaves invariant the family Λ generating linear shifts of θα0 with
constant frequencies ωαµ(U). We can therefore formulate here the following lemma.
Lemma 2.3′. Let the values Uν of the functionals Iν on a complete regular family Λ of m-phase
solutions of system (2.3) be functionally independent and give a complete set of parameters on Λ,
excluding the initial phases. Let the flow generated by the functional
Ĩ =
∫
P̃ (ϕ,ϕx, . . . )dx (2.49)
leave invariant the family Λ generating a linear shift of θα0 with constant frequencies ω̃α(U).
24 A.Ya. Maltsev
Consider the functionals
J̃(X) =
∫ 2π
0
· · ·
∫ 2π
0
P̃
(
ϕ, εϕX , ε
2ϕXX , . . .
) dmθ
(2π)m
. (2.50)
Then for the Poisson brackets of the functionals kα(U(X)) and J̃(Y ) on K we have the relation{
kα(U(X)), J̃(Y )
}∣∣
K = ε [ω̃α(U(X))δ(X − Y )]X +O
(
ε2
)
.
Proof of Lemma 2.3′ completely repeats the proof of Lemma 2.3.
Lemma 2.4. Let the values Uν of the functionals Iν on a complete regular family Λ of m-phase
solutions of system (2.3) be functionally independent and give a complete set of parameters on Λ,
excluding the initial phases. Then for the constraints gi(θ, X) imposed by (2.32) and smooth
compactly supported function q(X) as well as smooth 2π-periodic in each θα function Q(θ, X)
we have the following relations on the submanifold K{
gi(θ, X),
∫
q(Y )Jµ(Y )dY
}∣∣∣∣
K
= O(ε),[∫ 2π
0
· · ·
∫ 2π
0
Q
(
S(X)
ε
+ θ, X
){
gi(θ, X), Jµ(Y )
} dmθ
(2π)m
]∣∣∣∣
K
= O(ε). (2.51)
Proof. Indeed, by (2.39), (2.40), and Lemma 2.3 we have{
gi(θ, X),
∫
q(Y )Jµ(Y )dY
}∣∣∣∣
K[0]
= Φi
θα
(
S(X)
ε
+ θ,U(X)
)
ωαµ(X)q(X)
− 1
2
Φi
θα
(
S(X)
ε
+ θ,U(X)
)∫
sgn(X − Z) (ωαµ(Z)q(Z))Z dZ ≡ 0,[∫ 2π
0
· · ·
∫ 2π
0
Q
(
S(X)
ε
+ θ, X
){
gi(θ, X), Jµ(Y )
} dmθ
(2π)m
]∣∣∣∣
K[0]
= ωαµ(X)
∫ 2π
0
· · ·
∫ 2π
0
Q(θ, X)Φi
θα(θ,U(X))
dmθ
(2π)m
δ(X − Y )
−
∫ 2π
0
· · ·
∫ 2π
0
Q(θ, X)Φi
θα(θ,U(X))
dmθ
(2π)m
1
2
sgn(X − Z) [ωαµ(Z)δ(Z − Y )]Z dZ ≡ 0. �
Similarly to the previous case, we can formulate also the following lemma.
Lemma 2.4′. Let the values Uν of the functionals Iν on a complete regular family Λ of m-phase
solutions of system (2.3) be functionally independent and give a complete set of parameters on Λ,
excluding the initial phases. Let the flow generated by the functional (2.49) leave invariant
the family Λ generating a linear shift of θα0 with constant frequencies ω̃α(U). Then for the
constraints gi(θ, X) and the functionals J̃(X) imposed by (2.50) we have the following relations
on the submanifold K{
gi(θ, X),
∫
q(Y )J̃(Y )dY
}∣∣∣∣
K
= O(ε),[∫ 2π
0
· · ·
∫ 2π
0
Q
(
S(X)
ε
+ θ, X
){
gi(θ, X), J̃(Y )
} dmθ
(2π)m
]∣∣∣∣
K
= O(ε).
Let us consider now the Dirac restriction of the bracket (2.27) on the submanifold K.
Whitham’s Method and Dubrovin–Novikov Bracket 25
For the Dirac restriction of the bracket (2.27) on the submanifold K we have to modify the
functionals Uµ(Y ) by linear combinations of the constraints gi(θ, X)
Ũµ(Y ) = Uµ(Y ) +
∫ +∞
−∞
∫ 2π
0
· · ·
∫ 2π
0
gi(θ, X)αµi
(
S[U](X)
ε
+ θ, [U], X, Y
)
dmθ
(2π)m
dX,
so that the flows generated by the functionals Ũµ(Y ) leave K invariant and then use the func-
tionals Ũµ(Y ) to construct the Dirac bracket on K.
Since the functionals Uµ(Y ) are defined with the aid of the functionals J(Z), technically it
is more convenient to modify the functionals Jµ(Y )
J̃µ(Y ) = Jµ(Y ) +
∫ +∞
−∞
∫ 2π
0
· · ·
∫ 2π
0
gi(θ, X)βµi
(
S[U[J]](X)
ε
+ θ, [U[J]], X, Y
)
dmθ
(2π)m
dX,
so that the flows generated by the functionals J̃µ(Y ) leave K invariant and then put
Ũµ(Y ) = J̃µ(Y ) +
∑
l≥1
εlŨµ(l)(Y )
just using the functionals J̃(Y ) instead of J(Y ) in (2.31). Both the approaches give, certainly,
the same result for the Dirac restriction of the bracket (2.27) on K.
The distributions βµi (S(X)/ε+ θ, X, Y ) must satisfy the relation∫ +∞
−∞
∫ 2π
0
· · ·
∫ 2π
0
{
gi(θ, X), gj(θ′, Z)
}∣∣
K β
µ
j
(
S(Z)
ε
+ θ′, Z, Y
)
dmθ′
(2π)m
dZ
+
{
gi(θ, X), Jµ(Y )
}∣∣
K = 0 (2.52)
on K and are defined at each “point” of K up to linear combinations of the distributions
δJζ(W )/δϕi(θ, X) by virtue of the original dependence (2.33) of the constraints (2.32).
To obtain a local Poisson bracket on K we are trying to find the functions βµi (θ, X, Y ) in the
form
βµi (θ, X, Y ) =
∑
l≥1
εlβµi(l)(θ, X, Y ), (2.53)
where the functions βµi(l)(θ, X, Y ) represent the local distributions
βµi(l)(θ, X, Y ) =
l∑
p=0
βµi(l),p(θ, X) δ(p)(X − Y )
with the total grading l in ε, considering that the derivatives of the delta function δ(p)(X − Y )
have by definition the degree p.
Thus, we assume that all the functions βµi(l),p(θ, X) on K are local functionals of (U(X),UX ,
. . . ) at every θ, polynomial in the derivatives (UX , . . . ) and having degree l − p under the
previous definition. Such a structure of βµi (θ, X, Y ) is obviously equivalent to saying that the
functionals
J[q] =
∫ +∞
−∞
Jµ(Y )qµ(Y )dY
can be modified by linear combinations of constraints (2.32) with the coefficients
Bi[q](θ, X) =
∑
l≥1
εlBi[q](l)(θ, X) =
∑
l≥1
εl
l∑
p=0
βµi(l),p(θ, X)
dpqµ(X)
dXp
,
26 A.Ya. Maltsev
so that the corresponding flows leave invariant the submanifold K. Under this scheme, the
derivatives dpqµ/dX
p of slowly varying functions qµ(X) also have grading p as the derivatives
of the functionals U(X).
The pairwise Poisson brackets of the constraints gi(θ, X), gj(θ′, Y ) on K can be written in
the form
{gi(θ, X), gj(θ′, Y )}|K
= {ϕi(θ, X), ϕj(θ′, Y )}|K − {ϕi(θ, X), Uλ(Y )}|K Φj
Uλ
(
S(Y )
ε
+ θ′,U(Y )
)
− Φi
Uν
(
S(X)
ε
+ θ,U(X)
)
{Uν(X), ϕj(θ′, Y )}|K
+ Φi
Uν
(
S(X)
ε
+ θ,U(X)
)
{Uν(X), Uλ(Y )}|K Φj
Uλ
(
S(Y )
ε
+ θ′,U(Y )
)
− 1
2ε
∫
{ϕi(θ, X), kβ(W )}|K sgn(Y −W )dW Φj
θ′β
(
S(Y )
ε
+ θ′,U(Y )
)
− 1
2ε
Φi
θα
(
S(X)
ε
+ θ,U(X)
)∫
sgn(X − Z){kα(Z), ϕj(θ′, Y )}|KdZ
+
1
2ε
Φi
θα
(
S(X)
ε
+ θ,U(X)
)
×
∫
sgn(X − Z){kα(Z), Uλ(Y )}|KdZ Φj
Uλ
(
S(Y )
ε
+ θ′,U(Y )
)
+
1
2ε
Φi
Uν
(
S(X)
ε
+ θ,U(X)
)
×
∫
{Uν(X), kβ(W )}|K sgn(Y −W )dW Φj
θ′β
(
S(Y )
ε
+ θ′,U(Y )
)
+
1
4ε2
Φi
θα
(
S(X)
ε
+ θ,U(X)
)
(2.54)
×
∫
sgn(X − Z){kα(Z), kβ(W )}|K sgn(Y −W )dZdW Φj
θ′β
(
S(Y )
ε
+ θ′,U(Y )
)
.
For the purposes of this chapter, we need only “approximate” restriction of the bracket (2.27)
on the submanifold K. Specifically, we will need to prove the existence of only the first term
βµi(1)(θ, X, Y ) in the expansion (2.53) having the form
βµi(1)(θ, X, Y ) = βµi(1),1(θ,U(X))δ′(X − Y ) + βµi(1),0λ(θ,U(X))UλXδ(X − Y ), (2.55)
which is equivalent to the approximation
Bi[q](1)(θ, X) = βµi(1),1(θ,U(X))qµX(X) + βµi(1),0λ(θ,U(X))UλXqµ(X) (2.56)
for the functions Bi[q](θ, X).
To find it we need only the leading term of the convolution of the Poisson brackets of con-
straints on K with Bj[q](1)(S(Y )/ε+θ′, Y ) and the first (in ε) non-vanishing term of the brackets
of constraints and the functional J[q].
Note that according to the relations (2.43), (2.44), (2.47), and (2.48) all the terms of the
bracket (2.54) behave as the values of the order of at most O(1) for ε → 0 in the convolution
with Bj[q](1)(S(Y )/ε+ θ′, Y ).
We impose additional conditions∫ 2π
0
· · ·
∫ 2π
0
Φj
θα(θ,U(X))βµj(1)(θ, X, Y )
dmθ
(2π)m
≡ 0, α = 1, . . . ,m, (2.57)
Whitham’s Method and Dubrovin–Novikov Bracket 27
or equivalently∫ 2π
0
· · ·
∫ 2π
0
Φj
θα(θ,U(X))Bj[q](1)(θ, X)
dmθ
(2π)m
≡ 0, α = 1, . . . ,m, (2.58)
to be confirmed aposteriori for the functions βµj(1)(θ, X, Y ) and Bj[q](1)(θ, X). The Poisson
brackets of constraints (2.54) on K can then be reduced to the form
{gi(θ, X), gj(θ′, Y )}|K → {gi(θ, X), gj(θ′, Y )}|eff
K
= {ϕi(θ, X), ϕj(θ′, Y )}|K − {ϕi(θ, X), Uλ(Y )}|K Φj
Uλ
(
S(Y )
ε
+ θ′,U(Y )
)
− Φi
Uν
(
S(X)
ε
+ θ,U(X)
)
{Uν(X), ϕj(θ′, Y )}|K
+ Φi
Uν
(
S(X)
ε
+ θ,U(X)
)
{Uν(X), Uλ(Y )}|K Φj
Uλ
(
S(Y )
ε
+ θ′,U(Y )
)
− 1
2ε
Φi
θα
(
S(X)
ε
+ θ,U(X)
)∫
sgn(X − Z){kα(Z), ϕj(θ′, Y )}|KdZ
+
1
2ε
Φi
θα
(
S(X)
ε
+ θ,U(X)
)∫
sgn(X − Z){kα(Z), Uλ(Y )}|KdZ
× Φj
Uλ
(
S(Y )
ε
+ θ′,U(Y )
)
(2.59)
in finding Bi[q](1)(θ, X).
To find the leading term of the convolution of the Poisson brackets of constraints on K and
Bj[q](1)(S(Y )/ε+θ′, Y ) we can simplify further the relation (2.59). Namely, taking into account
relations (2.41), (2.47) we can omit in the leading order the second and the last terms of (2.59)
in the convolution with Bj[q](1)(S(Y )/ε+θ′, Y ). By (2.42) and (2.58), we can omit in the leading
order also the third term of (2.59) in the convolution with Bj[q](1)(S(Y )/ε+ θ′, Y ). The fourth
term of (2.59), obviously, is of the order O(ε) and also can be omitted in the leading order.
As a result, the principal term of the convolution of (2.54) with the functions
εBj[q](1)(S(Y )/ε+ θ′, Y ) can be written in the form[∫
{gi(θ, X), gj(θ′, Y )}|KεBj[q](1)
(
S(Y )
ε
+ θ′, Y
)
dmθ′
(2π)m
dY
]
[1]
≡
∑
s≥0
Bij
(s)
(
Φ
(
S(X)
ε
+ θ,U(X)
)
, kγ(X)Φθγ
(
S(X)
ε
+ θ,U(X)
)
, . . .
)
× kα1(X) · · · kαs(X)Bj[q](1)θα1 ···θαs
(
S(X)
ε
+ θ, X
)
− 1
2
Φi
θα
(
S(X)
ε
+ θ,U(X)
)∫
sgn(X − Z)
×
[∫ {
kα(Z), ϕj(θ′, Y )
}
|KεBj[q](1)
(
S(Y )
ε
+ θ′, Y
)
dmθ′
(2π)m
dY
]
[2]
dZ. (2.60)
Let us remind here that the indexes [1], [2] mean as before the terms of the corresponding
graded expansion on K having degree 1 and 2 respectively.
We have also
{gi(θ, X), J[q]}|K[1] = {ϕi(θ, X), J[q]}|K[1] − Φi
Uν
(
S(X)
ε
+ θ,U(X)
)
{Uν(X), J[q]}|K[1]
28 A.Ya. Maltsev
− 1
2
Φi
θα
(
S(X)
ε
+ θ,U(X)
)∫
sgn(X − Z){kα(Z), J[q]}|K[2]dZ. (2.61)
Let us prove here the following lemma which allows to investigate the further construction of
the functions Bi[q](1)(θ, X).
Lemma 2.5. Let the functions Bi[q](1)(θ, X) satisfy the conditions (2.58), i.e.∫ 2π
0
· · ·
∫ 2π
0
Φj
θα(θ,U(X))Bj[q](1)(θ, X)
dmθ
(2π)m
≡ 0
and the relation
B̂ij
[0](X)Bj[q](1)(θ, X) +Ai[1][q](θ, X) = 0, (2.62)
where
B̂ij
[0](X) (2.63)
=
∑
s≥0
Bij
(s) (Φ(θ,U(X)), kα(X)Φθα(θ,U(X)), . . . ) kα1(X) · · · kαs(X)
∂s
∂θα1 · · · ∂θαs
,
Ai[1][q]
(
S(X)
ε
+ θ, X
)
(2.64)
= {ϕi(θ, X), J[q]}|K[1] − Φi
Uν
(
S(X)
ε
+ θ,U(X)
)
{Uν(X), J[q]}|K[1].
Then the functions Bi[q](1)(θ, X) satisfy the relation[∫
{gi(θ, X), gj(θ′, Z)}|K εBj[q](1)
(
S(Z)
ε
+ θ′, Z
)
dmθ′
(2π)m
dZ
]
[1]
+ {gi(θ, X), J[q]}|K[1] = 0. (2.65)
Proof. The fulfillment of condition (2.58) allows to reduce the expression∫
{gi(θ, X), gj(θ′, Z)}|K εBj[q](1)
(
S(Z)
ε
+ θ′, Z
)
dmθ′
(2π)m
dZ
according to formula (2.60) up to the terms O(ε2). From (2.62) we can get then∫
{gi(θ, X), gj(θ′, Z)}|K εBj[q](1)
(
S(Z)
ε
+ θ′, Z
)
dmθ′
(2π)m
dZ + {gi(θ, X), J[q]}|K
= − ε
2
Φi
θα
(
S(X)
ε
+ θ,U(X)
)∫
sgn(X − Z)
×
[∫
{kα(Z), ϕj(θ′,W )}|K εBj[q](1)
(
S(W )
ε
+ θ′,W
)
dmθ′
(2π)m
dW
]
[2]
dZ
− ε
2
Φi
θα
(
S(X)
ε
+ θ,U(X)
)∫
sgn(X − Z){kα(Z), J[q]}|K[2]dZ +O
(
ε2
)
. (2.66)
The proof of the Lemma follows then from the orthogonality conditions of the left-hand part
of (2.66) to the variational derivatives
δkβ(J(X ′))
δϕi(θ, X)
∣∣∣∣
K
=
∂kβ
∂Uλ
(J(X ′))
δJλ(X ′)
δϕi(θ, X)
∣∣∣∣
K
, β = 1, . . . ,m (2.67)
Whitham’s Method and Dubrovin–Novikov Bracket 29
according to (2.33) in the order ε2 on K. Here we use the functions kβ(U) to define the
functionals kβ(J(X ′)). Let us remind also that the functionals U(X) and J(X) are connected
by the transformations (2.30) and (2.31).
Indeed, the right-hand part of (2.66) has the form
ε
2
Φi
θα
(
S(X)
ε
+ θ,U(X)
)∫
sgn(X − Z)Qα[q](Z)dZ +O
(
ε2
)
(2.68)
with some functions Qα[q](Z).
The convolution with the quantities (2.67) can be represented as the action of the operator∫ 2π
0
· · ·
∫ 2π
0
dmθ
(2π)m
∂kβ
∂Uλ
(
J(X ′)
)
(2.69)
×
∑
l≥0
Πλ
i(l)
(
Φ
(
S(X ′)
ε
+ θ,U(X ′)
)
, ε
∂
∂X ′
Φ
(
S(X ′)
ε
+ θ,U(X ′)
)
, . . .
)
εl
dl
dX ′l
on the functions of S(X ′)/ε+ θ and X ′.
Expanding the action of the operator (2.69) in powers of ε we can write it’s main part in the
form ∫ 2π
0
· · ·
∫ 2π
0
dmθ
(2π)m
∂kβ
∂Uλ
(
U(X ′)
)
×
∑
l≥0
Πλ
i(l) (Φ(. . . ), kαΦθα(. . . ), . . . ) kα1(X ′) · · · kαl(X ′) ∂l
∂θα1 · · · ∂θαl
.
We can see then after the integration by parts that the main part of the operator (2.69)
reduces to the convolution (w.r.t. θ) with the vectors
∂kβ
∂Uλ
(
U(X ′)
)
ζ
(λ)
i[U(X′)]
(
S(X ′)
ε
+ θ
)
,
which are identically equal to zero according to Lemma 2.1.
We can conclude, therefore, that terms of the order of O(ε2) in the right-hand part of (2.66)
do not play any role in the convolution with the values (2.67) in the order ε2.
Let us consider now the action of the operator (2.69) on the terms of the order of ε in the
expression (2.68). We can see that the terms in which the operators ds/dX ′s apply only to the
functions Φθα(S(X ′)/ε+ θ,U(X ′)) yield the quantity
ε
2
∂kβ
∂Uλ
(
J(X ′)
) ∫ 2π
0
· · ·
∫ 2π
0
∂
∂θα
P λ(ϕ, εϕX′ , . . . )
∣∣∣
K
dmθ
(2π)m
∫
sgn(X ′ − Z)Qα[q](Z)dZ
identically equal to zero.
The terms of the order of ε2 occur at the same time with a single differentiation of the function
sgn(X ′−Z) and application of the remaining derivations ∂/∂X ′ to Φθα(S(X ′)/ε+θ,U(X ′)) in
the leading order. In the required approximation the functionals J(X) should be also replaced
by U(X) in the argument of ∂kβ/∂Uλ.
The corresponding values are then equal to
ε2
∂kβ
∂Uλ
(
U(X ′)
)
×
∫ 2π
0
· · ·
∫ 2π
0
∑
l≥1
Πλ
i(l) (Φ, kγΦθγ , . . . ) lk
α1 · · · kαl−1Φi
θα1 ···θαl−1θα
dmθ
(2π)m
Qα[q](X
′).
30 A.Ya. Maltsev
Using again the relations (2.13), we obtain that the orthogonality of the right-hand part
of (2.66) to the values (2.67) in the order ε2 is equivalent to the relations Qβ[q](X
′) ≡ 0. Given
that the main part of (2.68) represents the difference between the left-hand sides of (2.65)
and (2.62) we obtain the assertion of the lemma. �
For a regular Hamiltonian family Λ and a complete Hamiltonian set of integrals (I1, . . . , IN )
we can also prove the following lemma.
Lemma 2.6. Let the functions Bj[q](1)(θ, X) satisfy conditions (2.62). Then the functions
Bj[q](1)(θ, X) automatically satisfy (2.58).
Proof. We first prove the following statement.
The values ζ
(ν)
i[U(X)](θ) are orthogonal (for any X) to the values Ai[1][q](θ, X), i.e.∫ 2π
0
· · ·
∫ 2π
0
ζ
(ν)
i[U(X)](θ)Ai[1][q](θ, X)
dmθ
(2π)m
≡ 0. (2.70)
Indeed, as we know, the values {gi(θ, X), J[q]}|K identically vanish under action of the op-
erator (2.34) (with the replacement of Z to X). In the leading order (∼ ε) it’s action is given
by the convolution w.r.t. θ of the values {gi(θ, X), J[q]}|K[1], imposed by (2.61), with the values
ζ
(ν)
i[U(X)](S(X)/ε + θ). We know also that the values ζ
(ν)
i[U(X)](S(X)/ε + θ) are automatically
orthogonal to the functions Φi
θα(S(X)/ε+ θ,U(X)), so we get the relation (2.70).
Thus, the implementation of (2.62) implies the conditions∫ 2π
0
· · ·
∫ 2π
0
ζ
(ν)
i[U(X)](θ)B̂ij
[0](X)Bj[q](1)(θ, X)
dmθ
(2π)m
= 0,
which is equivalent to
ωαν(U(X))
∫ 2π
0
· · ·
∫ 2π
0
Φj
θα(θ,U(X))Bj[q](1)(θ, X)
dmθ
(2π)m
= 0
view the skew-symmetry of B̂ij
[0](X).
From condition (2) of the definition of a complete Hamiltonian family of commuting func-
tionals we immediately obtain now conditions (2.58). �
It is not difficult to see that from the relation (2.70) follow also the conditions∫ 2π
0
· · ·
∫ 2π
0
κ
(q)
i[U(X)](θ)Ai[1][q](θ, X)
dmθ
(2π)m
≡ 0 (2.71)
by Proposition 2.1.
We come, therefore, to investigation of system (2.62) to construct an “approximate” restric-
tion of the Poisson bracket (2.27) on the submanifold K. In the remaining part of the article the
investigation of system (2.62) will play the basic role for the justification of the main results.
In what follows we consider separately the single-phase (m = 1) and the multiphase (m ≥ 2)
cases.
The following lemma can be formulated for the single-phase case m = 1.
Lemma 2.7. Let Λ be a regular Hamiltonian family of single-phase solutions of (2.3) and
(I1, . . . , IN ) be a complete Hamiltonian set of the first integrals of the form (2.18). Then the
functions Bj[q](1)(θ, X) can be found from system (2.62) and can be written in the form (2.56)
with smooth dependence on parameters U(X).
Whitham’s Method and Dubrovin–Novikov Bracket 31
Proof. System (2.62) in the single-phase case is a system of ordinary differential equations
in θ with skew-symmetric operator B̂ij
[0](X). It is easy to see also that the right-hand side of
system (2.62) has the form
ξiµ (θ,U(X)) qµ,X(X) + ηiµλ (θ,U(X))UλXqµ(X)
with periodic in θ functions ξiµ(θ,U(X)), ηiµλ (θ,U(X)).
The orthogonality conditions (2.71) imply the orthogonality of both the sets of functions
ξiµ(θ,U(X)) and ηiµλ (θ,U(X)) to the functions κ
(q)
i[U(X)](θ), such that system (2.62) can be split
into independent inhomogeneous systems
B̂ij
[0](X)βµj(1),1 (θ,U(X)) = ξiµ (θ,U(X)) , (2.72)
B̂ij
[0](X)βµj(1),0λ (θ,U(X)) = ηiµλ (θ,U(X)) (2.73)
defining functions (2.55).
Both the systems (2.72), (2.73) are systems of ordinary linear differential equations with
periodic coefficients and the skew-symmetric operator B̂ij
[0](X). The zero modes of the opera-
tor B̂ij
[0](X) are given by the variational derivatives of annihilators of the bracket (2.16) rep-
resented as functions of θ on the manifold of single-phase solutions and are orthogonal to the
right-hand parts of (2.72) and (2.73) according to (2.26) and (2.71). Eigenfunctions of B̂ij
[0](X)
form a basis in the space of 2π-periodic functions ϕ(θ). Besides that, the nonzero eigenvalues
of B̂ij
[0](X) are separated from zero in this case. Thus, the 2π-periodic functions βµj(1),1(θ,U(X)),
βµj(1),0λ(θ,U(X)) can be found from systems (2.72), (2.73) up to the variational derivatives of
the annihilators of the bracket (2.16). If we impose additional conditions of orthogonality of
βµj(1)(θ,X, Y ) to the variational derivatives of the annihilators of the bracket (2.16) on the
manifold of single-phase solutions we can suggest a unique procedure of construction of the
functions βµj(1)(θ,X, Y ). The functions βµj(1)(θ,X, Y ) satisfy (2.52) in the order of O(ε) and
conditions (2.57) and smoothly depend on the parameters U(X), which implies the required
properties for the functions Bj[q](1)(θ,X). �
Let us formulate now the theorem that gives the justification of the Dubrovin–Novikov pro-
cedure for a regular Hamiltonian family of single-phase solutions of system (2.3).
Theorem 2.1. Let Λ be a regular Hamiltonian family of single-phase solutions of system (2.3).
Let (I1, . . . , IN ) be a complete Hamiltonian set of commuting first integrals of system (2.3)
on Λ having the form (2.18). Then the Dubrovin–Novikov procedure gives a Poisson bracket of
hydrodynamic type on the space of parameters {Uν(X), ν = 1, . . . , N}, where Uν ≡ 〈P ν〉.
Proof. The antisymmetry of the Dubrovin–Novikov bracket follows immediately from the an-
tisymmetry of the original bracket. Let us prove the existence of the Jacobi identity under the
conditions of the theorem.
Consider the Poisson bracket (2.27) and the functionals
J̃ν(1)(X) = Jν(X) + ε
∫∫ 2π
0
gj(θ, Z)βνj(1)
(
S(Z)
ε
+ θ, Z,X
)
dθ
2π
dZ
in the neighborhood of the submanifold K. (Note that the functions βνj(1) depending on the
values U(X) on K, can be considered as functionals in the neighborhood of K by the definition
of the functionals U(X)).
The pairwise brackets of the functionals J̃ν(1)(X), J̃µ(1)(Y ) can be written as{
J̃ν(1)(X), J̃µ(1)(Y )
}
=
{
Jν(X), Jµ(Y )
}
32 A.Ya. Maltsev
+ ε
∫∫ 2π
0
{
βνi(1)
(
S(Z)
ε
+ θ, Z,X
)
gi(θ, Z), Jµ(Y )
}
dθ
2π
dZ
+ ε
∫∫ 2π
0
{
Jν(X), gj(θ,W )βµj(1)
(
S(W )
ε
+ θ,W, Y
)}
dθ
2π
dW (2.74)
+ ε2
∫ {
βνi(1)
(
S(Z)
ε
+ θ, Z,X
)
gi(θ, Z), gj(θ′,W )βµj(1)
(
S(W )
ε
+ θ′,W, Y
)}
dθ
2π
dθ′
2π
dZdW.
We note here that despite the presence of non-local terms in the brackets of constraints
{gi(θ, Z), gj(θ′,W )}, and {Jν(X), gj(θ,W )}, {gi(θ, Z), Jµ(Y )}, the bracket of the functionals
{J̃ν(1)(X), J̃µ(1)(Y )} on K has a purely local form because of the conditions (2.57). We can also
see here that expression (2.74) is a regular at ε→ 0 distribution after the restriction on K.
From the form of the constraints and condition (2.57) it is easy to obtain that the pairwise
Poisson brackets of the functionals J̃ν(1)(X), J̃µ(1)(Y ) on K can be written as
{
J̃ν(1)(X), J̃µ(1)(Y )
}∣∣
K =
∑
l≥1
εl
l∑
q=0
aνµ[l−q](U,UX , . . . )δ
(q)(X − Y ), (2.75)
where aνµ[l−q] are some functions of U,UX , . . . , polynomial in the derivatives and having deg-
ree l − q. By virtue of relations (2.51) we can state also that the leading term in ε in the
expression (2.75) is still the same Dubrovin–Novikov bracket on K, since the corrections given
by the last three terms of (2.74) are of the order of O(ε2) on K. We thus have
{
J̃ν(1)(X), J̃µ(1)(Y )
}∣∣
K = ε {Uν(X), Uµ(Y )}DN +
∑
l≥2
εl
l∑
q=0
aνµ[l−q](U,UX , . . . )δ
(q)(X − Y ),
where
{Uν(X), Uµ(Y )}DN = 〈Aνµ1 〉 (U(X)) δ′(X − Y ) +
∂〈Qνµ〉
∂Uγ
UγXδ(X − Y )
is the Dubrovin–Novikov bracket on the space of fields U(X). Thus, we can write
δ
δUγ(W )
{
J̃ν(1)(X), J̃µ(1)(Y )
}∣∣
K = ε
δ {Uν(X), Uµ(Y )}DN
δUγ(W )
+
δ
δUγ(W )
∑
l≥2
εl
l∑
q=0
aνµ[l−q](U,UX , . . . )δ
(q)(X − Y ) (2.76)
on the submanifold K.
For further discussion it is convenient to introduce “regularized” functionals. Namely, for any
smooth compactly supported vector-valued function q(X) = (q1(X), . . . , qN (X)) we introduce
the functionals
J[q] =
∫
qν(X)Jν(X)dX, J̃(1)[q] =
∫
qν(X)J̃ν(1)(X)dX, U[q] =
∫
qν(X)Uν(X)dX.
We can write then{
J̃(1)[q], J̃(1)[p]
}
=
{
J[q], J[p]
}
+ ε
∫ +∞
−∞
∫ 2π
0
{
Bi[q](1)
(
S(Z)
ε
+ θ, Z
)
gi(θ, Z), J[p]
}
dθ
2π
dZ
+ ε
∫ +∞
−∞
∫ 2π
0
{
J[q], g
j(θ,W )Bj[p](1)
(
S(W )
ε
+ θ,W
)}
dθ
2π
dW (2.77)
Whitham’s Method and Dubrovin–Novikov Bracket 33
+ ε2
∫ {
Bi[q](1)
(
S(Z)
ε
+ θ, Z
)
gi(θ, Z), gj(θ′,W )Bj[p](1)
(
S(W )
ε
+ θ′,W
)}
dθ
2π
dθ′
2π
dZdW.
We can see again that the last three terms of (2.77) give corrections of the order of ε2 on the
submanifold K.
For the expansion of the brackets (2.77) in the neighborhood of the submanifold K it is
convenient to use the expressions
ϕi(θ,X) = gi(θ,X) + Φi
(
S(X)
ε
+ θ,U(X)
)
(2.78)
making it possible to easily expand any functionals of ϕ(θ,X) with respect to the values of g.
For example, using the expression (2.20), we can write for the brackets {J[q], J[p]} in the neigh-
borhood of K{
J[q], J[p]
}
=
{
J[q], J[p]
}∣∣
K
+
∫∫ 2π
0
∑
s≥0
[
δ
δϕj(θ,W )
∫∫ 2π
0
εsAνµs
(
ϕ(θ′, X), εϕX(θ′, X), . . .
)
qν(X)pµ,sX(X)
dθ′
2π
dX
]∣∣∣∣
K
× gj(θ,W )
dθ
2π
dW +O
(
g2
)
,
where the values {J[q], J[p]}|K as well as all the other values on K are calculated on the functions
ϕ̃i(θ,W ) ∈ K corresponding to the same values of the functionals [U(Z)] as the original functions
ϕi(θ,W ). We can see that the first term of the expansion of {J[q], J[p]} near K also has the
order O(ε) as well as the values {J[q], J[p]} on K view the condition (2.35).
Considering relations (2.36), we can write a similar representation for the remaining terms
of the expression (2.77).
We can see that the last term of the right-hand side of (2.77) has the order O(ε2) in the
whole space which gives the same order in ε for the terms of its expansion after the substitution
of (2.78).
By Lemma 2.4, the second and the third terms of (2.77) are of the order O(ε2) on the
submanifold K. However, this can not be said about the next terms of the expansion of these
brackets in the values of g. To find the values of these terms near K we write e.g.∫∫ 2π
0
Bi[q](1)
(
S(Z)
ε
+ θ, Z
)
gi(θ, Z)
dθ
2π
dZ
=
∫∫ 2π
0
Bi[q](1)
(
S(Z)
ε
+ θ, Z
)
ϕi(θ, Z)
dθ
2π
dZ −
∫∫ 2π
0
Bi[q](1)(θ, Z)Φi(θ,U(Z))
dθ
2π
dZ.
After rather simple calculations it’s not difficult to get then for the second term of (2.77)
near K∫∫ 2π
0
{
Bi[q](1)
(
S(Z)
ε
+ θ, Z
)
gi(θ, Z), J[p]
}
dθ
2π
dZ
=
∫∫ 2π
0
{
Bi[q](1)
(
S(Z)
ε
+ θ, Z
)
gi(θ, Z), J[p]
}∣∣∣∣
K
dθ
2π
dZ
+
∫ [∫
Bi[q](1)
(
S(Z)
ε
+ θ, Z
) δCiµ(0)(ϕ(θ, Z), εϕZ(θ, Z), . . . )
δϕj(θ′,W )
∣∣∣∣∣
K
pµ(Z)
dθ
2π
dZ +O(ε)
]
× gj(θ′,W )
dθ′
2π
dW
34 A.Ya. Maltsev
+
∫ [∫
Bi[q](1)θ
(
S(Z)
ε
+ θ, Z
)
1
2ε
sgn(Z −W )
{
k(W ), J[p]
}∣∣
K dW +O(ε)
]
× gi(θ, Z)
dθ
2π
dZ +O
(
g2
)
=
∫∫ 2π
0
{
Bi[q](1)
(
S(Z)
ε
+ θ, Z
)
gi(θ, Z), J[p]
}∣∣∣∣
K
dθ
2π
dZ
+
∫ [∫
Bj[q](1)
(
S(Z)
ε
+ θ′, Z
)
δSjµ(ϕ(θ′, Z), εϕZ(θ′, Z), . . . )
δϕi(θ,W )
∣∣∣∣
K
pµ(Z)
dθ′
2π
dZ +O(ε)
]
× gi(θ,W )
dθ
2π
dW
+
∫ [∫
Bi[q](1)θ
(
S(Z)
ε
+ θ, Z
)
ωµ(Z)pµ(Z) +O(ε)
]
gi(θ, Z)
dθ
2π
dZ +O
(
g2
)
.
In general, considering the expansion of the brackets (2.77) in the neighborhood of K, we can
write the relations{
J̃(1)[q], J̃(1)[p]
}
=
{
J̃(1)[q], J̃(1)[p]
}∣∣
K
+ ε
∫∫ 2π
0
T̃i[q,p]
(
S(W )
ε
+ θ,W, ε
)
gi(θ,W )
dθ
2π
dW +O
(
g2
)
(2.79)
with some regular at ε→ 0, 2π-periodic in θ, functions T̃i[q,p](θ,W, ε).
Now for the bracket (2.27) we consider at the “points” of K the Jacobi identity of the form{{
J̃(1)[q], J̃(1)[p]
}
, J̃(1)[r]
}
+ c.p. ≡ 0 (2.80)
with some smooth compactly supported q(X), p(X), r(X).
We have on K the relations{
gi(θ,W ), J̃(1)[r]
}∣∣
K = O
(
ε2
)
for the bracket of the functionals gi(θ,W ) and J̃(1)[r].
It’s not difficult to get also{
Uγ(W ), J̃(1)[r]
}∣∣
K =
{
Jγ(W ), J̃(1)[r]
}∣∣
K +O
(
ε2
)
=
{
Jγ(W ), J[r]
}∣∣
K +O
(
ε2
)
= ε
{
Uγ(W ), U[r]
}
DN
+O
(
ε2
)
.
Now using relations (2.76) and expansion (2.79), we can see that the Jacobi identities (2.80)
on K coincide in the leading order in ε (ε2) with the similar Jacobi identities for the Dubrovin–
Novikov bracket∫
δ{U[q], U[p]}DN
δUγ(W )
{Uγ(W ), U[r]}DN dW + c.p. ≡ 0
on the space of fields U(X). �
Let us formulate here also the Theorems justifying the invariance of the Dubrovin–Novikov
procedure and the Hamiltonian properties of the Whitham system (2.10) with respect to the
averaged bracket in the single-phase case.
Theorem 2.2. Let Λ be a regular Hamiltonian family of single-phase solutions of (2.3). Let
(I1, . . . , IN ) and (I ′1, . . . , I ′N ) be two different complete Hamiltonian sets of commuting first
integrals of (2.3) having the form (2.18). Then the Dubrovin–Novikov brackets obtained using
the sets (I1, . . . , IN ) and (I ′1, . . . , I ′N ) coincide with each other.
Whitham’s Method and Dubrovin–Novikov Bracket 35
Theorem 2.3. Let Λ be a regular Hamiltonian family of single-phase solutions of (2.3). Let
(I1, . . . , IN ) be a complete Hamiltonian set of commuting first integrals of (2.3) having the
form (2.18) and H be the Hamiltonian function for the system (2.3) having the form (2.17).
Then the Whitham system (2.10) is Hamiltonian with respect to the corresponding Dubrovin–
Novikov bracket (2.21) with the Hamiltonian function (2.22)
Hav =
∫ +∞
−∞
〈PH〉 (U(X)) dX.
Proofs of Theorems 2.2 and 2.3 are identical for the single-phase and the multiphase cases,
so we prove them in general m-phase case in the next chapter.
3 Dubrovin–Novikov bracket and the multiphase case
We turn now to the general multiphase situation. We note that the difference with the single-
phase case occur only in Theorem 2.1, since finding the functions Bj[q](1)(θ, X) here is a more
complicated problem. Let us return to the system (2.62), i.e.
B̂ij
[0](X)Bj[q](1)(θ, X) +Ai[1][q](θ, X) = 0, (3.1)
where the operator B̂ij
[0](X) and the functions Ai[1][q](θ, X) are given by (2.63) and (2.64) respec-
tively. It is easy to see that the operator B̂ij
[0](X) is a differential operator on the torus, which acts
along a constant direction kα∂/∂θα. The corresponding vector field generates one-dimensional
trajectories which are everywhere dense in the torus Tm for generic vectors k(X). For special
values of k(X) the closures of the trajectories may be lower-dimensional tori, in particular, the
closed one-dimensional trajectories in Tm. Note that the operator B̂ij
[0](X) ≡ B̂ij
[0](U(X)) is
exactly the Hamiltonian operator for the bracket (2.16) on the manifold of m-phase solutions Λ.
The operator B̂ij
[0](U) has in general finite number of “regular” eigenvectors with zero eigen-
values defined for all values of the parameters U and smoothly depending on the parameters.
However, for special values of U the set of eigenvectors with zero eigenvalues is infinite and
determined by the dimension of closures of the straight-line trajectories in Tm, defined by the
vector k(U).3
Let us define in the space of the parameters U the set M, such that for all U ∈M the tra-
jectories of the vector field (k1(U), . . . , km(U)) are everywhere dense in Tm. From the condition
rank ||∂kα/∂Uν || = m
it follows that the set M is everywhere dense in the parameter space U and, moreover, has the
full measure.
3For some special brackets (2.16) the differential part can be absent in the operator B̂ij[0](U). The opera-
tor B̂ij[0](U) reduces then to an ultralocal operator acting independently at every point of Tm. As a rule, the
matrix Bij[0](U) is non-degenerate in this case. For example, for the ultralocal Poisson bracket
{ψ(x), ψ̄(y)} = iδ(x− y)
for the NLS equation
iψt = ψxx + ν|ψ|2ψ
we have exactly this situation. Easy to see that system (3.1) represents a simple algebraic system in this case and
is trivially solvable. The multiphase situation is not different here from the single-phase one, so all the arguments
of the previous chapter can be used also for the multiphase case. However, for arbitrary brackets (2.16) the
operators B̂ij[0](U) have more general form described above.
36 A.Ya. Maltsev
In the study of the solubility of (3.1) we must first require the orthogonality of the functions
Ai[1][q](θ, X) to the “regular” eigenvectors of B̂ij
[0](X) with zero eigenvalues. Let us prove here
the following lemma.
Lemma 3.1. Let Λ be a regular Hamiltonian family of m-phase solutions of (2.3) and (I1, . . . ,
IN ) be a complete Hamiltonian set of commuting first integrals of (2.3) having the form (2.18).
Let for U ∈M
v
(l)
[U](θ) =
(
v
(l)
1[U](θ), . . . , v
(l)
n[U](θ)
)
, l = 1, . . . , s
be the complete set of linearly independent eigenvectors of the operator B̂ij
[0](U) on the torus with
zero eigenvalues, smoothly depending on θ. Then
1) the number of the vectors v
(l)
[U](θ) is equal to the number of annihilators of the bracket (2.16)
on the submanifold of m-phase solutions of (2.3);
2) the functions Ai[1][q](θ, X) are orthogonal to all the vectors v
(l)
[U(X)](θ), i.e.∫ 2π
0
· · ·
∫ 2π
0
v
(l)
i[U(X)](θ)Ai[1][q](θ, X)
dmθ
(2π)m
≡ 0.
Proof. Consider the values of v
(l)
[U](θ) on any of the trajectories of the vector field kα∂/∂θα on
the torus Tm. According to the definition of regular Hamiltonian family of m-phase solutions
of (2.3) the corresponding functions v
(l)
i[U](kx+θ0) should be the variational derivatives of some
linear combination of annihilators of the bracket (2.16). We have then for a fixed value of θ0
v
(l)
i[U](kx+ θ0) =
∑
p
αlp(U,θ0)
δNp
δϕi(x)
∣∣∣∣
ϕ=Φ(kx+θ0,U)
.
From relation (2.26) we have then
v
(l)
i[U](k(U)x+ θ0) =
∑
p,q
αlp(U,θ0)npq(U)κ
(q)
i[U] (k(U)x+ θ0) . (3.2)
By definition, for U ∈M the trajectories of the field kα∂/∂θα are everywhere dense in Tm.
Since both the left- and the right-hand parts of (3.2) are smooth functions on Tm, we get then
that they coincide on Tm. We can put then αlp(U,θ0) = αlp(U) and write
v
(l)
i[U](θ) ≡
∑
p,q
αlp(U)npq(U)κ
(q)
i[U] (θ) . (3.3)
It is easy to see also that any linear combination of the form (3.3) gives a regular eigenvector
of the operator B̂ij
[0](U(X)) with zero eigenvalue.
Statement (2) follows then from relation (2.71) in view of representation (3.3). �
However, despite the presence of Lemma 3.1, study of system (3.1) is much more complicated
in the multiphase case if compared with the single-phase case. Thus, the presence of “resonances”
for some values of the parameters U(X) may lead to insolubility of (3.1) in the space of periodic
(in all θα) functions. The set of the “resonance” values of U, as a rule, has measure zero.
Nevertheless, the possibility of resonant parameters do not allow to transfer directly the methods
of the previous chapter to the multiphase case. The following theorem shows, however, that the
procedure of the bracket averaging is in fact insensitive to the appearance of the resonant values
of U and can be used in most multiphase cases, as well as in the single-phase case.
Whitham’s Method and Dubrovin–Novikov Bracket 37
Theorem 3.1. Let the system (2.3) be a local Hamiltonian system generated by the func-
tional (2.17) in the local field-theoretic Hamiltonian structure (2.16). Let Λ be a regular Hamil-
tonian family of m-phase solutions of (2.3) and (I1, . . . , IN ) be a complete Hamiltonian set of
commuting integrals (2.18) for this family.
Let the parameter space U of the family Λ have a dense set S ⊂M on which the system (3.1),
or, equivalently, the multiphase system (2.72), (2.73) is solvable in the space of smooth 2π-
periodic in each θα functions. Then the Dubrovin–Novikov bracket, obtained with the aid of the
functionals (I1, . . . , IN ), satisfies the Jacobi identity.
Proof. As before, for a smooth compactly supported vector-valued function q(X) = (q1(X),
. . . , qN (X)) we define the functional
J[q] =
∫ +∞
−∞
Jν(X)qν(X)dX.
Then, for arbitrary smooth, compactly supported in X and 2π-periodic in each θα functions
Q̃(θ, X) = (Q̃1(θ, X), . . . , Q̃n(θ, X)) we define the functionals
Qi(θ, X) = Q̃i(θ, X)
− Φi
θβ (θ,U(X))Mβγ(U(X))
∫ 2π
0
· · ·
∫ 2π
0
Q̃j(θ
′, X)Φj
θ′γ (θ′,U(X))
dmθ′
(2π)m
,
where the matrix Mβγ(U) is the inverse of the matrix∫ 2π
0
· · ·
∫ 2π
0
n∑
i=1
Φi
θβ (θ,U)Φi
θγ (θ,U)
dmθ
(2π)m
,
which is always defined according to the definition of a complete regular family of m-phase
solutions of system (2.3).
By definition, the functions Qi(θ, X) are local functionals of U(X)
Qi(θ, X) ≡ Qi(θ, X,U(X))
depending also on the arbitrary fixed functions Q̃(θ, X). Everywhere below we will assume that
Q(θ, X) is a functional of this type defined with some function Q̃(θ, X).
Easy to see that the values of Qi(θ, X) with arbitrary Q̃(θ, X) represent for fixed values of
the functionals U(Z) all possible smooth, compactly supported in X and 2π-periodic in each θα
functions with the only restriction∫ 2π
0
· · ·
∫ 2π
0
Qi(θ, X)Φi
θα(θ,U(X))
dmθ
(2π)m
= 0, ∀X,α = 1, . . . ,m. (3.4)
For the functionals Qi(θ, X) we define the functionals
g[Q] =
∫ +∞
−∞
∫ 2π
0
· · ·
∫ 2π
0
gi(θ, X)Qi
(
S(X)
ε
+ θ, X
)
dmθ
(2π)m
dX.
Now, for fixed functions q(X), p(X), and functional Q(θ, X) consider the Jacobi identity of
the form{
g[Q],
{
J[q], J[p]
}}
+
{
J[p],
{
g[Q], J[q]
}}
+
{
J[q],
{
J[p], g[Q]
}}
≡ 0. (3.5)
38 A.Ya. Maltsev
Expanding the values of the brackets {Jν(X), Jµ(Y )} in the neighborhood of the submani-
fold K, as in the single-phase case, we can write
{Jν(X), Jµ(Y )} = {Jν(X), Jµ(Y )}|K
+
∑
s≥0
[
δ
δϕk(θ,W )
∫ 2π
0
· · ·
∫ 2π
0
Aνµs
(
ϕ(θ′, X), εϕX(θ′, X), . . .
) dmθ′
(2π)m
]∣∣∣∣
K
× gk(θ,W )
dmθ
(2π)m
dWεsδ(s)(X − Y ) +O
(
g2
)
,
where all the values on the submanifold K are calculated at the same values of the functionals
[U(Z)] as for the original function ϕ(θ, X).
Let us introduce by definition
δ{J[q], J[p]}
δgk(θ,W )
∣∣∣∣
K
(3.6)
≡
∑
s≥0
∫ [
δ
δϕk(θ,W )
∫ 2π
0
· · ·
∫ 2π
0
εsAνµs
(
ϕ(θ′, X), εϕX(θ′, X), . . .
) dmθ′
(2π)m
]∣∣∣∣∣
K
qν(X)pµ,sXdX.
Note that the notations δ/δgk(θ,W ), generally speaking, are not natural in our situation
because of the dependence of the chosen system of constraints. Nevertheless, the preservation
of these notations can better clarify the algebraic structure of the further calculations.
The values defined by (3.6) can be represented in the form of the graded decompositions at
ε→ 0 on the submanifold K. By virtue of (2.35) it is easy to conclude that the expansion in ε
of the quantities (3.6) begins with the first degree in ε
δ{J[q], J[p]}
δgk(θ,W )
∣∣∣∣
K
= ε
δ{J[q], J[p]}
δgk(θ,W )
∣∣∣∣
K[1]
+ ε2
δ{J[q], J[p]}
δgk(θ,W )
∣∣∣∣
K[2]
+ · · · . (3.7)
The leading term of (3.7) can be divided into two parts, corresponding to the functions
εQνµX (ϕ, εϕX , . . . ) and Aνµ1 (ϕ, εϕX , . . . ) and containing the quantities qν(W )pµ(W ) and
qν(W )pµ,W (W ) as local factors, respectively.
The values of {J[q], J[p]}, are obviously invariant under transformations of the form
ϕ(θ, X)→ ϕ(θ + ∆θ, X). (3.8)
From the invariance of the functionals U(X) in such transformations, we can write for the
corresponding increments of the constraints (2.32)
δgk(θ, X) = ϕk(θ + δθ, X)− ϕk(θ, X).
As a consequence, we can write on the submanifold K∫∫ 2π
0
· · ·
∫ 2π
0
δ{J[q], J[p]}
δgk(θ,W )
∣∣∣∣
K
Φk
θα
(
S(W )
ε
+ θ,U(W )
)
dmθ
(2π)m
dW ≡ 0, (3.9)
α = 1, . . . ,m.
The above relation is satisfied to all orders in ε. By the arbitrariness of the functions qν(X),
the relation (3.9) in the leading order can be strengthened. Namely, according to the remark
about the form of the leading term of (3.7) we can write for any W∫ 2π
0
· · ·
∫ 2π
0
δ{J[q], J[p]}
δgk(θ,W )
∣∣∣∣
K[1]
Φk
θα
(
S(W )
ε
+ θ,U(W )
)
dmθ
(2π)m
≡ 0, α = 1, . . . ,m. (3.10)
Whitham’s Method and Dubrovin–Novikov Bracket 39
Quite similarly, we have the relation{
g[Q], J[q]
}
=
∫
Qi
(
S(X)
ε
+ θ, X
){
gi(θ, X), J[q]
} dmθ
(2π)m
dX
+
∫
gi(θ, X)Qi,θα
(
S(X)
ε
+ θ, X
)
1
2ε
sgn(X − Y )
{
kα(Y ), J[q]
} dmθ
(2π)m
dY dX
+
∫
gi(θ, X)Qi,Uν
(
S(X)
ε
+ θ, X
){
Uν(X), J[q]
} dmθ
(2π)m
dX
=
∫
Qi
(
S(X)
ε
+ θ, X
)∑
l≥0
εlCiµ(l) (ϕ(θ, X), εϕX(θ, X), . . . )
dmθ
(2π)m
qµ,lX(X)dX
−
∫
Qi
(
S(X)
ε
+ θ, X
)
Φi
Uν
(
S(X)
ε
+ θ,U(X)
){
Uν(X), J[q]
} dmθ
(2π)m
dX
−
∫
Qi
(
S(X)
ε
+ θ, X
)
Φi
θα
(
S(X)
ε
+ θ,U(X)
)
× 1
2ε
sgn(X − Y )
{
kα(Y ), J[q]
} dmθ
(2π)m
dY dX
+
∫
gi(θ, X)Qi,θα
(
S(X)
ε
+ θ, X
)
1
2ε
sgn(X − Y )
{
kα(Y ), J[q]
} dmθ
(2π)m
dY dX
+
∫
gi(θ, X)Qi,Uν
(
S(X)
ε
+ θ, X
){
Uν(X), J[q]
} dmθ
(2π)m
dX.
According to relations (3.4), we can actually see here that the third term in the above
expression is identically equal to zero.
According to the form of the constraints, we have again in the neighborhood of K
{g[Q], J[q]} = {g[Q], J[q]}
∣∣
K
+
∫ Qi
(
S(X)
ε
+ θ′, X
)∑
l≥0
εl
δCiµ(l)(ϕ(θ′, X), εϕX(θ′, X), . . . )
δϕk(θ,W )
dmθ′
(2π)m
qµ,lX(X)dX
∣∣∣∣∣∣
K
× gk(θ,W )
dmθ
(2π)m
dW
−
∫
Qi(θ
′, X)Φi
Uν
(
θ′,U(X)
) dmθ′
(2π)m
δ{Uν(X), J[q]}
δϕk(θ,W )
∣∣∣∣
K
dX × gk(θ,W )
dmθ
(2π)m
dW
+
∫
Qk,θα
(
S(W )
ε
+ θ,W
)
1
2ε
sgn(W − Y )
{
kα(Y ), J[q]
}∣∣
K dY × gk(θ,W )
dmθ
(2π)m
dW
+
∫
Qi,Uν
(
S(W )
ε
+ θ,W
) {
Uν(W ), J[q]
}∣∣
K × g
k(θ,W )
dmθ
(2π)m
dW +O
(
g2
)
provided that all the values on the submanifold K are calculated at the same values of the
functionals [U(Z)].
For the quantities {g[Q], J[q]} we can introduce by definition
δ{g[Q], J[q]}
δgk(θ,W )
∣∣∣∣
K
≡
∫
Qi
(
S(X)
ε
+ θ′, X
)∑
l≥0
εl
δCiµ(l)(ϕ(θ′, X), εϕX(θ′, X), . . . )
δϕk(θ,W )
qµ,lX(X)
∣∣∣∣∣∣
K
dmθ′
(2π)m
dX
−
∫
Qi(θ
′, X)Φi
Uν
(
θ′,U(X)
) dmθ′
(2π)m
δ{Uν(X), J[q]}
δϕk(θ,W )
∣∣∣∣
K
dX
40 A.Ya. Maltsev
+Qk,θα
(
S(W )
ε
+ θ,W
)
1
2ε
∫
sgn(W − Y )
{
kα(Y ), J[q]
}∣∣
K dY
+Qi,Uν
(
S(W )
ε
+ θ,W
) {
Uν(W ), J[q]
}∣∣
K . (3.11)
The quantities δ{g[Q], J[q]}/δgk(θ,W )|K have the order O(1) at ε→ 0. We have also
Ciµ(0)
(
ϕ(θ′, X), εϕX(θ′, X), . . .
)
≡ Siµ
(
ϕ(θ′, X), εϕX(θ′, X), . . .
)
according to relation (2.37).
Let us introduce the functions
Siµk(l)(ϕ,ϕx, . . . ) ≡
∂Siµ(ϕ,ϕx, . . . )
∂ϕklx
.
Using now Lemma 2.3 and relations (3.7) we can write for the leading term of (3.11)
δ{g[Q], J[q]}
δgk(θ,W )
∣∣∣∣
K[0]
=
∑
l≥0
(−1)lkα1(W ) · · · kαl(W )
∂l
∂θα1 · · · ∂θαl
×
[
Qi
(
S(W )
ε
+ θ,W
)
Siµk(l) (ϕ(θ,W ), εϕW (θ,W ), . . . )
∣∣∣
K[0]
]
qµ(W )
+ ωαµ(W )qµ(W )Qk,θα
(
S(W )
ε
+ θ,W
)
, (3.12)
where
Siµk(l) (ϕ(θ,W ), εϕW (θ,W ), . . . )
∣∣∣
K[0]
≡ Siµk(l)
(
Φ
(
S(W )
ε
+ θ,U(W )
)
, kγΦθγ
(
S(W )
ε
+ θ,U(W )
)
, . . .
)
.
Note one more property of the values δ{g[Q], J[q]}/δgk(θ,W )|K[0]. As we saw earlier, the
values {g[Q], J[q]} are of the order O(ε) at ε → 0 on the submanifold K. This property is
preserved also under the overall shift of the initial phase (3.8).
Indeed, for
ϕi(θ, X) = Φi
(
S(X)
ε
+ θ + ∆θ,U(X)
)
(3.13)
we can write{
g[Q], J[q]
}
=
∫
Qi,θα
(
S(Z)
ε
+ θ, Z
)
1
2ε
sgn(Z −W )
×
{
kα(W ), J[q]
}
Φi
(
S(Z)
ε
+ θ + ∆θ,U(Z)
)
dmθ
(2π)m
dZdW
+
∫
Qi
(
S(Z)
ε
+ θ, Z
)
Ciµ(0)
(
Φ
(
S(Z)
ε
+ θ + ∆θ,U(Z)
)
, kγΦθγ
(
S(Z)
ε
+ θ + ∆θ,U(Z)
)
, . . .
)
× qµ(Z)
dmθ
(2π)m
dZ +O(ε).
Because of the invariance under translations (3.8) the value of the bracket {kα(W ), J[q]} on
the functions (3.13) is equal to its value on K{
kα(W ), J[q]
}
= ε (ωαν(W )qν(W ))W +O
(
ε2
)
.
Whitham’s Method and Dubrovin–Novikov Bracket 41
Similarly, we have on the functions (3.13)
Ciµ(0)
(
Φ
(
S(Z)
ε
+ θ + ∆θ,U(Z)
)
, kγΦθγ
(
S(Z)
ε
+ θ + ∆θ,U(Z)
)
. . .
)
= ωαµ(Z)Φi
θα
(
S(Z)
ε
+ θ + ∆θ,U(Z)
)
.
We then obtain{
g[Q], J[q]
}
=
∫
qµ(Z)ωαµ(Z)
∂
∂θα
[
Qi(θ, Z)Φi(θ + ∆θ,U(Z))
] dmθ
(2π)m
dZ +O(ε) = O(ε).
As a consequence, we can write∫∫ 2π
0
· · ·
∫ 2π
0
δ{g[Q], J[q]}
δgk(θ,W )
∣∣∣∣
K[0]
Φk
θα
(
S(W )
ε
+ θ,U(W )
)
dmθ
(2π)m
dW ≡ 0, (3.14)
α = 1, . . . ,m
for the main part of δ{g[Q], J[q]}/δgk(θ,W ) on K.
Using again the fact that the relation (3.14) contains arbitrary functions qµ(W ), appearing
in the integrand expression in the form of local factors, we can rewrite (3.14) in a stronger form∫ 2π
0
· · ·
∫ 2π
0
δ{g[Q], J[q]}
δgk(θ,W )
∣∣∣∣
K[0]
Φk
θα
(
S(W )
ε
+ θ,U(W )
)
dmθ
(2π)m
≡ 0, (3.15)
∀W, α = 1, . . . ,m.
We now turn back to the Jacobi identity (3.5) for the functionals g[Q], J[q], and J[p]. It is not
difficult to see that after the restriction on K the leading term (∼ε) of the relation (3.5) can be
written as∫ {
g[Q], g
k(θ,W )
}∣∣
K[0]
δ{J[q], J[p]}
δgk(θ,W )
∣∣∣∣
K[1]
dmθ
(2π)m
dW
+
∫ {
J[p], g
k(θ,W )
}∣∣
K[1]
δ{g[Q], J[q]}
δgk(θ,W )
∣∣∣∣
K[0]
dmθ
(2π)m
dW
−
∫ {
J[q], g
k(θ,W )
}∣∣
K[1]
δ{g[Q], J[p]}
δgk(θ,W )
∣∣∣∣
K[0]
dmθ
(2π)m
dW ≡ 0.
The above identity can again be written in a stronger form. Namely, making the change
Q̃i(θ,W ) → Q̃i(θ,W )µi(W ), we get the corresponding change, Qi(θ,W ) → Qi(θ,W )µi(W ),
where µi(W ) are arbitrary smooth functions of W . By virtue of (3.4), (3.10), and (3.12) it is
easy to see then that the integrands are smooth functions of θ and W , containing µi(W ) in the
form of local factors. By the arbitrariness of µi(W ), we can omit the integration over W in the
above integrals and write for every W∫ 2π
0
· · ·
∫ 2π
0
{
g[Q], g
k(θ,W )
}∣∣
K[0]
δ{J[q], J[p]}
δgk(θ,W )
∣∣∣∣
K[1]
dmθ
(2π)m
+
∫ 2π
0
· · ·
∫ 2π
0
{
J[p], g
k(θ,W )
}∣∣
K[1]
δ{g[Q], J[q]}
δgk(θ,W )
∣∣∣∣
K[0]
dmθ
(2π)m
−
∫ 2π
0
· · ·
∫ 2π
0
{
J[q], g
k(θ,W )
}∣∣
K[1]
δ{g[Q], J[p]}
δgk(θ,W )
∣∣∣∣
K[0]
dmθ
(2π)m
≡ 0.
42 A.Ya. Maltsev
Finally, using the relations (2.61) and (3.15), we can write the above identity as∫ 2π
0
· · ·
∫ 2π
0
{
g[Q], g
k(θ,W )
}∣∣
K[0]
δ{J[q], J[p]}
δgk(θ,W )
∣∣∣∣
K[1]
dmθ
(2π)m
−
∫ 2π
0
· · ·
∫ 2π
0
Ak[1][p]
(
S(W )
ε
+ θ,W
)
δ{g[Q], J[q]}
δgk(θ,W )
∣∣∣∣
K[0]
dmθ
(2π)m
+
∫ 2π
0
· · ·
∫ 2π
0
Ak[1][q]
(
S(W )
ε
+ θ,W
)
δ{g[Q], J[p]}
δgk(θ,W )
∣∣∣∣
K[0]
dmθ
(2π)m
≡ 0,
where the functions Ak[1][q](θ,W ) are introduced by formula (2.64).
Now assume that the values Ak[1][q](θ,W ) for U(W ) ∈ S, according to (3.1), can be repre-
sented in the form
Ak[1][q](θ,W ) = −B̂kj
[0](W )Bj[q](1)(θ,W ) (3.16)
with some smooth in θ, 2π-periodic in each θα functions Bj[q](1)(θ,W ).
We can then write for U(W ) ∈ S∫ 2π
0
· · ·
∫ 2π
0
{
g[Q], g
k(θ,W )
}∣∣
K[0]
δ{J[q], J[p]}
δgk(θ,W )
∣∣∣∣
K[1]
dmθ
(2π)m
(3.17)
+
∫ 2π
0
· · ·
∫ 2π
0
[
B̂kj
[0][S](W )Bj[p](1)
(
S(W )
ε
+ θ,W
)]
δ{g[Q], J[q]}
δgk(θ,W )
∣∣∣∣
K[0]
dmθ
(2π)m
−
∫ 2π
0
· · ·
∫ 2π
0
[
B̂kj
[0][S](W )Bj[q](1)
(
S(W )
ε
+ θ,W
)]
δ{g[Q], J[p]}
δgk(θ,W )
∣∣∣∣
K[0]
dmθ
(2π)m
≡ 0,
where
B̂ij
[0][S](W ) ≡
∑
s≥0
Bij
(s)
(
Φ
(
S(W )
ε
+ θ,U(W )
)
, kγ(W )Φθγ
(
S(W )
ε
+ θ,U(W )
)
, . . .
)
× kα1(W ) · · · kαs(W )
∂s
∂θα1 · · · ∂θαs
.
Using the expression (2.54) for the bracket of constraints on K, as well as the relations (2.42),
(3.4), (3.10), we can obtain also the following relation∫ 2π
0
· · ·
∫ 2π
0
{
g[Q], g
k(θ,W )
}∣∣
K[0]
δ{J[q], J[p]}
δgk(θ,W )
∣∣∣∣
K[1]
dmθ
(2π)m
= −
∫ 2π
0
· · ·
∫ 2π
0
[
B̂kj
[0][S](W )Qj
(
S(W )
ε
+ θ,W
)]
δ{J[q], J[p]}
δgk(θ,W )
∣∣∣∣
K[1]
dmθ
(2π)m
.
Expanding also the remaining terms of the identity (3.17) according to (3.12), we get for
U(W ) ∈ S∫ 2π
0
· · ·
∫ 2π
0
[
B̂kj
[0][S](W )Qj
(
S(W )
ε
+ θ,W
)]
δ{J[q], J[p]}
δgk(θ,W )
∣∣∣∣
K[1]
dmθ
(2π)m
−
∫ 2π
0
· · ·
∫ 2π
0
dmθ
(2π)m
[
B̂kj
[0][S](W )Bj[p](1)
(
S(W )
ε
+ θ,W
)]
×
[∑
l≥0
(−1)lkα1(W ) · · · kαl(W )
∂l
∂θα1 · · · ∂θαl
Whitham’s Method and Dubrovin–Novikov Bracket 43
×
[
Qi
(
S(W )
ε
+ θ,W
)
Siµk(l) (ϕ(θ,W ), . . . )
∣∣∣
K[0]
]
qµ(W )
+ ωαµ(W )qµ(W )Qk,θα
(
S(W )
ε
+ θ,W
)]
+
∫ 2π
0
· · ·
∫ 2π
0
dmθ
(2π)m
[
B̂kj
[0][S](W )Bj[q](1)
(
S(W )
ε
+ θ,W
)]
×
[∑
l≥0
(−1)lkα1(W ) · · · kαl(W )
∂l
∂θα1 · · · ∂θαl
×
[
Qi
(
S(W )
ε
+ θ,W
)
Siµk(l) (ϕ(θ,W ), . . . )
∣∣∣
K[0]
]
pµ(W )
+ ωαµ(W )pµ(W )Qk,θα
(
S(W )
ε
+ θ,W
)]
≡ 0. (3.18)
Consider now the Jacobi identity of the form{
g[P],
{
g[Q], J[q]
}}
+
{
g[Q],
{
J[q], g[P]
}}
+
{
J[q],
{
g[P], g[Q]
}}
≡ 0 (3.19)
for arbitrary fixed functions q(X), and the functionals P(θ, X) and Q(θ, X) defined as before
with the aid of arbitrary functions P̃(θ, X), Q̃(θ, X).
According to the relations{
g[Q,P], U
µ(W )
}∣∣
K = O(ε),
{
J[q], g
k(θ,W )
}∣∣
K = O(ε),
{
J[q], U
µ(W )
}∣∣
K = O(ε),
it’s not difficult to see that after the restriction on K the major term (in ε) of (3.19) will be
written as∫ {
g[P], g
k(θ,W )
}∣∣
K[0]
δ{g[Q], J[q]}
δgk(θ,W )
∣∣∣∣
K[0]
dmθ
(2π)m
dW
−
∫ {
g[Q], g
k(θ,W )
}∣∣
K[0]
δ{g[P], J[q]}
δgk(θ,W )
∣∣∣∣
K[0]
dmθ
(2π)m
dW ≡ 0.
Again recalling that qν(W ) are arbitrary functions of W appearing in the integrand in the
form of local factors, we can write the above relation in a stronger form. That is, for every W∫ 2π
0
· · ·
∫ 2π
0
{
g[P], g
k(θ,W )
}∣∣
K[0]
δ{g[Q], J[q]}
δgk(θ,W )
∣∣∣∣
K[0]
dmθ
(2π)m
−
∫ 2π
0
· · ·
∫ 2π
0
{
g[Q], g
k(θ,W )
}∣∣
K[0]
δ{g[P], J[q]}
δgk(θ,W )
∣∣∣∣
K[0]
dmθ
(2π)m
≡ 0. (3.20)
As well as in the case of the identity (3.18), using the relations (2.54), (2.42), (3.4), (3.15),
we can write the identity (3.20) in the form∫ 2π
0
· · ·
∫ 2π
0
dmθ
(2π)m
[
B̂kj
[0][S](W )Pj
(
S(W )
ε
+ θ,W
)]
×
[∑
l≥0
(−1)lkα1(W ) · · · kαl(W )
∂l
∂θα1 · · · ∂θαl
×
[
Qi
(
S(W )
ε
+ θ,W
)
Siµk(l) (ϕ(θ,W ), . . . )
∣∣∣
K[0]
]
qµ(W )
44 A.Ya. Maltsev
+ ωαµ(W )qµ(W )Qk,θα
(
S(W )
ε
+ θ,W
)]
−
∫ 2π
0
· · ·
∫ 2π
0
dmθ
(2π)m
[
B̂kj
[0][S](W )Qj
(
S(W )
ε
+ θ,W
)]
×
[∑
l≥0
(−1)lkα1(W ) · · · kαl(W )
∂l
∂θα1 · · · ∂θαl
×
[
Pi
(
S(W )
ε
+ θ,W
)
Siµk(l) (ϕ(θ,W ), . . . )
∣∣∣
K[0]
]
qµ(W )
+ ωαµ(W )qµ(W )Pk,θα
(
S(W )
ε
+ θ,W
)]
≡ 0. (3.21)
Note now that the values of Q(θ, X) and P(θ, X) are arbitrary 2π-periodic functions of θ,
satisfying the conditions (3.4). In particular, we can put in (3.21) at U(W ) ∈ S
P(θ,W ) = B[p](1)(θ,W ) or P(θ,W ) = B[q](1)(θ,W ). (3.22)
By analogy with (3.12) we introduce for convenience the notation for U(W ) ∈ S
δ{g[εB[p](1)], J[q]}
δgk(θ,W )
∣∣∣∣∣
K[1]
≡
[∑
l≥0
(−1)lkα1(W ) · · · kαl(W )
∂l
∂θα1 · · · ∂θαl
×
[
Bi[p](1)
(
S(W )
ε
+ θ,W
)
Siµk(l) (ϕ(θ,W ), . . . )
∣∣∣
K[0]
]
qµ(W )
+ ωαµ(W )qµ(W )Bk[p](1),θα
(
S(W )
ε
+ θ,W
)]
(3.23)
for arbitrary smooth functions q(X), p(X).
Note that the functional g[εB[p](1)] is not defined on the whole functional space, so the rela-
tion (3.23) plays just a role of a formal notation for U(W ) ∈ S.
Using now (3.21) for the functions (3.22), we can rewrite (3.18) in the form∫ 2π
0
· · ·
∫ 2π
0
dmθ
(2π)m
[
B̂kj
[0][S](W )Qj
(
S(W )
ε
+ θ,W
)]
×
δ{J[q], J[p]}
δgk(θ,W )
∣∣∣∣
K[1]
−
δ{g[εB[p](1)], J[q]}
δgk(θ,W )
∣∣∣∣∣
K[1]
+
δ{g[εB[q](1)], J[p]}
δgk(θ,W )
∣∣∣∣∣
K[1]
≡ 0
provided that U(W ) ∈ S.
Using the skew-symmetry of the Hamiltonian operator B̂kj
[0][S](W ) on Λ we can then write
∫ 2π
0
· · ·
∫ 2π
0
dmθ
(2π)m
Qj
(
S(W )
ε
+ θ,W
)
×
[
B̂jk
[0][S](W )
(
δ{J[q], J[p]}
δgk(θ,W )
∣∣∣∣
K[1]
−
δ{g[εB[p](1)], J[q]}
δgk(θ,W )
∣∣∣∣∣
K[1]
+
δ{g[εB[q](1)], J[p]}
δgk(θ,W )
∣∣∣∣∣
K[1]
)]
≡ 0.
The values Qj(θ,W ) are arbitrary smooth 2π-periodic functions of θ with the only restric-
tion (3.4). We know also that the values in the brackets are smooth 2π-periodic in each θα
Whitham’s Method and Dubrovin–Novikov Bracket 45
functions of θ for U(W ) ∈ S. As a consequence, we can write for U(W ) ∈ S
B̂jk
[0][S](W )
δ{J[q], J[p]}
δgk(θ,W )
∣∣∣∣
K[1]
−
δ{g[εB[p](1)], J[q]}
δgk(θ,W )
∣∣∣∣∣
K[1]
+
δ{g[εB[q](1)], J[p]}
δgk(θ,W )
∣∣∣∣∣
K[1]
≡
m∑
α=1
aα(U(W ),UW (W ))Φj
θα
(
S(W )
ε
+ θ,U(W )
)
with some coefficients aα(U(W ),UW (W )).
The values in parentheses are smooth 2π-periodic functions of θ at U(W ) ∈ S. At the
same time the trajectories of the vector field (k1(W ), . . . , km(W )) are completely irrational
windings of the torus Tm. We can therefore say in the case of a regular Hamiltonian family Λ,
that up to a linear combination of the regular eigenvectors v
(l)
[U(W )](S(W )/ε+θ) of B̂jk
[0][S](W ) =
B̂jk
[0][S](U(W )), corresponding to zero eigenvalues, the value in parentheses is a linear combination
of the variational derivatives (2.7), generating linear shifts of the phases on Λ. For complete
Hamiltonian set of the integrals (I1, . . . , IN ) we can then write according to (2.8) and (3.3)
δ{J[q], J[p]}
δgk(θ,W )
∣∣∣∣
K[1]
−
δ{g[εB[p](1)], J[q]}
δgk(θ,W )
∣∣∣∣∣
K[1]
+
δ{g[εB[q](1)], J[p]}
δgk(θ,W )
∣∣∣∣∣
K[1]
≡
∑
q
bq(U(W ),UW (W ))κ
(q)
k[U(W )]
(
S(W )
ε
+ θ
)
(3.24)
with some coefficients bq(U(W ),UW (W )) at U(W ) ∈ S.
Consider the Jacobi identity of the form{{
J[q], J[p]
}
, J[r]
}
+
{{
J[p], J[r]
}
, J[q]
}
+
{{
J[r], J[q]
}
, J[p]
}
≡ 0
with arbitrary smooth functions q(X), p(X), r(X).
In the main (∼ε2) order on K given identity leads to the relations∫
δ{J[q], J[p]}
δUγ(W )
∣∣∣∣
K[1]
{
Uγ(W ), J[r]
}∣∣
K[1]
dW + c.p.
+
∫∫ 2π
0
· · ·
∫ 2π
0
δ{J[q], J[p]}
δgk(θ,W )
∣∣∣∣
K[1]
{
gk(θ,W ), J[r]
}∣∣
K[1]
dmθ
(2π)m
dW + c.p. ≡ 0. (3.25)
Again, using the relations (2.61), (2.64) and (3.10), we can replace the identity (3.25) by the
following relation∫
δ{J[q], J[p]}
δUγ(W )
∣∣∣∣
K[1]
{
Uγ(W ), J[r]
}∣∣
K[1]
dW + c.p.
+
∫∫ 2π
0
· · ·
∫ 2π
0
δ{J[q], J[p]}
δgk(θ,W )
∣∣∣∣
K[1]
Ak[1][r]
(
S(W )
ε
+ θ,W
)
dmθ
(2π)m
dW + c.p. ≡ 0.
Using relations (2.71) and representations (3.16) and (3.24), we can write for U(W ) ∈ S∫ 2π
0
· · ·
∫ 2π
0
δ{J[q], J[p]}
δgk(θ,W )
∣∣∣∣
K[1]
Ak[1][r]
(
S(W )
ε
+ θ,W
)
dmθ
(2π)m
+ c.p.
≡
∫ 2π
0
· · ·
∫ 2π
0
[
B̂kj
[0][S](W )Bj[r](1)
(
S(W )
ε
+ θ,W
)]
46 A.Ya. Maltsev
×
δ{g[εB[q](1)], J[p]}
δgk(θ,W )
∣∣∣∣∣
K[1]
−
δ{g[εB[p](1)], J[q]}
δgk(θ,W )
∣∣∣∣∣
K[1]
dmθ
(2π)m
+
∫ 2π
0
· · ·
∫ 2π
0
[
B̂kj
[0][S](W )Bj[q](1)
(
S(W )
ε
+ θ,W
)]
×
δ{g[εB[p](1)], J[r]}
δgk(θ,W )
∣∣∣∣∣
K[1]
−
δ{g[εB[r](1)], J[p]}
δgk(θ,W )
∣∣∣∣∣
K[1]
dmθ
(2π)m
+
∫ 2π
0
· · ·
∫ 2π
0
[
B̂kj
[0][S](W )Bj[p](1)
(
S(W )
ε
+ θ,W
)]
×
δ{g[εB[r](1)], J[q]}
δgk(θ,W )
∣∣∣∣∣
K[1]
−
δ{g[εB[q](1)], J[r]}
δgk(θ,W )
∣∣∣∣∣
K[1]
dmθ
(2π)m
. (3.26)
Similarly to earlier arguments, substituting now in the identity (3.21)
Q(θ,W ) = B[r](1)(θ,W ), P(θ,W ) = B[p](1)(θ,W )
for U(W ) ∈ S we obtain the identity∫ 2π
0
· · ·
∫ 2π
0
[
B̂kj
[0][S](W )Bj[p](1)
(
S(W )
ε
+ θ,W
)]
δ{g[εB[r](1)], J[q]}
δgk(θ,W )
∣∣∣∣∣
K[1]
dmθ
(2π)m
(3.27)
−
∫ 2π
0
· · ·
∫ 2π
0
[
B̂kj
[0][S](W )Bj[r](1)
(
S(W )
ε
+ θ,W
)]
δ{g[εB[p](1)], J[q]}
δgk(θ,W )
∣∣∣∣∣
K[1]
dmθ
(2π)m
= 0.
Using the cyclic permutations of the functions q(X), p(X), and r(X) in the identity (3.27),
it’s not difficult to see that the right-hand part of the relation (3.26) is identically equal to zero
at U(W ) ∈ S. It’s not difficult to see also that the left-hand side of the expression (3.26) is
a smooth regular function of the parameters U(W ). Using the fact that the set S is everywhere
dense in the parameter space U, we can conclude that the left-hand side of equation (3.26) is
identically equal to zero under the conditions of the theorem.
We have, therefore, that under the conditions of the theorem, the identity (3.25) implies the
relation∫
δ{J[q], J[p]}
δUγ(W )
∣∣∣∣
K[1]
{
Uγ(W ), J[r]
}∣∣
K[1]
dW + c.p. ≡ 0.
Using the relations
δ{J[q], J[p]}
δUγ(W )
∣∣∣∣
K[1]
=
δ{U[q], U[p]}DN
δUγ(W )
,
{
Uγ(W ), J[r]
}∣∣
K[1]
=
{
Uγ(W ), U[r]
}
DN
,
we obtain the Jacobi identity for the Dubrovin–Novikov bracket on the space of the func-
tions U(X). �
Remark 3.1. According to relations (2.47) the functionals
Kα =
∫ +∞
−∞
kα (U(X)) dX
are annihilators of the Dubrovin–Novikov bracket (2.21) since we have {Kα, U[q]} = 0 for any
functional U[q]. As a consequence, we can claim that the functions kα(U) represent a part of the
Whitham’s Method and Dubrovin–Novikov Bracket 47
flat coordinates for the metric gνµ(U) = 〈Aνµ1 〉 connected with the Dubrovin–Novikov bracket
for the Whitham system. Besides that, we have{
kα(X), kβ(Y )
}
DN
= 0
according to relations (2.48).
The remaining part of the flat coordinates of gνµ(U) and the corresponding annihilators of
the bracket (2.21) are defined by concrete form of the initial bracket (2.16) and the family of
m-phase solutions of system (2.3).
Let us prove now the second theorem justifying the invariance of the Dubrovin–Novikov
procedure.
Theorem 3.2. Let Λ be a regular Hamiltonian family of m-phase solutions of (2.3). Let
(I1, . . . , IN ) and (I ′1, . . . , I ′N ) be two different complete Hamiltonian sets of commuting first
integrals of (2.3) having the form (2.18). Then the Dubrovin–Novikov brackets obtained using
the sets (I1, . . . , IN ) and (I ′1, . . . , I ′N ) coincide with each other.
Proof. Note that the sets (I1, . . . , IN ), (I ′1, . . . , I ′N ) correspond to the two systems of coordi-
nates (U1, . . . , UN ), (U ′1, . . . , U ′N ) on the family Λ, given by the averages of the functionals I
and I′. We have to show then that the Dubrovin–Novikov brackets obtained using the set
(I ′1, . . . , I ′N ) coincides with the bracket, obtained using the set (I1, . . . , IN ), after the corre-
sponding change of coordinates,
U ′ν = U ′ν(U).
Consider the values of the functionals
Jν(X) =
∫ 2π
0
· · ·
∫ 2π
0
P ν(ϕ, εϕX , . . . )
dmθ
(2π)m
,
and the values of constraints gi(θ, X), introduced with the aid of the functionals J(X) by
formula (2.32), as a coordinate system in the neighborhood of the submanifold K.
For values of the functionals J ′ν(X) on the submanifold K we then have the relations
J ′ν(X)
∣∣
K = U ′ν (J(X)) +
∑
l≥1
εlj′ν(l) (J,JX , . . . ) ,
where j′ν(l) are smooth functions of (J,JX , . . . ), polynomial in the derivatives and having degree l.
Expanding the values of J ′ν(X) in the neighborhood of the submanifold K, we can write
J ′ν(X) = U ′ν (J(X)) +
∑
l≥1
εlj′ν(l) (J,JX , . . . )
+
∫ +∞
−∞
∫ 2π
0
· · ·
∫ 2π
0
T ′νi (X,θ, Y, ε) gi(θ, Y )
dmθ
(2π)m
dY +O
(
g2
)
,
where, according to the form of the constraints (2.32), we can put
T ′νi (X,θ, Y, ε) =
∑
l≥0
Π′νi(l) (ϕ(θ, X), εϕX(θ, X), . . . )
∣∣∣
K
εlδ(l)(X − Y )
=
∑
l≥0
∂P ′ν
∂ϕilx
(ϕ(θ, X), εϕX(θ, X), . . . )
∣∣∣∣
K
εlδ(l)(X − Y ).
48 A.Ya. Maltsev
Considering the functionals J ′[q] =
∫
qν(X)J ′ν(X)dX with arbitrary smooth (compactly sup-
ported) functions qν(X), we can write in the vicinity of K
J ′[q] =
∫
qν(X)
[
U ′ν (J(X)) +
∑
l≥1
εlj′ν(l) (J,JX , . . . )
]
dX (3.28)
+
∫ ∑
l≥0
(−1)lεl
[
dl
dY l
qν(Y ) Π′νi(l) (ϕ(θ, Y ), εϕY (θ, Y ), . . . )
∣∣∣
K
]
gi(θ, Y )
dmθ
(2π)m
dY +O
(
g2
)
.
The leading term (in ε) in the second part of the expression (3.28) is given by the expression∫
qν(Y )
∑
l≥0
(−1)lkα1(Y ) · · · kαl(Y )Π′νi(l)θα1 ···θαl
(
Φ
(
S(Y )
ε
+ θ, Y
)
, . . .
)
gi(θ, Y )
dmθ
(2π)m
dY
and coincides with the value∫∫ 2π
0
· · ·
∫ 2π
0
qν(Y )ζ
′(ν)
i[U(Y )]
(
S(Y )
ε
+ θ
)
gi(θ, Y )
dmθ
(2π)m
dY,
where
ζ
′(ν)
i[U](θ) =
[
δ
δϕi(θ)
∫ 2π
0
· · ·
∫ 2π
0
P ′ν
(
ϕ, kβϕθβ , . . .
) dmθ
(2π)m
]∣∣∣∣
ϕ(θ)=Φ(θ,U)
.
As the values ζ
(ν)
i[U](θ), the values ζ
′(ν)
i[U](θ) represent regular left eigenvectors of the opera-
tor L̂ij[U] corresponding to the zero eigenvalues. In the case of a complete regular family of
m-phase solutions, we have therefore
ζ
′(ν)
i[U](θ) =
∑
q
Γνq (U)κ
(q)
i[U](θ)
for some functions Γνq (U).
We can write, therefore, up to quadratic terms in g(θ, X)∫
qν(X)J ′ν(X)dX =
∫
qν(X)
[
U ′ν (J(X)) +
∑
l≥1
εlj′ν(l) (J,JX , . . . )
]
dX
+
∫
qν(X)
[∑
q
Γνq (U)κ
(q)
i[U(X)]
(
S(X)
ε
+ θ
)
+O(ε)
]
gi(θ, X)
dmθ
(2π)m
dX +O
(
g2
)
.
Consider the Poisson brackets{
J ′[q], J
′
[p]
}∣∣
K =
{∫
qν(X)J ′ν(X)dX,
∫
pµ(Y )J ′µ(Y )dY
}∣∣∣∣
K
=
∫
qν(X)
∂U ′ν
∂Uλ
(X)
[{
Jλ(X),
∫
pµ(Y )J ′µ(Y )dY
}∣∣∣∣
K
+O
(
ε2
)]
dX
+
∫
qν(X)
∑
q
Γνq (U(X))
[
κ
(q)
i[U(X)]
(
S(X)
ε
+ θ
)
+O(ε)
]
×
{
gi(θ, X), J ′µ(Y )
}∣∣
K pµ(Y )
dmθ
(2π)m
dXdY.
Whitham’s Method and Dubrovin–Novikov Bracket 49
By Lemma 2.4′ we have the relation {gi(θ, X), J ′[p]}|K = O(ε) on the submanifold K. In
addition, completely analogous to the relation (2.71) holds the relation∫ 2π
0
· · ·
∫ 2π
0
κ
(q)
i[U(X)]
(
S(X)
ε
+ θ
){
gi(θ, X), J ′[p]
}
|K[1]
dmθ
(2π)m
≡ 0
by virtue of the original dependence of the constraints gi(θ, X). We thus obtain∫∫
qν(X)
{
J ′ν(X), J ′µ(Y )
}∣∣
Kpµ(Y )dXdY
=
∫
qν(X)
∂U ′ν
∂Uλ
(X)
{
Jλ(X),
∫
pµ(Y )J ′µ(Y )dY
}∣∣∣∣
K
dX +O
(
ε2
)
.
Repeating the arguments for the functional
∫
pµ(Y )J ′µ(Y )dY we finally obtain∫∫
qν(X)
{
J ′ν(X), J ′µ(Y )
}∣∣
Kpµ(Y )dXdY
=
∫∫
qν(X)
∂U ′ν
∂Uλ
(X)
{
Jλ(X), Jγ(Y )
}∣∣
K
∂U ′µ
∂Uγ
(Y )pµ(Y )dXdY +O
(
ε2
)
. (3.29)
Given that the principal (in ε) terms in the expressions {Jλ(X), Jγ(Y )}|K and
{J ′ν(X), J ′µ(Y )}|K coincide with the Dubrovin–Novikov brackets, obtained with the aid of the
sets (I1, . . . , IN ) and (I ′1, . . . , I ′N ) respectively, we conclude from (3.29) that
{
U ′ν(X), U ′µ(Y )
}′
DN
=
∂U ′ν
∂Uλ
(X)
{
Uλ(X), Uγ(Y )
}
DN
∂U ′µ
∂Uγ
(Y ),
which means the coinciding of the brackets {·, ·}DN and {·, ·}′DN. �
Finally, we prove the theorem about the Hamiltonian properties of the Whitham system (2.10)
under the same conditions as before.
Theorem 3.3. Let Λ be a regular Hamiltonian family of m-phase solutions of (2.3). Let
(I1, . . . , IN ) be a complete Hamiltonian set of commuting first integrals of (2.3) having the
form (2.18) and H be the Hamiltonian function of the system (2.3) having the form (2.17).
Then the Whitham system (2.10) is Hamiltonian with respect to the corresponding Dubrovin–
Novikov bracket (2.21) with the Hamiltonian function (2.22)
Hav =
∫ +∞
−∞
〈PH〉 (U(X)) dX.
Proof. By Theorem 3.2, without loss of generality we can assume that the Hamiltonian func-
tional H belongs to the set (I1, . . . , IN ), H = Iµ0 . It is easy to verify that the corresponding
Hamiltonian Hav generates in this case the system
UνT =
d
dX
[〈Qνµ0〉 (U(X))] ,
i.e. exactly system (2.10). �
As the simplest example of Theorems 3.1–3.3 consider the procedure of averaging of the
Gardner–Zakharov–Faddeev bracket for the KdV equation.
Consider the Gardner–Zakharov–Faddeev bracket
{ϕ(x), ϕ(y)} = δ′(x− y) (3.30)
50 A.Ya. Maltsev
for the KdV equation
ϕt = ϕϕx − ϕxxx
with the Hamiltonian functional
H =
∫ (
ϕ3
6
+
ϕ2
x
2
)
dx.
As is well known, the KdV equation has a family of m-phase solutions for any m, given
by the Novikov potentials, which represent the stationary points for the higher KdV flows.
According to the Hamiltonian structure (3.30) it is also equivalent to the extremality of a linear
combination of higher integrals of the KdV
c1δI
1 + c2δI
2 + · · ·+ cm+2δI
m+2 = 0, (3.31)
where
I1 = N =
∫
ϕdx
is the annihilator of the bracket (3.30),
I2 = P =
∫
ϕ2
2
dx
is the momentum functional of the bracket (3.30), I3 = H, and Ik, k ≥ 4 are higher integrals of
the KdV equation.
As was shown in [41], all the systems (3.31) are completely integrable finite-dimensional
systems with quasiperiodic solutions in the general case. As is well known, the correspond-
ing m-phase solutions of KdV can be represented in the form (1.2) with the theta-functional
expressions for the functions Φ(θ,U), satisfying the system [10–13,17,26,27]
ωα(U)Φθα = kα(U)ΦΦθα − kα(U)kβ(U)kγ(U)Φθαθβθγ . (3.32)
The parameters of the solution (E1, . . . , E2m+1), E1 < E2 < · · · < E2m+1 represent the
branching points of the Riemann surface of genus m, and, together with the initial phases
(θ1
0, . . . , θ
m
0 ), completely determine the corresponding solution ϕ(x, t) [41].
The eigenmodes of the linearized operator (3.32), as well as the adjoint operator, were studied
in detail [4, 5, 28–30]. In particular, we can state that the families of m-phase solutions of KdV
are complete regular families in the sense of Definition 1.1 and the corresponding systems (2.10)
for any independent set of 2m+1 higher integrals of KdV (Ik1 , . . . , Ik2m+1) represent the regular
Whitham systems for the KdV equation in the sense defined above. It is well known [22, 49],
that the Whitham system for KdV in m-phase case can be written in the diagonal form where
the parameters E1, . . . , E2m+1 are its Riemann invariants.
Regular zero modes of the linearized operator (3.32) on the torus Tm are given by the functions
Φθα(θ,U), θα = 1, . . . ,m and by the function ∂Φ/∂n (at constant k and ω), where n is the
value of the functional N on the corresponding solution. Similarly, regular zero modes of the
adjoint linear operator on the torus Tm are given by any m+ 1 linearly independent variational
derivatives of the higher integrals of KdV on the family of the functions Φ(θ,U).
The families of Novikov potentials are obviously regular Hamiltonian families with respect to
the bracket (3.30), while any set of 2m+1 independent higher integrals of KdV is a complete set
of commuting functionals (2.18) on such a family. We also note that the variational derivative
of the annihilator of the bracket (3.30) always appears on the family of m-phase solutions in the
Whitham’s Method and Dubrovin–Novikov Bracket 51
form of a linear combination of m+1 variational derivatives of higher integrals of KdV according
to the original construction of the Novikov potentials.
Investigation of the system (3.1) can be carried out in this case as follows. We first note that
the operator B̂[0](X) is given in this case as
B̂[0](X) = k1(X)
∂
∂θ1
+ · · ·+ km(X)
∂
∂θm
and, thus, is the derivative along the vector field (k1(X), . . . , km(X)) on the torus Tm.
The regular zero mode of the operator B̂[0](X) is a constant function on the torus and is
orthogonal to the functions A[1][q](θ, X) by Lemma 3.1. This fact is easily verified also by direct
computation. Indeed, the functions A[1][q](θ, X) have in this case the form
A[1][q]
(
S(X)
ε
+ θ, X
)
=
{
ϕ(θ, X), J[q]
}∣∣
K[1]
− ΦUν
(
S(X)
ε
+ θ,U(X)
){
Uν(X), J[q]
}∣∣
K[1]
.
Let us assume for simplicity that the functional
N(X) =
∫ 2π
0
· · ·
∫ 2π
0
ϕ(θ, X)
dmθ
(2π)m
is included in the coordinate system U(X), N(X) = U1(X), which, by Theorem 3.2, does not
affect the Dubrovin–Novikov bracket. We have then∫ 2π
0
· · ·
∫ 2π
0
ΦU1(θ,U(X))
dmθ
(2π)m
≡ 1,
∫ 2π
0
· · ·
∫ 2π
0
ΦUν (θ,U(X))
dmθ
(2π)m
≡ 0, ν 6= 1.
It is easy to see then that∫ 2π
0
· · ·
∫ 2π
0
A[1][q](θ, X)
dmθ
(2π)m
=
{
N(X), J[q]
}∣∣
K[1]
−
{
N(X), J[q]
}∣∣
K[1]
≡ 0.
The operator B̂[0](X), however, has also irregular zero modes which arise in the cases when
the closure of trajectories of the vector field (k1(X), . . . , km(X)) on Tm represent the lower-
dimensional tori Tk ⊂ Tm. Modes of this type arise obviously on a set of measure zero in the
parameter space U, i.e. {U}/M.
We assume here, without proof, following easily verifiable from the theta-functional represen-
tation fact. The Fourier harmonics An1...nm
[1][q] (X) of the functions A[1][q](θ, X) in the variables θ
A[1][q](θ, X) ≡
∑
n1,...,nm
An1...nm
[1][q] (X) exp
(
in1θ
1 + · · ·+ inmθ
m
)
decay faster than any power of |n| at |n| → ∞, where
|n| ≡
√
n2
1 + · · ·+ n2
m.
Let us define the Diophantine conditions for an arbitrary set of values (k1(U), . . . , km(U)).
The vector (k1(U), . . . , km(U)) is a Diophantine vector with the index ν > 0 and the coefficient
A > 0, if∣∣n1k
1(U) + · · ·+ nmk
m(U)
∣∣ ≥ A|n|−ν
for all (n1, . . . , nm) ∈ Zm ((n1, . . . , nm) 6= (0, . . . , 0)).
Let us use now the following well known theorem (see, e.g., [3, 45]):
52 A.Ya. Maltsev
For ν > m− 1 the measure of the set of non-Diophantine vectors (k1, . . . , km) in Rm is equal
to zero.
Using the condition rank ||∂kα/∂Uν || = m we can also state that the measure of the corre-
sponding parameters U, such that k(U) are non-Diophantine vectors with index ν > m− 1, is
also zero in the parameter space U.
We can now define the sets Sν in the space of the parameters U, such that the wave vectors
(k1(U), . . . , km(U)) are Diophantine with index ν if U ∈ Sν . In this case all Sν with index
ν > m− 1 are everywhere dense, and also have the full measure in the parameter space U.
It is easy to see also that the system (3.1) is resolvable in the space of 2π-periodic in all
θα functions for each U ∈ Sν . Indeed, due to the absence of the zero Fourier harmonic of the
function A[1][q](θ, X) and conditions of decreasing of Fourier coefficients An1...nm
[1][q] (X) we can put
Bn1...nm
[q](1) (X) =
1
in1k1(U) + · · ·+ inmkm(U)
An1...nm
[1][q] (X)
for (n1, . . . , nm) 6= (0, . . . , 0) and restore the solution
B[q](1)(θ, X) =
′∑
n1,...,nm
Bn1...nm
[q](1) (X) exp(in1θ
1 + · · ·+ inmθ
m),
(n1, . . . , nm) 6= (0, . . . , 0)
as a smooth m-periodic function of θ.
Assuming, therefore, S = Sν for any ν > m− 1, we meet all the conditions of Theorem 3.1.
We can thus formulate the following proposition.
Proposition 3.1. The Dubrovin–Novikov procedure is well justified in the averaging of the
Gardner–Zakharov–Faddeev bracket on the m-phase solutions of KdV for any m and provides
a local Hamiltonian structure of hydrodynamic type for the corresponding regular Whitham sys-
tem.
We note here that both local and weakly nonlocal Hamiltonian structures of the Whitham
hierarchy for KdV were investigated in detail in the papers [2, 44].
In the more general case the constructing of the set S requires the study of the eigenfunctions
and the eigenvalues of the Hamilton operator B̂ij
[0](U) on the manifold of m-phase solutions
depending on the values of the parameters U. For m-phase solutions, given by the algebraic-
geometric families, in this case are very convenient methods similar to those used in [4,5,28,30,47]
to study the spectrum of auxiliary linear operators of integrable hierarchies.
Acknowledgements
In conclusion the author expresses his deep gratitude to Professors S.P. Novikov, B.A. Dubrovin,
I.M. Krichever, S.Yu. Dobrokhotov, and M.V. Pavlov for fruitful discussions. This work was
financially supported by the Russian Federation Government Grant No. 2010-220-01-077, Grant
of the President of Russian Federation NSh-4995.2012.1, and Grant RFBR No. 11-01-12067-ofi-
m-2011.
References
[1] Ablowitz M.J., Benney D.J., The evolution of multi-phase modes for nonlinear dispersive waves, Stud. Appl.
Math. 49 (1970), 225–238.
[2] Alekseev V.L., On non-local Hamiltonian operators of hydrodynamic type connected with Whitham’s equa-
tions, Russian Math. Surveys 50 (1995), 1253–1255.
http://dx.doi.org/10.1070/RM1995v050n06ABEH002640
Whitham’s Method and Dubrovin–Novikov Bracket 53
[3] Arnol’d V.I., Geometrical methods in the theory of ordinary differential equations, Grundlehren der Mathe-
matischen Wissenschaften, Vol. 250, 2nd ed., Springer-Verlag, New York, 1988.
[4] Dobrokhotov S.Yu., Resonance correction of an adiabatically perturbed finite-gap almost periodic solution
of the Korteweg–de Vries equation, Math. Notes 44 (1988), 551–555.
[5] Dobrokhotov S.Yu., Resonances in asymptotic solutions of the Cauchy problem for the Schrödinger equation
with rapidly oscillating finite-zone potential, Math. Notes 44 (1988), 656–668.
[6] Dobrokhotov S.Yu., Krichever I.M., Multi-phase solutions of the Benjamin–Ono equation and their avera-
ging, Math. Notes 49 (1991), 583–594.
[7] Dobrokhotov S.Yu., Maslov V.P., Finite-gap almost periodic solutions in the WKB approximation, J. Sov.
Math. 15 (1980), 1433–1487.
[8] Dobrokhotov S.Yu., Maslov V.P., Multiphase asymptotics of nonlinear partial differential equations with
a small parameter, in Mathematical Physics Reviews, Soviet Sci. Rev. Sect. C Math. Phys. Rev., Vol. 3,
Harwood Academic Publ., Chur, 1982, 221–311.
[9] Dobrokhotov S.Yu., Minenkov D.S., Remark on the phase shift in the Kuzmak–Whitham ansatz, Theoret.
Math. Phys. 166 (2011), 303–316.
[10] Dubrovin B.A., Inverse problem for periodic finite-zoned potentials in the theory of scattering, Funct. Anal.
Appl. 9 (1975), 61–62.
[11] Dubrovin B.A., Theta functions and non-linear equations, Russian Math. Surveys 36 (1981), no. 2, 11–92.
[12] Dubrovin B.A., Matveev V.B., Novikov S.P., Nonlinear equations of Korteweg–de Vries type, finite-band
linear operators and Abelian varieties, Russian Math. Surveys 31 (1976), no. 1, 59–146.
[13] Dubrovin B.A., Novikov S.P., A periodic problem for the Korteweg–de Vries and Sturm–Liouville equations.
Their connection with algebraic geometry, Soviet Math. Dokl. 15 (1976), 1597–1601.
[14] Dubrovin B.A., Novikov S.P., Hamiltonian formalism of one-dimensional systems of the hydrodynamic type
and the Bogolyubov–Whitham averaging method, Soviet Math. Dokl. 27 (1983), 665–669.
[15] Dubrovin B.A., Novikov S.P., Hydrodynamics of soliton lattices, Sov. Sci. Rev. Sect. C 9 (1992), no. 4,
1–136.
[16] Dubrovin B.A., Novikov S.P., Hydrodynamics of weakly deformed soliton lattices. Differential geometry and
Hamiltonian theory, Russian Math. Surveys 44 (1989), no. 6, 35–124.
[17] Dubrovin B.A., Novikov S.P., Periodic and conditionally periodic analogs of the many-soliton solutions of
the Korteweg–de Vries equation, Soviet Physics JETP 67 (1974), 1058–1063.
[18] Ferapontov E.V., Differential geometry of nonlocal Hamiltonian operators of hydrodynamic type, Funct.
Anal. Appl. 25 (1991), 195–204.
[19] Ferapontov E.V., Nonlocal Hamiltonian operators of hydrodynamic type: differential geometry and appli-
cations, in Topics in Topology and Mathematical Physics, Amer. Math. Soc. Transl. Ser. 2, Vol. 170, Amer.
Math. Soc., Providence, RI, 1995, 33–58.
[20] Ferapontov E.V., Nonlocal matrix Hamiltonian operators, differential geometry and applications, Theoret.
Math. Phys. 91 (1992), 642–649.
[21] Ferapontov E.V., Restriction, in the sense of Dirac, of the Hamiltonian operator δij(d/dx) to a surface of
the Euclidean space with a plane normal connection, Funct. Anal. Appl. 26 (1992), 298–300.
[22] Flaschka H., Forest M.G., McLaughlin D.W., Multiphase averaging and the inverse spectral solution of the
Korteweg–de Vries equation, Comm. Pure Appl. Math. 33 (1980), 739–784.
[23] Haberman R., Standard form and a method of averaging for strongly nonlinear oscillatory dispersive traveling
waves, SIAM J. Appl. Math. 51 (1991), 1638–1652.
[24] Haberman R., The modulated phase shift for weakly dissipated nonlinear oscillatory waves of the Korteweg–
de Vries type, Stud. Appl. Math. 78 (1988), 73–90.
[25] Hayes W.D., Group velocity and nonlinear dispersive wave propagation, Proc. R. Soc. Lond. Ser. A 332
(1973), 199–221.
[26] Its A.R., Matveev V.B., Hill’s operator with finitely many gaps, Funct. Anal. Appl. 9 (1975), 65–66.
[27] Its A.R., Matveev V.B., Schrödinger operators with finite-gap spectrum and N -soliton solutions of the
Korteweg–de Vries equation, Theoret. Math. Phys. 23 (1975), 343–355.
[28] Krichever I.M., Perturbation theory in periodic problems for two-dimensional integrable systems, Sov. Sci.
Rev. Sect. C 9 (1992), no. 2, 1–103.
http://dx.doi.org/10.1007/978-1-4612-1037-5
http://dx.doi.org/10.1007/978-1-4612-1037-5
http://dx.doi.org/10.1007/BF01159126
http://dx.doi.org/10.1007/BF01156581
http://dx.doi.org/10.1007/s11232-011-0025-y
http://dx.doi.org/10.1007/s11232-011-0025-y
http://dx.doi.org/10.1007/BF01078183
http://dx.doi.org/10.1007/BF01078183
http://dx.doi.org/10.1070/RM1981v036n02ABEH002596
http://dx.doi.org/10.1070/RM1976v031n01ABEH001446
http://dx.doi.org/10.1070/RM1989v044n06ABEH002300
http://dx.doi.org/10.1007/BF01085489
http://dx.doi.org/10.1007/BF01085489
http://dx.doi.org/10.1007/BF01017341
http://dx.doi.org/10.1007/BF01017341
http://dx.doi.org/10.1007/BF01075056
http://dx.doi.org/10.1002/cpa.3160330605
http://dx.doi.org/10.1137/0151084
http://dx.doi.org/10.1098/rspa.1973.0021
http://dx.doi.org/10.1007/BF01078185
http://dx.doi.org/10.1007/BF01038218
54 A.Ya. Maltsev
[29] Krichever I.M., The averaging method for two-dimensional “integrable” equations, Funct. Anal. Appl. 22
(1988), 200–213.
[30] Krichever I.M., The “Hessian” of integrals of the Korteweg–de Vries equation and perturbations of finite-gap
solutions, Sov. Math. Dokl. 270 (1983), 757–761.
[31] Luke J.C., A perturbation method for nonlinear dispersive wave problems, Proc. R. Soc. Lond. Ser. A 292
(1966), 403–412.
[32] Maltsev A.Ya., Conservation of Hamiltonian structures in Whitham’s averaging method, Izv. Math. 63
(1999), 1171–1201.
[33] Maltsev A.Ya., Deformations of the Whitham systems in the almost linear case, in Geometry, Topology,
and Mathematical Physics, Amer. Math. Soc. Transl. Ser. 2, Vol. 224, Amer. Math. Soc., Providence, RI,
2008, 193–212, arXiv:0709.4618.
[34] Maltsev A.Ya., The averaging of nonlocal Hamiltonian structures in Whitham’s method, Int. J. Math. Math.
Sci. 30 (2002), 399–434, solv-int/9910011.
[35] Maltsev A.Ya., Whitham systems and deformations, J. Math. Phys. 47 (2006), 073505, 18 pages,
nlin.SI/0509033.
[36] Maltsev A.Ya., Novikov S.P., On the local systems Hamiltonian in the weakly non-local Poisson brackets,
Phys. D 156 (2001), 53–80, nlin.SI/0006030.
[37] Maltsev A.Ya., Pavlov M.V., On Whitham’s averaging method, Funct. Anal. Appl. 29 (1995), 6–19,
nlin.SI/0306053.
[38] Mokhov O.I., Ferapontov E.V., Non-local Hamiltonian operators of hydrodynamic type related to metrics
of constant curvature, Russian Math. Surveys 45 (1990), 218–219.
[39] Newell A.C., Solitons in mathematics and physics, CBMS-NSF Regional Conference Series in Applied Ma-
thematics, Vol. 48, Society for Industrial and Applied Mathematics (SIAM), Philadelphia, PA, 1985.
[40] Novikov S.P., Geometry of conservative systems of hydrodynamic type. The averaging method for field-
theoretic systems, Russian Math. Surveys 40 (1985), no. 4, 85–98.
[41] Novikov S.P., The periodic problem for the Korteweg–de Vries equation, Funct. Anal. Appl. 8 (1974),
236–246.
[42] Novikov S.P., Manakov S.V., Pitaevskĭı L.P., Zakharov V.E., Theory of solitons. The inverse scattering
method, Contemporary Soviet Mathematics, Plenum, New York, 1984.
[43] Pavlov M.V., Elliptic coordinates and multi-Hamiltonian structures of hydrodynamic-type systems, Russian
Acad. Sci. Dokl. Math. 50 (1995), 374–377.
[44] Pavlov M.V., Multi-Hamiltonian structures of the Whitham equations, Russian Acad. Sci. Dokl. Math. 50
(1995), 220–223.
[45] Schmidt W.M., Diophantine approximation, Lecture Notes in Mathematics, Vol. 785, Springer-Verlag,
Berlin – Heidelberg – New York, 1980.
[46] Tsarev S.P., On Poisson bracket and one-dimensional systems of hydrodynamic type, Soviet Math. Dokl. 31
(1985), 488–491.
[47] Vorob’ev Y.M., Dobrokhotov S.Yu., Completeness of the system of eigenfunctions of a nonelliptic operator on
the torus, generated by a Hill operator with a finite-zone potential, Funct. Anal. Appl. 22 (1988), 137–139.
[48] Whitham G.B., A general approach to linear and non-linear dispersive waves using a Lagrangian, J. Fluid
Mech. 22 (1965), 273–283.
[49] Whitham G.B., Linear and nonlinear waves, Pure and Applied Mathematics, Wiley-Interscience, New York,
1974.
[50] Whitham G.B., Non-linear dispersive waves, Proc. R. Soc. Lond. Ser. A 283 (1965), 238–261.
http://dx.doi.org/10.1007/BF01077626
http://dx.doi.org/10.1098/rspa.1966.0142
http://dx.doi.org/10.1070/im1999v063n06ABEH000269
http://arxiv.org/abs/0709.4618
http://dx.doi.org/10.1155/S0161171202106120
http://dx.doi.org/10.1155/S0161171202106120
http://arxiv.org/abs/solv-int/9910011
http://dx.doi.org/10.1063/1.2217648
http://arxiv.org/abs/nlin.SI/0509033
http://dx.doi.org/10.1016/S0167-2789(01)00280-9
http://arxiv.org/abs/nlin.SI/0006030
http://dx.doi.org/10.1007/BF01077037
http://arxiv.org/abs/nlin.SI/0306053
http://dx.doi.org/10.1070/RM1990v045n03ABEH002351
http://dx.doi.org/10.1137/1.9781611970227
http://dx.doi.org/10.1137/1.9781611970227
http://dx.doi.org/10.1070/RM1985v040n04ABEH003615
http://dx.doi.org/10.1007/BF01075697
http://dx.doi.org/10.1007/BF01077607
http://dx.doi.org/10.1017/S0022112065000745
http://dx.doi.org/10.1017/S0022112065000745
http://dx.doi.org/10.1098/rspa.1965.0019
1 Introduction
2 Lagrangian and Hamiltonian formulations of the Whitham method
3 Dubrovin-Novikov bracket and the multiphase case
References
|
| id | nasplib_isofts_kiev_ua-123456789-149185 |
| institution | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| issn | 1815-0659 |
| language | English |
| last_indexed | 2025-11-29T01:49:00Z |
| publishDate | 2012 |
| publisher | Інститут математики НАН України |
| record_format | dspace |
| spelling | Maltsev, A.Y. 2019-02-19T18:22:01Z 2019-02-19T18:22:01Z 2012 Whitham's Method and Dubrovin-Novikov Bracket in Single-Phase and Multiphase Cases / A.Y. Maltsev // Symmetry, Integrability and Geometry: Methods and Applications. — 2012. — Т. 8. — Бібліогр.: 50 назв. — англ. 1815-0659 2010 Mathematics Subject Classification: 37K05; 35B10; 35B15; 35B34; 35L65 DOI: http://dx.doi.org/10.3842/SIGMA.2012.103 https://nasplib.isofts.kiev.ua/handle/123456789/149185 In this paper we examine in detail the procedure of averaging of the local field-theoretic Poisson brackets proposed by B.A. Dubrovin and S.P. Novikov for the method of Whitham. The main attention is paid to the questions of justification and the conditions of applicability of the Dubrovin-Novikov procedure. Separate consideration is given to special features of single-phase and multiphase cases. In particular, one of the main results is the insensitivity of the procedure of bracket averaging to the appearance of ''resonances'' which can arise in the multi-phase situation. This paper is a contribution to the Special Issue “Geometrical Methods in Mathematical Physics”. The full collection is available at http://www.emis.de/journals/SIGMA/GMMP2012.html. In conclusion the author expresses his deep gratitude to Professors S.P. Novikov, B.A. Dubrovin, I.M. Krichever, S.Yu. Dobrokhotov, and M.V. Pavlov for fruitful discussions. This work was financially supported by the Russian Federation Government Grant No. 2010-220-01-077, Grant of the President of Russian Federation NSh-4995.2012.1, and Grant RFBR No. 11-01-12067-ofim-2011. en Інститут математики НАН України Symmetry, Integrability and Geometry: Methods and Applications Whitham's Method and Dubrovin-Novikov Bracket in Single-Phase and Multiphase Cases Article published earlier |
| spellingShingle | Whitham's Method and Dubrovin-Novikov Bracket in Single-Phase and Multiphase Cases Maltsev, A.Y. |
| title | Whitham's Method and Dubrovin-Novikov Bracket in Single-Phase and Multiphase Cases |
| title_full | Whitham's Method and Dubrovin-Novikov Bracket in Single-Phase and Multiphase Cases |
| title_fullStr | Whitham's Method and Dubrovin-Novikov Bracket in Single-Phase and Multiphase Cases |
| title_full_unstemmed | Whitham's Method and Dubrovin-Novikov Bracket in Single-Phase and Multiphase Cases |
| title_short | Whitham's Method and Dubrovin-Novikov Bracket in Single-Phase and Multiphase Cases |
| title_sort | whitham's method and dubrovin-novikov bracket in single-phase and multiphase cases |
| url | https://nasplib.isofts.kiev.ua/handle/123456789/149185 |
| work_keys_str_mv | AT maltsevay whithamsmethodanddubrovinnovikovbracketinsinglephaseandmultiphasecases |