Int J Theor Phys (2010) 49: 884–913

DOI 10.1007/s10773-010-0271-z

On General Solutions for Field Equations in Einstein

and Higher Dimension Gravity

Sergiu I. Vacaru

Received: 2 December 2009 / Accepted: 2 February 2010 / Published online: 2 March 2010
© Springer Science+Business Media, LLC 2010

Abstract We prove that the Einstein equations can be solved in a very general form for ar-
bitrary spacetime dimensions and various types of vacuum and non-vacuum cases following
a geometric method of anholonomic frame deformations for constructing exact solutions in
gravity. The main idea of this method is to introduce on (pseudo) Riemannian manifolds
an alternative (to the Levi-Civita connection) metric compatible linear connection which
is also completely defined by the same metric structure. Such a canonically distinguished
connection is with nontrivial torsion which is induced by some nonholonomy frame coef-
ficients and generic off-diagonal terms of metrics. It is possible to define certain classes of
adapted frames of reference when the Einstein equations for such an alternative connec-
tion transform into a system of partial differential equations which can be integrated in very
general forms. Imposing nonholonomic constraints on generalized metrics and connections
and adapted frames (selecting Levi-Civita configurations), we generate exact solutions in
Einstein gravity and extra dimension generalizations.

Keywords Einstein spaces and higher dimension gravity · Anholonomic frames · Exact
solutions · Nonholonomic manifolds

1 Introduction and Formulation of Main Result

The issue to construct exact solutions in Einstein gravity and high dimensional gravity
theories is not new. It has been posed in different ways and related to various problems
in multidimensional cosmology, black hole physics, nonlinear gravitational effects etc in
Kaluza–Klein gravity and string/brane generalizations. For instance, in brane gravity, one
faces the problem to generate solutions with wrapped configurations and possible quantum

S.I. Vacaru ()

Science Department, University “Al. I. Cuza” Iaşi, 54 Lascar Catargi street, 700107, Iaşi, Romania
e-mail: sergiu.vacaru@uaic.ro
url: http://www.scribd.com/people/view/1455460-sergiu
S.I. Vacaru
e-mail: Sergiu.Vacaru@gmail.com
Int J Theor Phys (2010) 49: 884–913 885

corrections and noncommutative modifications. Former elaborated approaches are model

dependent, usually for metric ansatz depending on 1–2 coordinates, for spherical/cylindrical
symmetries and chosen backgrounds with certain types of asymptotic boundary conditions.
In this context, and further application in modern physics, it is important to find a way to
elaborate methods of constructed exact solutions in very general form.
The problem of constructing most general classes of solutions in gravity has been posed
in a geometric language following the anholonomic deformation/frame method, see reviews
of results in Refs. [1–4].1 In brief, the method allows us to generate any spacetime metric
with prescribed, or reasonable, physical/geometric properties (as solutions of Einstein equa-
tions and modifications for different gravity theories) by performing certain types of non-
holonomic transforms/deformations from another well defined (pseudo) Riemannian met-
rics. Such a problem of constructing “almost general” solutions for Einstein spaces was
solved recently for four and five dimensional spaces [5] but straightforward extensions and
more sophisticate constructions should be provided to achieve such results for spacetimes
of arbitrary dimensions.2
In this paper, we show how the Einstein equations can be integrated in general form for
spacetimes of higher dimensions k n = n + m +1 m + · · · + k m > 5, when n = 2, or 3, and
0
m = m, 1 m, 2 m, . . . = 2; for k = 0, 1, 2, . . . . Such constructions provide a generalization
of the Main Result (Theorem 1) from Ref. [5] proved for dimensions 0 n = n + m = 4, or
5 (see details on geometric methods of constructing exact solutions in gravity in Refs. [1–
4]). The approach developed in this work may present a substantial interest for research in
higher dimensional (super) gravity theories (which, in general, may possess higher order
anisotropies [6–9]) and in higher order Lagrange–Finsler/Hamilton–Cartan geometry and
related gravity and mechanical models [10–15].
Let us consider a (pseudo) Riemannian manifold k V, dim k V = k n, provided with a met-
ric
k kα kβ
g = gk α k β (u γ )du ⊗ du (1)
of arbitrary signature k α = ((1) = ±1, (2) = ±1, . . . , ( n) = ±1). The local coordi-
k 3

nates on k V are parametrized “shell by shell” by increasing dimensions on 2 at every level

1 The geometry of nonholonomic distributions/deformations and frames should be not identified with the
Cartan’s moving frame method even in the first case “moving frames” can be also included. In our approach,
we consider arbitrary real/complex, in general, noncommutative/supersymmetric nonholonomic distributions
on certain manifolds and adapt the geometric constructions with respect to such distributions. Such con-
structions result in (nonlinear) deformations of connection and metric structures, which is not the case for
moving frames, when the same geometric objects are re-expressed with respect to moving/different systems
of reference. Selecting some convenient nonholonomic distributions, we obtain separations of equations and
reparametrizations of variables in some physically important nonlinear systems of partial differential equa-
tions which allows us to integrate such systems in general forms. Then constraining correspondingly some
general solutions, we select necessary subclasses of exact solutions, for instance, in general relativity and
extra dimensions.
2 Some conditions of theorems and formulas, for four and five dimensions, presented in Ref. [5] (that version
should be considered as a Letter version of this work, where we emphasized certain techniques for generating
solutions in general relativity) will be repeated for some our further constructions because they are used for
different type generalizations and simplify proofs for higher dimensions.
3 Notation Remarks: In our works, we follow conventions from [1, 4, 6, 7, 9] when left up/low indices are
used as labels for certain types of geometric spaces/manifolds and objects. The Einstein summation rule is
applied on repeating right left-up indices if it is not stated a contrary condition. Boldfaced letters are used
for spaces (geometric objects) enabled with (adapted to) some nonholonomic distributions/frames prescribed
on corresponding classes of manifolds. An abstract/coordinate index formalism is necessary for deriving in
explicit form some general/exact solutions for gravitational field equations in higher dimensional gravity.
886 Int J Theor Phys (2010) 49: 884–913

(equivalently, shell). We begin with denotations for 0 V, dim 0 V = 0 n, when uα = (x i , y a ),

0 0 
for uα = u α and y a = y a with k = 0, where x i = (x 1 , x i ) and y a = (v, y), i.e. y 4 = v,
y = y. Indices i, j, k, . . . = 1, 2, 3; ı̂, jˆ, k̂, . . . = 2, 3 and a, b, c, . . . = 4, 5 are used for a
5

conventional (3 + 2)-splitting of dimension and general abstract/coordinate indices when

α, β, . . . run values 1, 2, . . . , 5. For four dimensional (in brief, 4–d) constructions, we can

write uα = (x i , y a ), when the coordinate x 1 and values for indices like α, i, . . . = 1 are
not considered. In brief, we shall denote some partial derivatives ∂α = ∂/∂uα in the form
s • = ∂s/∂x 2 , s  = ∂s/∂x 3 , s ∗ = ∂s/∂y 4 . At the next level, 1 V, dim 1 V = 1 n = n + m + 1 m,
1 1
the coordinates are labeled u α = (x i , y a , y a ), for 1 a, 1 b, . . . = 6, 7. For the “2-anisotropy”,
2 1 2
2
V, dim 2 V = 2 n = n + m + 1 m + 2 m, the coordinates are labeled u α = (x i , y a , y a , y a ),
for 2 a, 2 b, . . . = 8, 9; and (recurrently) for the “k-anisotropy”, k V, dim k V = k n, the coordi-
k 1 2 k
nates are labeled u α = (x i , y a , y a , y a , . . . , y a ), for k a, k b, . . . = 4 + 2k, 5 + 2k.4
k
We shall write k ∇ = { kαβ k γ } for the Levi-Civita connection, with coefficients stated
kβ
with respect to an arbitrary local frame basis e k α = (e k−1 α , e k a ) and its dual basis e =
k−1 k k
(e β , e b ). Using the Riemannian curvature tensor k R = {R kαβ k γ k δ } defined by k
∇,
kα
one constructs the Ricci tensor, k Ric = {Rk β k δ  R kβ kα kδ }, and scalar curvature k R 
kβ kδ kβ kδ
g Rk β k δ , where g is inverse to gk αk β . The Einstein equations on k V, for an energy-
k
momentum source Tαβ , are written in the form

1
Rk β k δ − gk β k δ k R = κTk β k δ , (2)
2
where κ = const. For the Einstein spaces defined by a cosmological constant λ, such grav-
k k k
itational field equations can be represented as R αk β = λδ kαβ , where δ kαβ is the Kronecher
symbol. The vacuum solutions are obtained for λ = 0.
The goal of our work is to provide (see further sections) the proof of:

Theorem 1.1 (Main Theorem) The gravitational field equations in the k n-dimensional Ein-
stein gravity (2) represented by frame transforms as
kα kα
R kβ =ϒ kβ , (3)

k
for any given ϒ αk β = diag[ϒ1 , ϒ2 , ϒ2 = ϒ3 , ϒ4 , ϒ5 = ϒ4 , ϒ6 = 1 ϒ2 , ϒ7 = ϒ6 , . . . , ϒ4+2k =
k
ϒ2 , ϒ5+2k = ϒ4+2k ] with

ϒ 1 1 = ϒ1 = ϒ2 + ϒ4 , ϒ 2 2 = ϒ 3 3 = ϒ2 (x k , v),

ϒ 4 4 = ϒ 5 5 = ϒ4 (x k ), ϒ 6 6 = ϒ 7 7 = 1 ϒ2 (uα , 1 v), (4)
1 k−1 α
ϒ 8 8 = ϒ 9 9 = 2 ϒ2 (u α , 2 v), ..., 4+2k = ϒ 5+2k = ϒ2 (u
ϒ 4+2k 5+2k k
, k v),

Unfortunately, such denotations can not be introduced in a more simple form if we aim to present certain
general results on exact solutions derived for some “multi-level” systems of nonlinear partial differential
equations.
4 We use the term anisotropy/anisotropic for some nonholonomically (equivalently, anholonomically) con-
strained variables/coordinates on a (pseudo) Riemannian manifold subjected to certain non-integrable condi-
tions; such anisotropies should be not confused with those when geometric objects depend on some “direc-
tions and velocities”, for instance, in Finsler geometry.
Int J Theor Phys (2010) 49: 884–913 887

for y 4 = v, y 6 = 1 v, y 8 = 2 v, . . . , y 4+2k = k v (where k labels the shell’s number), can be

solved in general form by metrics of type
  
k
g = 1 dx 1 ⊗ dx 1 + gi (x k )dx i ⊗ dx i + ω2 (x j , y b )ha (x k , v)ea ⊗ea
1 1 1a
+ 1 ω2 (uα , y b )h1 a (uα , 1 v) e a ⊗ e
1 2 1 2 2
+ 2 ω2 (u α , y b )h2 a (u α , 2 v)e a ⊗ e a + · · ·
k−1 α k k−1 α k k
+ k ω2 (u , y b )h2 a (u , k v)e a ⊗ e a , (5)

for

e4 = dy 4 + wi (x k , v)dx i , e5 = dy 5 + ni (x k , v)dx i ,
e6 = dy 6 + wβ (uα , 1 v)duβ , e7 = dy 7 + nβ (uα , 1 v)duβ ,
1 1 1 1
e8 = dy 8 + w1 β (u α , 2 v)du β , e9 = dy 9 + n1 β (u α , 2 v)du β ,
...
k−1 α k−1 β
e4+2k
= dy 4+2k + wk−1 β (u , k v)du ,
k−1 α k−1 β
e5+2k = dy 5+2k + nk−1 β (u , k v)du ,
k−1
where coefficients are defined by generating functions f (x i , v), ∂f/∂v = 0, . . . , k f (u α ,
k−1 k
k
v), ∂ k f/∂ k v = 0 and ω(x j , y b ), . . . , k ω(u α , y b ) = 0 and integration functions
k−1 k−1 k−1
0
f (x i ), . . . , 0k f (u α ), 0 h(x i ), . . . , 0k h(u α ), 1 nj (x i ), . . . , nk−1 β (u α ), 2 nj (x i ), . . . ,
2 k−1 α
k−1 β nj (u ) following recurrent formulas (when a next “shell” extends in a compatible
form the previous ones; i.e. containing the previous constructions), being computed as

for 2 ψ •• + 3 ψ  = ϒ4 ;
k
gi = i eψ(x ) ,
(6)
h4 = 4 0 h(x i )[∂v f (x i , v)]2 |ς (x i , v)|, h5 = 5 [f (x i , v) − 0 f (x i )]2 ;

wi = −∂i ς (x i , v)/∂v ς (x i , v),


nk = 1 nk (x i ) + 2 nk (x i ) dv ς (x i , v)

[∂v f (x i , v)]2 /[f (x i , v) − 0 f (x i )]3 ,


4 0
for ς = ς (x ) −
0 i
h(x ) dv ϒ2 (x k , v)
i
8
∂v f (x i , v)[f (x i , v) − 0 f (x i )];
h6 = 6 01 h(uα )[∂1 v 1 f (uα , 1 v)]2 | 1 ς (uα , 1 v)|,
h7 = 7 [1 f (uα , 1 v) − 01 f (uα )]2 ;
wβ = −∂β 1 ς (uα , 1 v)/∂1 v 1 ς (uα , 1 v),

nβ = 1 nβ (uα ) + 2 nβ (uα ) d 1 v 1 ς (uα , 1 v)
888 Int J Theor Phys (2010) 49: 884–913

[∂1 v 1 f (uα , 1 v)]2 /[1 f (uα , 1 v) − 01 f (uα )]3 ,


6 0
for 1 ς = 01 ς (uα ) − α
1 h(u ) d 1 v 1 ϒ2 (uα , 1 v)
8
[∂1 v 1 f (uα , 1 v)][1 f (uα , 1 v) − 01 f (uα )];
1 1 1
h8 = 8 02 h(u α ) [∂2 v 2 f (u α , 2 v)]2 | 2 ς (u α , 2 v)|,
1 1
h9 = 9 [2 f (u α , 2 v) − 02 f (u α )]2 ;
1 1
w1 β = −∂1 β 2 ς (u α , 2 v)/∂2 v 2 ς (u α , 2 v),

1 1 1
n1 β = 1 n1 β (u α ) + 2 n1 β (u α ) d 2 v 2 ς (u α , 2 v)
1 1 1
[∂2 v 2 f (u α , 2 v)]2 /[2 f (u α , 2 v) − 02 f (u α )]3 ,

1 8 0 1α 1
for 2 ς = 02 ς (u α ) − 1 h(u ) d 2 v 2 ϒ2 (u α , 2 v)
8
1 1 1
[∂2 v 2 f (u α , 2 v)][2 f (u α , 2 v) − 02 f (u α )];
...
k−1 α k−1 α k−1 α
h4+2k = 4+2k 0k h(u )[∂k v k f (u , k v)]2 |k ς (u , k v)|,
k−1 α k−1 α
h5+2k = 5+2k [k f (u , k v) − 0k f (u )]2 ;
k−1 α k−1
wk−1 β = −∂k−1 β k ς (u , k v)/∂k v k ς (u α , k v),

k−1 α k−1 α k−1
nk−1 β = nk−1 β (u
1
) + nk−1 β (u
2
) d k v k ς (u α , k v)
k−1 α k−1 α k−1 α
[∂k v k f (u , k v)]2 /[k f (u , k v) − 0k f (u )]3 ,

k−1 α 4+2k 0 k−1 k−1 α
for k ς = 0k ς (u )− h(u α ) d k v k ϒ2 (u , k v)
8 k
k−1 α k−1 α k−1 α
[∂k v k f (u , k v)][k f (u , k v) − 0k f (u )];

and

ek ω = ∂k ω + wk ∂v ω + nk ∂ω/∂y 5 = 0,
eα 1 ω = ∂α 1 ω + wα ∂ 1 ω/∂ 1 v + nα ∂ 1 ω/∂y 5 = 0,
e1 α 2 ω = ∂1 α 2 ω + w1 α ∂ 2 ω/∂ 2 v + n1 α ∂ 2 ω/∂y 7 = 0, (7)
...
ek−1 α ω = ∂k−1 α k ω + wk−1 α ∂ k ω/∂ k v + nk−1 α ∂ k ω/∂y 5+2k = 0,
k

when the solutions for the Levi-Civita connection are selected by additional constraints

∂v wi = ei ln |h4 |, ek wi = ei wk , ∂v ni = 0, ∂i n k = ∂k n i ;
∂1 v wα = eα ln |h6 |, eα wβ = eβ wα , ∂1 v nα = 0, ∂α n β = ∂β n α ;
Int J Theor Phys (2010) 49: 884–913 889

∂2 v w1 α = e1 α ln |h8 |, e1 α w1 β = e1 β w1 α ,
∂2 v n1 α = 0, ∂1 α n 1 β = ∂1 β n 1 α ; (8)
...
∂k v wk−1 α = e1 α ln |h4+2k |, ek−1 α wk−1 β = ek−1 β wk−1 α ,
∂k v nk−1 α = 0, ∂k−1 α nk−1 β = ∂k−1 β nk−1 α .

Following above Theorem,5 we express the solutions of Einstein equations in high di-
mensional gravity in a most general form, presenting in formulas all classes of generating
and integration functions and stating all constraints selecting Einstein spaces. We choose a
“two by two” increasing of spacetime dimensions in formulas because this provides us a
simplest way for generating “non-Killing” solutions characterized by certain types of two
dimensional conformal factors depending on two “anisotropic” coordinates and the rest ones
being considered as parameters. This allows us to associate to the constructed classes of so-
lutions certain hierarchies of two-dimensional conformal symmetries with corresponding
invariants and to derive associated solitonic hierarchies and bi-Hamiltonian structures as
we elaborated for four dimensional spaces in Refs. [16, 17]. The length of this article does
not give us a possibility to present such nonlinear wave developments for high dimensional
gravity.
We have to state certain boundary/symmetry/topology conditions and define in explicit
form the integration functions and systems of first order partial differential equations of
type (8). This is necessary when we are interested to construct some explicit classes of exact
solutions of Einstein equations (3) which are related to some physically important four and
higher dimensional metrics. For instance, such high dimension solutions can be constructed
to contain wormholes [18, 19] and/or to model (non) holonomic Ricci flows of various types
of gravitational solitonic pp-wave, ellipsoid etc. configurations [20–24]. Perhaps all classes
of exact solutions presented in the above mentioned references and (for instance, reviewed
in) Refs. [1–4, 28, 29] can be found as certain particular cases of metrics (5) or certain
equivalently redefined ones. In this article, we shall emphasize the geometric background of
the anholonomic deformation method for construction high dimensional exact solutions in
gravity and refer readers to cited works, for details and physical applications.
k 
Any (pseudo) Riemannian metric gk α k β  (u γ ) (1) depending in general on all 5 + 2k
k
local coordinates on V can be parametrized in a form gk α k β (u γ ) (5) using transforms of
k k  k β kγ 
coefficients of metric gk α k β (u γ ) = e kαα e k β gk α k β (u ) under vielbein transforms ek α =
k α 6
e e preserving a chosen shell structure. If the coefficients of such metrics satisfy the
k α k α
conditions of Main Theorem, they define general solutions of Einstein equations for any type
of sources which can be parametrized in a formally diagonalized (with shell conditions)
form (4), with respect to certain classes of nonholonomic frames of reference. By frame
transforms such parametrizations can be defined for various types of physically important

5 A similar Theorem was formulated in Ref. [5] for four and five dimensions; some formulas and conditions
have to be repeated in this work because they are used for high dimension generalizations.
6 We have to solve certain systems of quadratic algebraic equations and define some ek α  (uk β ), for given
kα
k  k  k
coefficients of any (5) and (1), choosing a convenient system of coordinates u α = u α (u β ); to present
the a “shell structure” is convenient for purposes to simplify proofs of theorems; in general, under arbitrary
frame/coordinate transform, all “shells” mix each with others.
890 Int J Theor Phys (2010) 49: 884–913

energy-momentum tensors, cosmological constants (in general, with anisotropic polarizia-

tions), and for vacuum configurations.
For the case k = 0, the proof of Theorem 1.1 is outlined in Ref. [5] using the anholonomic
deformation method which was originally proposed in Refs. [25–27]. There were published
a series of reviews and generalizations of the method, see [1–4], when the solutions of
gravitational field equations in different types of commutative and noncommutative gravity
and Ricci flow theories contain at least one Killing vector symmetry. Such higher dimension
Einstein metrics with Killing symmetries, are generated if in the conditions of Theorem 1.1
there are considered ω, 1 ω, . . . , k ω = 1.
Summarizing the results provided in Sects. 2–4 (they may be considered also as a review
for the ‘higher dimension’ version of the anholonomic deformation method of constructing
exact solutions in gravity7 ), we get a proof of Main Theorem 1.1.
Finally, we note that even we shall present a number of key results and some technical
details, we shall not repeat explicit computations for coefficients of tensors and connections
presented for the “Killing case” k = 0 in our previous works [30–32], see also reviews and
generalizations in [1, 2].

2 Higher Order Nonholonomic Manifolds

In this section, we outline the geometry of higher order nonholonomic manifolds which, for
simplicity, will be modelled as (pseudo) Riemannian manifolds with higher order “shell”
structure of dimensions k n = n + m + 1 m + · · · + k m. Certain geometric ideas and construc-
tions originate from the geometry of higher order Lagrange–Finsler and Hamilton–Cartan
spaces defined on higher order (co) tangent classical and quantum bundles [10–13]. Such
nonholonomic structures were investigated for models of (super) strings in higher order
anisotropic (super) spaces [6] and for anholonomic higher order Clifford/spinor bundles [7–
9]. In explicit form, the (super) gravitational gauge field equations and conservations laws
were analyzed in Refs. [33–36].

2.1 Higher Order N-adapted Frames and Metrics

Our geometric spacetime arena is defined by (pseudo) Riemannian manifolds k V enabled

with nonholonomic distributions (which can be prescribed in any convenient for our pur-
poses form like we can fix any system of reference/coordinates).

Definition 2.1 A manifold k V, dim k V = k n, is higher order nonholonomic (equivalently,

k-anholonomic) if its tangent bundle T k V is enabled with a Whitney sum of type

T k V = h k V ⊕ v k V ⊕ 1 v k V ⊕ 2 v k V ⊕ · · · ⊕ k v k V. (9)

For k = 0, we get a usual nonholonomic manifold (or, in this case, N-anholonomic) enabled
with nonlinear connection (N-connection) structure. We say that a distribution (9) defines a
higher order N-connection (equivalently, k N-connection) structure.

7 This work is organized as following: In Sect. 2, there are outlined some necessary results from the geometry
of nigher order nonholonomic manifolds. Section 3 contains the system of partial differential equations (PDE)
to which the Einstein equations can be transformed under nonholonomic frame transforms and deformations.
In Sect. 4, we prove that it is possible to construct very general classes of exact solutions with Killing sym-
metries for such PDE in high dimensional gravity. We also show that it is possible to generate “non-Killing”
solutions in most general forms if we consider nonholonomically deformed conformal symmetries. We con-
clude and discuss the results in Sect. 5.
Int J Theor Phys (2010) 49: 884–913 891

1 2 k
Locally, a k N-connection is defined by its coefficients k N = {Nia , Nα a , N1 αa , . . . , Nk−1a α },
1a 2a ka
with {Nia } ⊂ {Nα } ⊂ {N } ⊂ · · · ⊂ N 1α k−1 α }, when

∂ 1 1 ∂
0
N = Nia (uα )dx i ⊗ , 1
N = Nβ a (u α )duβ ⊗ ,
∂y a ∂y 1 a
2 2 1β ∂ k k k−1 β ∂
2
N = N1 βa (u α )du ⊗ , ..., k
N = Nk−1
a
(u α )du ⊗ .
∂y 2 a β
∂y k a

It should be noted that for general coordinate transforms on k V, there is a mixing of

coefficients and coordinates.8 For simplicity, we can work with adapted coordinates when
some sets of coordinates on a shell of lower order are contained in a subset of coordinates
k−s k−s+1 α k−s k−s+1 a
on shells of higher order by trivial extensions like u α → u = (u α , y ).

Proposition 2.1 There is a class of N-adapted frames and dual (co-) frames (equivalently,
vielbeins) on k V which depend linearly on coefficients of k N-connection.

Proof We construct such frames following recurrent formulas for k = 0, 1, . . . , when

 
∂ ka ∂ ∂
ek ν = ek−1 v = k−1 − Nk−1 v k a , ek a = ∂k a = k a , (10)
∂u v ∂y ∂y
k  k−1 k−1 k k ka k−1 
e μ = e μ = du μ , e a = dy a + Nk−1 v
du v . (11)

The vielbeins (10) satisfy the nonholonomy relations

kγ
[ek α , ek β ] = ek α ek β − ek β ek α = wk α k β ek γ (12)

k k ka
b
with (antisymmetric) nontrivial anholonomy coefficients wk−1 α ka
= ∂Nk−1b α /∂u and
kb kb
w k−1 α k−1 β = k−1 α k−1 β , where

kb  kb   kb 
k−1 α k−1 β
= ek−1 β Nk−1 α
− ek−1 α Nk−1 β
(13)

are the coefficients of curvature k  of N-connection k N. The particular holonomic/integrable

kb
case is selected by the integrability conditions wk−1 α ka
= 0.
Any (pseudo) Riemannian metric g on k V can be written in N-adapted form, we shall
write in brief that g = k g = {gk β k γ }, for
k−1 α k−1 β k−1 γ k k kb
k
g = gk−1 β k−1 γ (u )e ⊗e + hk a k b (u α )e a ⊗ e
1 1 1b
= gij (x k )ei ⊗ ej + hab (uα )ea ⊗ eb + h1 a 1 b (u α )e a ⊗ e
2 2 2 k k k
+h2 a 2 b (u α )e a ⊗ e b + · · · + h k a k b (u α )e a ⊗ e b , (14)

8 We use boldface symbols for spaces (and geometric objects on such spaces) enabled with a structure of
N-coefficients.
892 Int J Theor Phys (2010) 49: 884–913

k
for some N-adapted coefficients gk β k γ = [gij , hab , h1 a 1 b , . . . , hk a k b ] and Nk−1
a
α
. For con-
structing exact solutions in high dimensional gravity, it is convenient to work with such
N-adapted formulas for tensors’ and connections’ coefficients.
For instance, we get from (14) a metric with a parametrization of type (5) when all k ω = 1
if we choose

gij = diag[1 , gi (x k )], hab = diag[ha (x i , v)],
Nk4 = wk (x i , v), Nk5 = nk (x i , v);
h1 a 1 b = diag[h1 a (uα , 1 v)],
Nβ6 = wβ (uα , 1 v), Nβ7 = nβ (uα , 1 v);
1
h2 a 2 b = diag[h2 a (u α , 2 v)], (15)
1 1
N 81 β = w1 β (u α , 2 v), N 91 β = n1 β (u α , 2 v);
...
k−1 α
hk a k b = diag[hk a (u , k v)],
k−1 α k−1 α
k−1 β = wk−1 β (u
N 4+2k , k v), k−1 β = nk−1 β (u
N 5+2k , k v).

Such a metric has symmetries of k + 1 Killing vectors, e5 = ∂/∂y 5 , e7 = ∂/∂y 7 , . . . , e5+2k =

∂/∂y 5+2k , because its coefficients do not depend on y 5 , y 7 , . . . , y 5+2k . Introducing nontrivial
k 2 kα
ω (u ) depending also on y 5+2k , as multiples before hk a , we get N-adapted parametriza-
tions, up to certain frame/coordinate transforms, for all metrics on k V. In Sect. 3, we shall
define the equations which must satisfy the coefficients of N-adapted metrics when (15) will
generate exact solutions of Einstein equations.

2.2 N-adapted Deformations of the Levi-Civita Connection

By straightforward computations, it is a cumbersome task to prove using the Levi-Civita

connection9 k ∇) that the Einstein equations (3) on higher dimensional spacetimes are solved
by metrics of type (5). We are going to show explicitly that general solutions for k ∇ can be
constructed passing three steps: (1) to adapt our constructions to N-adapted frames of type
eα (10) and eμ (11); (2) to use as an auxiliary tool (we emphasize, in Einstein gravity and its
generalizations on high dimensional (pseudo) Riemannian manifolds) a new type of linear
connection k D = {ˆ k βα k γ }, also uniquely defined by the metric structure; (3) To constrain
k

the integral varieties of general solutions in such a form that k 

D → k ∇.

Definition 2.2 A distinguished connection k D (in brief, d-connection) on k V is a linear

connection preserving under parallelism a conventional horizontal and k-vertical splitting
(in brief, h- and v-splitting) induced by k N-connection structure (9).
k k
We note that the Levi-Civita connection k ∇, for which k ∇ k g = 0 and k T α  k ∇e α = 0,
is not a d-connection because, in general, it is not adapted to a N-splitting defined by a Whit-
ney sum (9). So, in order to elaborate self-consistent geometric/physical models adapted to
a N-connection it is necessary to work with d-connections.

9 A unique one, which is metric compatible and with zero torsion, and completely defined by the metric
structure.
Int J Theor Phys (2010) 49: 884–913 893

Theorem 2.1 There is a unique canonical d-connection k 

D satisfying the condition k  D kg =
0 and with vanishing “pure” horizontal and vertical torsion coefficients, i.e. Tj k = 0 and
i

T ka k = 0, see (below) formulas (20).

k
b c

Proof Let us define k 

D as a 1-form
kα

 kβ =

kα
kβ kγ
kγ
e (16)

kγ
with  ij k , L
abk , C
ji c , C
bc
a  a α 1 a 1 ; . . . ; L
k a k−1 , C
k−1 α k , C
k a k ), where
1 1 k k−1 k
 kα kβ = (L ; L1 bα , Cβ 1 c , C b c b α β c b c
k−1
11 α 1 = 
L α
βγ ij k , L
= (L abk , C
ji c , C
bc
a 22 α 2 = 
); L
1
 kk α k = 
1 βα 1 γ ; . . . ; L
α
 k−1 β k−1 γ , for
β γ β γ β γ

1
gαβ = [gij (x k ), hab (uγ )], g1 α 1 β = [gαβ (uγ ), h1 a 1 b (u γ )],
1 2
g2 α 2 β = [g1 α 1 β (u γ ), h2 a 2 b (u γ )], . . . ,
k−1 γ k
gk α k β = [gk−1 α k−1 β (u ), hk a k b (u γ )],

where

ij k = 1 g ir (ek gj r + ej gkr − er gj k ),

L
2
 
abk = eb (Nka ) + 1 hac ek hbc − hdc eb Nkd − hdb ec Nkd ,
L
2
1
ji c = g ik ec gj k , bc 1
C C a
= had (ec hbd + ec hcd − ed hbc ),
2 2
11 a = e1 b (Nα1 a ) + 1 h1 a 1 c (eα h1 b 1 c − h1 d 1 c e1 b Nα1 d − h1 d 1 b e1 c Nα1 d ),
L bα 2
1
α 1 = g αγ e1 c gβγ ,
C β c
2
1
1 1 = h1 a 1 d (e1 c h1 b 1 d + e1 c h1 c 1 d − e1 d h1 b 1 c ),
1a
C b c 2
22 a 1 = e2 b (N 21a ) + 1 h2 a 2 c (e1 α h2 b 2 c − h2 d 2 c e2 b N 2 d1 − h2 d 2 b e2 c N 2 d1 ),
L (17)
b α α 2 α α

C11 α 2 = 1 g 1 α 1 γ e2 c g1 β 1 γ ,
β c 2
1
2 2 = h2 a 2 d (e2 c h2 b 2 d + e2 c h2 c 2 d − e2 d h2 b 2 c ),
2a
C b c 2
...
kk a k−1 = ek b (Nk−1ka 1 k k k k
L b α α
) + h a c (ek−1 α hk b k c − hk d k c ek b N dk−1 α − hk d k b ek c N dk−1 α ),
2
k−1
k−1 α 1 k−1 k−1
C β kc
= g α γ ek c gk−1 β k−1 γ ,
2
1
k k = hk a k d (ek c hk b k d + ek c hk c k d − ek d hk b k c ).
ka
C b c
2
894 Int J Theor Phys (2010) 49: 884–913

It follows by straightforward verifications that k 

D k g = 0, where this N-adapted metric com-
patibility condition splits into
j gkl = 0,
D a gkl = 0,
D j hab = 0,
D a hbc = 0,
D
γ gαβ = 0,
D 1 a gαβ = 0,
D γ h1 a 1 b = 0,
D 1 a h1 b 1 c = 0,
D
1 γ g1 α 1 β = 0,
D 2 a g1 α 1 β = 0,
D 1 γ h2 a 2 b = 0,
D 2 a h2 b 2 c = 0,
D
(18)
...
k−1 γ gk−1 α k−1 β = 0,
D k a gk−1 α k−1 β = 0,
D
k−1 γ hk a k b = 0,
D k a hk b k c = 0,
D

where the covariant derivatives are computed using corresponding coefficients, step by step,
on every shell. The canonical d-connection contains an induced torsion (completely deter-
mined by the coefficients of metric, and respective N-connection coefficients)

 =  k
De α = de α + 
kα kα kβ kγ k k k k
T T kβ kγ e ∧e  αk β ∧ e β , (19)

with coefficients
Tij k = L
i j k − L
i kj , Tij a = −Tiaj = C
ij a , T aj i = −aj i ,
∂Nia a
Tabi = − Taib = − L bi , Tabc = C
abc − C
acb ;
∂y b
Tαβγ = L
α βγ − L
α γβ , Tαβ 1 a = −Tα1 aβ = C
α 1 , 1 1
β a
T βαa = − βαa ,
1
∂Nα a  1 a
T 1abα = − Tα a1 b = T1 ba 1 c = C
1 a1 1 − C
1 a1 1 ;
1 1 1
− L 1 bα ,
∂y 1 b b c c b

 1 α1 1 − L
T 1αβ 1 γ = L  11α T 1αβ 2 a = −T2 aα 1 β = C
 11α 2 ,
1 1 1
β γ γ 1β , β a
2
∂N1 αa
T 2ab 1 α = −T 1aα 2 b =  a2 1 ,
2a 2a 2 2 2
T 1β 1α = − 1β 1α , −L
∂y 2b b α (20)

2 a2 2 − C
T2 ba 2 c = C 2 a2 2 ;
2
b c c b

...

T k−1αβ k−1 γ = L
k−1 k−1 α
k−1
k−1 β k−1 γ − L k−1 γ
α
k−1 β ,

T k−1αβ k a = − Tk a k−1 k−1

k−1 k−1 α ka ka
β
=C α
k−1 β k a , T k−1 β k−1 α = − k−1 β k−1 α ,
ka
∂N
T kab k−1 α = − T k−1  ak k−1 ,
k ka k−1 α k
α kb
= kb
−L b α
∂y
T  
ka ka ka
kb kc =C kb kc −C kc kb .

Introducing values (17) into (20) we get that Tij k = 0 and C

k ak k = 0 which satisfy the
b c
conditions of this theorem. In general, other N-adapted torsion coefficients (for instance,
Tij a , Taj i and Tabi ) are not zero. 
Int J Theor Phys (2010) 49: 884–913 895

The torsion (19) is very different from that, for instance, in Einstein–Cartan, string, or
gauge gravity because we do not consider additional field equations (algebraic or dynamical
ones), see discussions in [2, 3]. In our case, the nontrivial torsion coefficients are related to
kγ
anholonomy coefficients wk α k β in (12).
We can distinguish the covariant derivative k  D determined by formulas (16) and (17), in
N-adapted to (9) form, as Dk α = (Di , D
a , D
1 a , . . . , D
k a ), where D
k a are shell operators.
From Theorem 2.1, we get:

Corollary 2.1 Any geometric construction for the canonical d-connection k 

kγ
D = { kα kβ }
can be re-defined equivalently into a similar one with the Levi-Civita connection k ∇ =
kγ
{ k α k β } following formulas

kγ kγ kγ
 kα kβ =
 kα kβ +Z kα kβ , (21)

kγ
,
kγ
where the N-adapted coefficients of linear connections,  kα kβ  kα kβ , and the distortion
kγ
tensor Z kα kβ are determined in unique forms by the coefficients of a metric gk α k β .

Proof It is similar to that presented for vector bundles in Refs. [14, 15] but in our case
adapted for (pseudo) Riemannian nonholonomic manifolds, see details in [1, 3, 4] and, in
higher order form, in [6, 7, 9]. Here we present the N-adapted components of the distortion
kγ
tensor Z k α k β computed as

ji b gik hab − 1 ajk ,

Zjak = −C i
Zbk
1
= cj k hcb g j i − ih j
j k Chb ,
2 2
c 1 j
a
Zbk = + ab
cd Tkb ,
i
Zkb = ajk hcb g j i + ihj k Chb , Zji k = 0,
2

c g ij  c
Zjab = −− ad cb Tj d ,
a
Zbc = 0, i
Zab =− Tj a hcb + Tjcb hca ,
2
1a
α1 gαγ h a b − βγa ,
1 1 1 1 1 1
β
Zβγ = −C β b
Z α1 bγ = βγc h1 c 1 b g βα − ατ βγ Cτ 1 b ,
2 2
1 1
= + 1 ac 1 db Tγ 1cb , β
1a 1 1 1
Z1 bγ Zβα 1 b = βγa h1 c 1 b g βα + ατ βγ Cτ 1 b ,
2
Zβ a1 b = −− 1 ac 1 bd Tβ 1cd ,
1 1 1 1 1
α
Zβγ = 0, Z1 ba 1 c = 0,

g αβ   1 c
Tβ 1 a h1 c 1 b + Tβ 1cb h1 c 1 a ,
1
Z α1 a 1 b = −
2

Z
2a
1β 1γ = −C11 α 2 g1 α 1 γ h2 a 2 b − 1 21 a 1 , 1
54j = 1 ∂v nj ,
Z2 bα 1 γ = L
β b
2 β γ 2
1 2c 1 β 2 , Z2 ba 1 γ = + 2 ac 2 db T1 γc 2 b ,
1 1 1 1 1 2 2 2 2
1 1 h2 2 g β α − 1 αβ 1 τγ C
2 β γ c b τ b

1 1 2 1 1 1 1
1 β 2 ,
1
Z1 βα 2 b = 1 aβ 1 γ h2 c 2 b g β α + 1 αβ 1 τγ C τ b
2
Z1 βa 2 b = −− 2 ac 2 bd T1 βc 2 d ,
1 2 2 2 2 2
Z1 βα 1 γ = 0, Z2 ba 2 c = 0, (22)
896 Int J Theor Phys (2010) 49: 884–913


1α 1β
g
T1 βc 2 a h2 c 2 b + T1 βc 2 b h2 c 2 a ,
1 2 2
Z2 aα 2 b = −
2
...
ka
k−1 α k gk−1 α k−1 γ h
k−1 ka kb 1 ka
Z k−1 β k−1 γ = −C β b
− k−1 β k−1 γ
,
2
k−1 α 1 kc k−1 k−1 k−1 k−1
k−1 β k ,
k−1
Zk b k−1 γ
= k−1 β k−1 γ
hk c k b g β α − k−1 αβ k−1 τγ C τ b
2
Zk ba k−1 γ = + k ac k db Tk−1c γ
k k k k
kb ,
k−1 1 ka k−1 β k−1 α k−1 k−1
k−1 β k ,
k−1
Zk−1 βα k b = k−1 β k−1 γ hk c k b g + k−1 αβ k−1 τγ C τ b
2
= −− k ac k bd Tk−1c β k d ,
k−1 k k k k k
Zk−1 βα k−1 γ = 0, a
Zk−1 β kb
Zk ba k c = 0,

 kc
k−1 α k−1 β
g
Tk−1 β k a hk c k b + Tk−1c β k b hk c k a ,
k−1 k
Zk a kαb = −
2
1 ± ab 1
for ih j k = (δj δk − gj k g ), cd = (δca δdb ± hcd hab ),
i h ih
2 2
± 1a 1b 1 1 1 1 1
1 c 1 d = (δ1 ca δ1 db ± h1 c 1 d h a b ), . . . ,
2
± ka kb 1 k k k k
k c k d = (δk ca δk db ± hk c k d h a b ),
2
where the necessary torsion coefficients are computed as in (20). 

Remark 2.1 Hereafter, we shall omit certain details on shell components of formulas and
computations if that will not result in ambiguities. Such constructions are similar to those
presented in above Theorems and in Refs. [1–4, 6–9, 30–32, 34–36]. Some additional nec-
essary formulas are given in Appendix.

In four dimensions, the Einstein gravity can be equivalently formulated in the so-called
almost Kähler and Lagrange–Finsler variables, as we considered in Refs. [4, 37–39]. Sim-
ilarly, for higher dimensions, we can use the canonical d-connection k  D and its nonholo-
nomic deformations for equivalent reformulations of extra dimension gravity theories and as
tools for generating constructing exact solutions. In particular, imposing necessary type con-
straints, it is possible to generate exact solutions of the Einstein equations for the Levi-Civita
connection k ∇.

3 N-adapted Einstein Equations

In this section, we define the Riemannian, Ricci and Einstein tensors for the canonical
d-connection k 
D and metric k g (14) and derive the corresponding gravitational field equa-
tions. We also formulate the general conditions when the Einstein tensor for k 
D is equal to
that for k ∇.

3.1 Curvature of the Canonical d-connection

As for any linear connection, we can introduce:

Int J Theor Phys (2010) 49: 884–913 897

Definition 3.1 The curvature of  

D is a 2-form R D
 = d
 −
 ∧
.

In explicit form, the N-adapted coefficients can be computed using the 1-form (16),
kα kα kα

D = d ∧ =
kγ
 −
kα kα kγ k
R kβ  kβ  kβ  kβ  kγ R kβ kγ kτ e ∧ e τ, (23)

The N-adapted coefficients of curvature are parametrized in the form:

k
R αk β k γ kτ
αβγ τ = {R
= R ihj k , R
abj k , R
ij ka , R
cbka , R
ij bc , R
abcd };

1 α1 1
R 1τ
αβγ τ , R
= {R 11a , R α 1a
 c1
,R
1
1a
α 1 1 , R
,R 1 a1 1 1 };
β γ bγ τ βγ bγ β b c b c d

2 α2 2
R 2τ
 α1 1
= {R
1
1τ
 2a 1
,R
2
1τ
 α1 1
,R
1
2a
 c2 1
,R
2
2a ,
β γ β γ b γ β γ b γ

1 α1 2 2 , R
R 2 a2 2 2 };
β b c b c d

...
k αk k
R kτ
k−1
= {R α
k−1 β k−1 γ k−1 τ
 ka k−1
,R
k
k−1 τ

,R
k−1 α
k−1 β k−1 γ k a ,
β γ b γ


R
kc
k b k−1 γ k a

,R
k−1 α
k−1 β k b k c

,R
ka
kb kc kd }, (24)

where the values of such coefficients are provided in Appendix A, see Theorem A.1 and
formulas (51).

Definition 3.2 The Ricci tensor Ric(D) = {Rαβ } of a canonical d-connection  D is defined
Rαβγ δ (23), when 
by contracting respectively the N-adapted coefficients of  Rαβ  
Rταβτ .

We formulate:

Corollary 3.1 The Ricci tensor of 

D is characterized by N-adapted coefficients

 ij , R
Rk α k β = { R ia , R
ai , R
ab ; R
αβ , R
α 1 a , R
1 aβ , R
1 a 1 b ;

R 1 α 2 a , R
1 α 1 β , R 2 a 1 β , R
2 a 2 b ; . . . ;
k−1 α k−1 β , R
R k−1 α k a , R
 k a k−1 β , R
k a k b }, (25)

where

R kij k ,
ij  R ia  −R
R kika , ai  R
R baib , ab  R
R cabc ;

αβ  R
R γαβγ , α 1 a  −R
R γ 1a , 1 aα  R
R 1 b1 1 ,
αγ aα b

1 a 1 b  R
R 1 c1 1 1 ;
a b c
1 1
R  γ1 1 1 ,
1 α 1 β  R 1 α 2 a  −R
R  γ1 1 2a ,
α β γ α γ

2 b2 1 2 ,
2 a 1 α  R  2a 2b  R
2 c2 2 2 ; (26)
R a α b
R a b c

...
898 Int J Theor Phys (2010) 49: 884–913

k−1 α k−1 β  R
  k−1 α k a  −R

k−1 γ k−1 γ
R k−1 α k−1 β k−1 γ , R k−1 α k−1 γ k a ,

R k bk k−1 k ,
 k a k−1 α  R  ka kb  R
R k ck k k .
a α b a b c

Proof The formulas (26) follow from contractions of (24). To compute the N-adapted coef-
ficients of the Ricci tensor Ric(
D) for metric k g (14) we have to construct correspondingly
the formulas (51). 

 of 
Definition 3.3 The scalar curvature s R D is by definition

s
Rg
kα kβ

Rk α k β
ij + hab R
= g ij R ab + h 1a 1b
1 a 1 b + · · · + hk a k b R
R k a k b . (27)

E k α k β of 
Using values (26) and (27), we can compute the Einstein tensor  D,

 1
Ek α k β   
Rk α k β − gk α k β s R. (28)
2
In explicit form, for N-adapted coefficients, we get a proof for

Corollary 3.2 The Einstein tensor  Ek α k β splits into h- and k v-components  Ek α k β 

ij = R
{E ij − gij R,
1 s E ia = R ia , E
ai = R ai , Eab = R ab − 1 hab s R;  E α 1 a = R α 1 a , E
1 aβ =
2 2
1 aβ , E
R  1a 1b = R  E
1 a 1 b − 1 h1 a 1 b s R; 1 α 2 a = R 1α 2a , E 2a 1β = R  2a 1β , E  2a 2b =
2
R E
 2 a 2 b − 1 h 2 a 2 b s R; k−1 α k a = Rk−1 α k a , E
k a k−1 β = R
k a k−1 β , E
k a k b = R 
k a k b − 1 hk a k b s R}.
2 2

In different theories of (string/brane/gauge etc.) gravity, we can consider nonholonomi-

cally modified gravitational field equations


Ek α k β = κ 
Tk α k β , (29)

for a source 
Tk α k β defined by certain classical or quantum corrections and/or constraints on
dynamics of fields to usual energy-momentum tensors. Such equations are not equivalent,
in general, to the usual Einstein equations (2) for the Levi-Civita connection k ∇.10

Condition 3.1 A class of metrics k g (14) defining solutions of the gravitational field equa-
tions the canonical d-connection (29) are also solutions for the Einstein equations for the
Levi-Civita connection (2), if with respect to certain N-adapted frames (10) and (11) there
are satisfied the conditions

ji b = 0,
C aj i = 0, Tjca = 0; α 1 = 0,
C
1
βαa = 0, Tβ 1ca = 0;
1
β b

11 α 2 = 0,
C 
2a
1β 1α = 0, T1 βc 2 a = 0; . . . ;
2
(30)
β b

k−1
C
k−1 α
= 0, 
ka
= 0, Tk−1c β k a = 0;
k
β kb k−1 β k−1 α

10 As we noted in Refs. [4, 37–39], an equivalence of both types of filed equations would be possible, for

instance, if we introduce a generalized source 

Tk β k δ containing contributions of the distortion tensor (22).
Int J Theor Phys (2010) 49: 884–913 899

and κ Tk α k β includes the energy-momentum tensor for matter field in usual gravity and
distortions of the Einstein tensor determined by distortions of linear connections.

Proof We can see that both the torsion (20) and distortion tensor, see formulas (22), became
zero if the conditions (30) are satisfied. In such a case, the distortion relations (21) trans-
kγ
form into  k k =   k k (even, in general, 
kγ

α β α β D = ∇).11 Even such additional constraints

are imposed, the geometric constructions are with nonholonomic variables because the an-
b
holonomy coefficients are not obligatory zero (for instance, wia = ∂a Nib etc, see formulas
(12)). 

3.2 The System of N-adapted Einstein Equations

The goal of this work is to prove that we can solve in a very general form any system
of gravitational filed equations in high dimensional gravity, for instance, for the canonical
d-connection and/or the Levi-Civita connection if such equations can be written as a variant
of (29),
k k
R αk β = ϒ αk β , (31)
k
with a general source parametrize in the form (4), ϒ αk β = diag[ϒk γ ], including possible
contributions from energy-momentum and/or distortion tensors.
Let us denote partial derivatives in the form
k
∂2 = ∂/∂x 2 , . . . , ∂v = ∂/∂v, ∂k v = ∂/∂ k v, . . . , ∂α = ∂/∂uα , . . . , ∂k α = ∂/∂u α .
kγ
Theorem 3.1 The gravitational field equations (31) constructed for k  D = {  k α k β } with
coefficients (17) and computed for a metric g = {gk β k γ } (14) with coefficients (15) are
k

equivalent to this system of partial differential equations:

22 = R33 = 1 ∂2 g2 · ∂2 g3 (∂2 g3 )2

R + − ∂22 g3
2g2 g3 2g2 2g3

∂3 g2 · ∂3 g3 (∂3 g2 )2 
+ + − ∂32 g2 = −ϒ4 (x i ), (32)
2g3 2g2
 √ 
 
R 55 = ∂v h5 ∂v ln |h4 h5 |  = −ϒ2 (x i , v),
44 = R (33)
2h4 h5  ∂ v h5 

4i = −wi β − αi = 0,
R (34)
2h4 2h4

5i = − h5 [∂v2 ni + γ ∂v ni ] = 0,
R (35)
2h4
 √ 
 
77 = ∂1 v h7 ∂1 v ln |h6 h7 |  = −1 ϒ2 (uα , 1 v),
66 = R
R
2h h  ∂1 h 
6 7 v 6
1
6μ = −wμ β − αμ = 0,
R 7μ = − h7 [∂12 nμ + 1 γ ∂1 v nμ ] = 0,
R
2h6 2h6 2h6 v

11 This is possible because the laws of transforms for d-connections, for the Levi-Civita connection and dif-
ferent types of tensors being adapted, or not, to a N-splitting (9) are very different.
900 Int J Theor Phys (2010) 49: 884–913
 √ 
 
99 = ∂2 v h9 ∂2 v ln |h8 h9 |  = −2 ϒ2 (u1 α , 2 v),
88 = R
R
2h8 h9  ∂ 2 v h8 
2 α1
8 1 μ = −w1 μ β − μ = 0,
R
2h8 2h8

9 1 μ = − h9 [∂22 n1 μ + 2 γ ∂2 v n1 μ ] = 0,
R
2h8 v
...
 √ 
∂ k v h5+2k  |h4+2k h5+2k | 
 
R4+2k = R5+2k =
4+2k 5+2k
∂ k ln  ∂k v h4+2k 
2h4+2k h5+2k v
k−1 α
= −k ϒ2 (u , k v),
k α k−1 μ
4+2k k−1 μ = −wk−1 μ β
R − = 0,
2h4+2k 2h4+2k

5+2k k−1 μ = − h5+2k [∂k2 nk−1 μ + k γ ∂k v nk−1 μ ] = 0,

R
2h4+2k v

where, for ∂v h4 = 0 and ∂v h5 = 0; ∂1 v h6 =  0 and ∂1 v h7 = 0; ∂2 v h8 = 0 and ∂2 v h9 =

0; . . . ; ∂k v h4+2k = 0 and ∂k v h5+2k = 0;12
 
 ∂ v h5 
φ = ln √ , αi = ∂v h5 · ∂i φ,
|h4 h5 |  (36)
β = ∂v h4 · ∂v φ, γ = ∂v (ln |h5 |3/2 /|h4 |);
 
 ∂ 1 v h7 
1 
φ = ln √ , αμ = ∂1 v h7 · ∂μ 1 φ,
|h6 h7 | 
 
1
β = ∂1 v h6 · ∂1 v 1 φ, 1
γ = ∂1 v ln |h7 |3/2 /|h6 | ;
 
 ∂ 2 v h9 
2 
φ = ln √ , α1 μ = ∂2 v h9 · ∂12μ φ,
|h8 h9 | 
 
2
β = ∂2 v h8 · ∂2 v 2 φ, 2
γ = ∂2 v ln |h9 |3/2 /|h8 | ;
...
 
 ∂ k h5+2k 
k
φ = ln √ v , αk−1 μ = ∂k v h5+2k · ∂k−1 μ k φ,
|h4+2k h5+2k | 
 
k
β = ∂k v h4+2k · ∂k v k φ, k
γ = ∂k v ln |h5+2k |3/2 /|h4+2k | .

Proof of this theorem is sketched in Appendix B.

Finally, we emphasize that the system of equations constructed in Theorem 3.1 can be
integrated in very general forms. For instance, for any given ϒ4 and ϒ4 , (32) relates an
un-known function g2 (x 2 , x 3 ) to a prescribed g3 (x 2 , x 3 ), or inversely. Equation (33) con-
tains only derivatives on y 4 = v and allows us to define h4 (x i , v) for a given h5 (x i , v), or
inversely, for h∗4,5 = 0. Having defined h4 and h5 , we can compute the coefficients (36),

12 Solutions, for instance, with ∂ h = 0 and/or ∂ h = 0, should be analyzed as some special cases (for
v 4 v 5
simplicity, we omit such considerations in this work).
Int J Theor Phys (2010) 49: 884–913 901

which allows us to find wi from algebraic equations (34) and to compute ni by integrating
two times on v as follow from (35). Similar properties hold true for equations on higher
order shells.

4 General Solutions for Einstein Equations with Extra Dimensions

In this section, we show how general solutions of the gravitational field equations can be
constructed in explicit form. There are three key steps: The first one is to generate exact
solutions with Killing symmetries for the canonical d-connection. At the second one, we
shall analyze the constraints selecting solutions for the Levi-Civita connections. The final
(third) step will be in generalizing the constructions by eliminating Killing symmetries.

4.1 Exact Solutions with Killing Symmetries

We formulate for the Einstein equations for the canonical d-connection:

Theorem 4.1 The general class of solutions of nonholonomic gravitational equations (31)
with Killing symmetries on e5+2k = ∂/∂y 5+2k is defined by ansatz of type (5) with k ω2 = 1
and coefficients gi , hk a , wk−1 α , nk−1 α computed for k = 0, 1, 2, . . . following formulas (6).

Proof We sketch the proof giving more details for the shell k = 0 (higher order constructions
being similar):
• The general solution of (32) can be written in the form  = g[0] exp[a2 x 2 (x 2 , x 3 ) +
a3 x (x , x )], were g[0] , a2 and a3 are some constants and the functions
3 2 3
x (x , x 3 ) de-
2,3 2

fine any coordinate transforms x 2,3 → x 2,3 for which the 2D line element becomes con-
formally flat, i.e.

g2 (x 2 , x 3 )(dx 2 )2 + g3 (x 2 , x 3 )(dx 3 )2 →  (x 2 , x 3 ) (d
x 2 )2 + (d
x 3 )2 ,

where  = ±1 for a corresponding signature. It is convenient to write some partial deriv-

atives, in brief, in the form ∂2 g = g • , ∂3 g = g  , ∂4 g = g ∗ . In coordinates
x 2,3 , (32) trans-
 
form into  ( ¨ + )− ˙ −  = 2 ϒ4 (x̃ , x̃ ) or
2 2 3

ψ̈ + ψ  = 2ϒ4 (x̃ 2 , x̃ 3 ), (37)

for ψ = ln | |. The integrals of (37) depends on the source ϒ4 . As a particular case we

can consider that ϒ4 = 0.
k−1
• For k ϒ2 (u α , k v) = 0, (33), and its higher shell analogs, relates two functions
k−1 α k k−1
h4+2k (u , v) and h5+2k (u α , k v) following two possibilities:
(a) to compute
 k−1 k−1
|h5+2k | = 1 h5+2k (u α ) + 2 h5+2k (u α )
 
k−1
× |h4+2k (uk−1 α , k v)|dv, for ∂ k v h4+2k (u α , k v) = 0;
k−1 α k−1 α k
= 1 h5+2k (u ) + 2 h5+2k (u ) v,
k−1 α
for ∂k v h4+2k (u , k v) = 0, (38)
902 Int J Theor Phys (2010) 49: 884–913

k−1 k−1
for some functions 1 h5+2k (u α ) and 2 h5+2k (u α ) stated by boundary conditions;
(b) or, inversely, to compute h4+2k for respectively given h5+2k , with ∂k v h5+2k = 0,
 
k−1
|h4+2k | = 0k h(u α ) ∂k v |h5+2k (uk−1 α , k v)|, (39)
k−1
with 0k h(u α ) given by boundary conditions. We note that the (33) with zero source
k−1 k−1
is satisfied by arbitrary pairs of coefficients h4+2k (u α , k v) and 0 h5+2k (u α ). So-
lutions with k ϒ2 = 0 can be found by ansatz of type
k−1 α
h5+2k [k ϒ2 ] = h5+2k , h4 [ k ϒ2 ] = ς4+2k (u , k v)h4+2k , (40)

where h4+2k and h5+2k are related by formula (38), or (39). Substituting (40), we
obtain

k−1 k−1 h4+2k h5+2k k
ς4+2k (u α , k v) = 0 ς4+2k (u α ) − k ϒ2 d v, (41)
4∂k v h5+2k
k−1 α
where 0 ς4+2k (u ) are arbitrary functions.
• The exact solutions of (34) for β = 0 are defined from an algebraic equation, wi β + αi =
0, where the coefficients β and αi are computed as in formulas (36) by using the solutions
for (32) and (33). The general solution is
 
wk = ∂k ln[ |h4 h5 |/|h∗5 |]/∂v ln[ |h4 h5 |/|h∗5 |], (42)

with ∂v = ∂/∂v and h∗5 = 0. If h∗5 = 0, or even h∗5 = 0 but β = 0, the coefficients wk could
be arbitrary functions on (x i , v). For the vacuum Einstein equations this is a degenerated
case imposing the compatibility conditions β = αi = 0, which are satisfied, for instance,
if the h4 and h5 are related as in the formula (39) but with h[0] (x i ) = const.
• Having defined h4 and h5 and computed γ from (36), we can solve (35) by integrating on
variable “v” the equation n∗∗ ∗
i + γ ni = 0. The exact solution is
 

nk = nk[1] (x i ) + nk[2] (x i ) h4 /( |h5 |)3 dv, h∗5 = 0;

= nk[1] (x i ) + nk[2] (x i ) h4 dv, h∗5 = 0;

 
= nk[1] (x i ) + nk[2] (x i ) 1/( |h5 |)3 dv, h∗4 = 0, (43)

for some functions nk[1,2] (x i ) stated by boundary conditions.

• The generating and integration formulas in higher order formulas (40), (41), (42), (43)
etc. are redefined in a form as it was considered for k = 0 in review articles [1, 4] which
result in formulas (6) for k ω2 = 1. 

We note that the solutions constructed in Theorem 4.1 are very general ones and con-
tain as particular cases all known exact solutions for (non) holonomic Einstein spaces with
Killing symmetries. They also can be generalized to include arbitrary finite sets of parame-
ters, see Ref. [1].

Corollary 4.1 An ansatz (5) with k ω2 = 1 and coefficients gi , hk a , wk−1 β , nk−1 β computed
following formulas (6) define solutions with Killing symmetries on e5+2k = ∂/∂y 5+2k of the
Int J Theor Phys (2010) 49: 884–913 903

kγ
Einstein equations (3) for the Levi-Civita connection  kα kβ if the coefficients of metric are
subjected additionally to the conditions (8).

Proof By straightforward computations for ansatz (14) with coefficients (15), we get that the
kγ kγ
conditions (30) resulting in  k α k β = 
 k α k β are just those written as (8). For such ansatz,
one holds the conditions (54) and the N-connection and torsion coefficients vanish, i. e. the
values (52) and (55) became zero. We get nonholonomic configurations for the Levi-Civita
connection with nontrivial anholonomy coefficients (12). 

We conclude that in order to generate exact solutions with Killing symmetries in Ein-
stein gravity and its higher order generalizations, we should consider N-adapted frames and
nonholonomic deformations of the Levi-Civita connection to an auxiliary metric compati-
ble d-connection (for instance, to the canonical d-connection, k D), when the corresponding
system of nonholonomic gravitational field equations (32)–(35) can be integrated in general
form. Subjecting the integral variety of such solutions to additional constraints of type (30),
i.e. imposing the conditions (8) to the coefficients of metrics, we may construct new classes
of exact solutions of Einstein equations for the Levi-Civita connection k ∇.

4.2 General Non-Killing Solutions

Our final aim is to consider general classes of solutions of the nonholonomic gravitational
field equations (31), and (for more particular cases), of Einstein equations (3) metrics de-
k k
pending on all coordinates u α = (x i , y a ), i.e. the solutions will be without Killing symme-
tries.
k
Let us introduce some nontrivial multiples k ω2 (u α ) before coefficients hk a parametrized
in the form (15) and defining solutions with Killing symmetries. We get an ansatz

  
k
ωg = 1 e1 ⊗ e1 + g
j (x )e ⊗ e + ω (x , y )ha (x , v)e ⊗ e
k j j 2 i a i a a

1 1 1b
+ 1 ω2 (u α )h1 a 1 b (uα , 1 v)e a ⊗ e
2 1 2 2
+ 2 ω2 (u α )h2 a 2 b (u α , 2 v)e a ⊗ e b + · · ·
k k−1 α k k
+ k ω2 (u α ) hk a k b (u , k v)e a ⊗ e b , (44)
k
where the N-adapted basis e a (and N-connection) are the same as in (14). Under such
k
noholonomic conformal transform13 (defined by generating functions 2 ω2 (u α )) of metric,
kγ kγ kγ
k
g → k ω g, the canonical d-connection deforms as 
 k k → k ω
 k k , where k ω 
 k k =
α β α β α β
ij k , ω L
(L abk , C bc
ji c , ω C a 1 a , ω C
; 1ωL
1
α 1 , 1 ω C
1 a 1 ; . . . ;
1
kω kk a k−1 , k−1 ω C
L k−1
k−1 α
k a ),
,k C
bα β c b c b α β kc ω kb kc

for C α1 = · · · = k−1 ω C

ji c = ω C k−1
k−1 α
= 0, kω k ak−1
L
β c β kc b α

=L k a
kk a k−1 + zk L k a = C
,k C k a ,
kk a k + zk C
b α ω k b k−1 α ω k b k c b c ω kb kc

13 We use the term “nonholonomic” because such transforms/deformations are adapted to a N-splitting stated
by a prescribed nonholonomic distribution on a corresponding high dimension spacetime.
904 Int J Theor Phys (2010) 49: 884–913

kk a k−1 = 1 ka kc
with z
kω L b α k ω2
h [hk b k c e k−1 β (k ω2 )
2
k k
− hk b k c Nk−1d
∂k (k ω2 )] = δ kab ek−1 β ln |k ω|;
β d
(45)
z k a
 ka ka ka ke 
k ω C k b k c = δ k b ∂ k c + δ k c ∂ k b − hk b k e h ∂ k e ln |k ω|, (46)
k
are computed by introducing coefficients of ωg (44) into (17).

Proposition 4.1 For nonholonomic N-adapted transforms k g (14) → k ω g (44) with shell co-
efficients k ω satisfying respectively the conditions ek−1 β (k ω) = 0, the Ricci tensor transform
Rk α k β (25) → k ω 
 Rk α k β , where

kω  ij , R
Rk α k β = {R ia , R ab ; ω R
ai , ω R αβ , R
α1 a , R 1 a 1 b ;
1 aβ , 1 ω R

1ω 1 α 1 β , R
R 1 α 2 a , R 2 a 2 b ; . . . ;
2 a 1 β , 2 ω R

k−1 ω k−1 α k−1 β , R

R k−1 α k a , R k a k b },
k a k−1 β , k ω R (47)

k−1 α k−1 β computed recurrently using R

with k−1 ω R k a k b = R
k−1 α k−1 β and k ω R k a k b + zk Rk a k b ,
ω
z 
where the deformation tensor k ω Rk a k b is given by formula

z
kω
k a k b = (2 − k m)D
R k b ln |k ω| − hk a k b hk c k d
k a D

×D k d ln |k ω| − (2 − k m)(D
k c D k a ln | k ω|)D
k b ln | k ω|

+ (2 − k m)hk a k b h
kc kd
k c ln |k ω|)D
(D  k d ln |k ω|. (48)

Proof It follows from an explicit computation of N-adapted coefficients of (47) taking into
account the deformation relations (45) and (46) when the condition ek−1 β (k ω) = 0, which is
k
just (7) from the Main Theorem 1.1. Working with shell coordinates y a , the formulas for
curvature and Ricci tensors are the same as on usual (pseudo) Riemannian spaces for the
k−1
Levi-Civita connection, when coordinates of type x i and y a can be considered as some
parameters. For “pure vertical” components, we can apply usual formulas for conformal
transforms, like (46) and (48) outlined, for instance, in Appendix D of monograph [40]. 

Remark 4.1 (1) There are two reasons to consider two dimensional shells with k m = 2: The
first one is that this results in field equations of type (33) which can be integrated in general
form (it is a problem, at least technically, to find exact solutions for k m > 2). The second
one is that from (48) we get zk ω R  ln |k ω|, with k 
k ak = δ k ak k    hk c k d D
k c D
k d being a
b b
shell type d’Alambert operator defined by the canonical d-connection.
(2) We can impose additionally the conditions
k
 ln |k ω| = 0, (49)

or to include such terms in sources (4), redefining the nonholonomic distributions to have
k
ϒ2 (u
k−1 α k
 ln |k ω(u α )|,
, k v) = kω ϒ2 (u α ) − k 
k
(50)
k
for some well defined kω ϒ2 (u α ) when formulas of type (41) can be computed. The condi-
tions (49) or (50) can be selected also by corresponding integration functions, for instance,
in (39), (42) and/or (43) and/or their higher shell analogs.
Int J Theor Phys (2010) 49: 884–913 905

(3) Solutions with k m > 2 can be with different topologies and generalized nonholonomic
conformal symmetries. Locally such constructions may be performed in a simplest way by
considering formulas only with k m = 2 by increasing the number of “formal” shells.

As a result, we get the proof of

Lemma 4.1 Any metric parametrized in the form (44) with coefficients depending on
all variables on a (pseudo) Riemannian manifold k V (dim k V = 3, or 2, +2k; with k =
0, 1, 2, . . . two dimensional shells) defines a “non-Killing” solution of the Einstein equations
for the canonical d-connection k  D if the coefficients are given by data (31) as solutions with
Killing symmetries of (32)–(35), when the parameters of nonholonomic conformal defor-
k
mations k ω(u α ) are chosen as generating functions satisfying the conditions ek−1 β (k ω) = 0
and, for instance, k   ln |k ω| = 0. Imposing additional restrictions on integration functions
as in Corollary 4.1, we get general solutions for the Levi-Civita connection k ∇.

Finally, in this section we formulate:

Conclusion 4.1 (1) Summarizing the Theorems 2.1–4.1 and Lemma 4.1, we prove the Main
Result stated in Theorem 1.1.
(2) The general solutions defined by the conditions of Theorem 1.1 (and related results)
can be extended to include contributions of an arbitrary number of commutative and non-
commutative parameters. This is possible following the constructions with Killing symme-
tries, in our case for metrics (14) provided in Ref. [1], which can be similarly reconsidered
with higher order shells.

5 Summary and Discussion

This work was primarily motivated by the question if the Einstein equations can be inte-
grated in very general forms, for generic off-diagonal metrics depending on all possible
variables, in arbitrary dimensions. To the best of our knowledge, such a problem has not yet
been addressed in mathematical and physical literature being known the high complexity
of related systems of nonlinear partial differential equations. This is in spite of the fact that
there were elaborated a number of analytic and numerical methods of constructing exact and
approximate solutions and that various types of such solutions seem to be of crucial physical
importance in modern astrophysics and cosmology. Here we note that the bulk of former de-
rived solutions are for diagonalizabe metrics (by coordinate transforms), depending on one
and/or two (in some exceptional cases, on three) variables, with compactified dimensions,
imposed symmetries, boundary conditions etc.
In a series of works, see reviews of results in Refs. [1–3], one of the main our goals
was to formulate a geometric method which would allow us to construct exact solutions of
gravitational field equations. We applied the formalism of nonholonomic distributions with
generating and integration functions, when some of them are subjected to additional condi-
tions/constraints (written as certain types of first order partial equations, algebraic relations,
symmetry conditions etc.). That allowed us to elaborate a general scheme for deriving exact
solutions with one Killing vector symmetry and various types of parametric dependencies.
Finally, the so-called anholonomic deformation method was developed for general “non-
Killing” solutions in paper [5].
906 Int J Theor Phys (2010) 49: 884–913

Following the anholonomic deformation method, we define some “more convenient”

holonomic and nonholonomic variables (frames coefficients and coordinates), which for cer-
tain types well defined conditions transform the Einstein equations into exactly integrable
systems of equations. The key idea is to use additionally some auxiliary linear connec-
tions correspondingly adapted to nonholonomic distributions. Surprisingly, in our approach,
it was possible to reformulate (and, in general, to modify) the Einstein equations in such
forms, when general integral varieties can be constructed. Subjecting the coefficients of such
way defined solutions to additional constraints, we can determine some integral subvarieties
for standard gravity theories and generalizations. Here we note that our auxiliary connec-
tion (the so-called, canonical distinguished connection, in brief, d-connection) is also metric
compatible and uniquely defined by the metric coefficients. It contains a nonholonomically
induced torsion but such a geometric object is completely different from that, for instance,
in Einstein–Cartan/gauge/string theory. In our approach, we do not need any additional field
equations because we work with torsion coefficients induced by certain off-diagonal co-
efficients of metric. All geometric constructions can be equivalently performed using the
Levi-Civita connection or, alternatively, the canonical d-connection.
Of course, our findings should be considered only in a line of qualitative understand-
ing of the concept of general exact solutions in Einstein and high dimensional gravity. For
such generic nonlinear systems, it is not possible to formulate any general uniqueness and
completeness criteria for solution if we do not introduce any additional suppositions on
classes of generating functions, symmetries, horizons, singularities, asymptotic conditions
etc. Only in some more special/restricted cases, we can provide certain physical meaning for
such general classes of solutions; to put, for instance, the Cauchy problem, construct some
evolution models, determine symmetries of interactions etc. Our constructions are general
ones because “almost” any solution in gravity theories can be parametrization in such a
form at least locally even very different classes of metrics and connections can be stated
globally for different topologies, boundary conditions, with various types of horizons and
singularities etc. It is not our aim to perform such studies in this article.
Finally, we emphasize that the bulk of exact solutions in gravity theories (in Einstein
gravity and various supersymmetric/noncommutative sting, brane, gauge, Kaluza–Klein,
Lagrange–Finsler, generalizations etc.) can be represented in a form similar to (5). In this
paper, we do not analyze possible explicit symmetries and physical properties of such so-
lutions. We consider such problems in our recent papers [20–24, 38, 41–43] and plan to
provide further developments and applications in our future works.

Acknowledgement Author is grateful to M. Anastasiei for discussions and support.

Appendix A: Coefficients of N-adapted Curvature

In this section, we outline some formulas which play an important role in finding systems
of partial differential equations which are equivalent to the Einstein equations with non-
holonomic variables of arbitrary dimensions, see details in Refs. [1–4]. The formulas for
coefficients of curvature R of the canonical d-connection  D are written with respect to
N-adapted frames (10) and (11).
Int J Theor Phys (2010) 49: 884–913 907

Theorem A.1 The curvature R  (23) of the canonical d-connection  D computed with respect
to N-adapted frames (10) and (11) is characterized by coefficients
hj
R i i i m i m i i
k = ek L hj − ej L hk + L hj L mk − L hk L mj − C ha  kj ,
a

R abj − ej L
abj k = ek L abk + L ack − L
cbj L acj − C
cbk L abc ckj ,

R i j k − D
ij ka = ea L ij a + C
k C ij b Tbka ,
(51)
cbka = ea L
R cbk − Dk C cbd Tcka ,
cba + C
ij bc = ec C
R ij b − eb C hj b C
ij c + C hj c C
ihc − C ihb ,

R abc − ec C
abcd = ed C ebc C
abd + C ebd C
aed − C aec ;

ατβγ = eγ L
R ατβ − eβ L
ατ γ + L αμγ − L
μτβ L αμβ − C
μτ γ L α 1a
a
 γβ
1
,
τ

11a = eγ L
R 11a − eβ L
11a + L
11c L 11a − L
11c L 11a − C
11a 1 1γβ
c
,
bβγ bβ bγ bβ cγ bγ cβ b c

α
R 1a
αβγ − D
= e1 a L α 1 + C
γ C α 1 T1 b1 ,
βγ β a β b γ a

1 c1
R 1a
1 c − D
= e1 a L
1
 c1 1 + C
γ C 1 c 1 T c1 ,
1 1 1
bγ bγ b a b d γ a

α 1 1 = e1 c C
R α 1 − e1 b C
α 1 + C α − C
μ 1 C α ,
μ 1 C
β b c β b β c β b μ 1c β c μ 1b

R 1 a1 1 − e1 c C
1 a1 1 1 = e1 d C 1 a1 1
b c d b c b d

1 e1 1 C
+C 1 a1 1 − C
1 e1 1 C1 a1 1 ;
b c e d b d e c

1 α1  α1  α1  μ1 1 α1 1
1 1 1
R τ 1β 1γ = e1 γ L τ 1β − e1 β L τ 1γ +L τ 1β L μ γ

 μ1 1 α1 1 − C
1 α1
1 2a
−L τ 1γ L μ β τ 2a  1γ 1β ,

R 22a 1 − e1 β L
22a 1 1 = e1 γ L 22a 1
b β γ b β b γ

 2c 1 L
+L
2
 2a 1 − L
 2c 1 L
2
 2a 1 − C
 2a 2 
2 2 2 2c
1γ 1β ,
b β c γ b γ c β b c

1 α1 1
R 2a = e2 a L
1
1 γ C
 α1 1 − D  α1 2 T 1b
 α1 2 + C 1 1 2
2a ,
β γ β γ β a β b γ

2 c2 1
R 2a
2 c2 1 − D
= e2 a L 2 c2 2 + C
1 γ C 22 c 2 T21c 2a ,
b γ b γ b a b d γ

1 α1 2 2 = e2 c C
R α 1 2 − e2 b C
1 α1 2
β b c β b β c

 μ1 2 C α1 2 − C
 μ1 2 C α1 2 ,
1 1 1 1
+C β b μ c β c μ b

R 2 a2 2 − e 2 c C
2 a2 2 2 = e2 d C 2 a2 2
b c d b c b d

2 e2 2 C
+C 2 a2 2 − C
2 e2 2 C2 a2 2 ;
b c e d b d e c

...

R
k−1 α
k−1 τ k−1 β k−1 γ

= e k−1 γ L
k−1 α
k−1 τ k−1 β

− ek−1 β L
k−1 α
k−1 τ k−1 γ

 k−1  k−1
k−1 μ k−1 μ
+L k−1 τ k−1 β L k−1 μ k−1 γ − L
α
k−1 τ k−1 γ L α
k−1 μ k−1 β
908 Int J Theor Phys (2010) 49: 884–913


−C
k−1 α
k−1 τ k a 
ka
k−1 γ k−1 β ,

kka k−1 k−1 = ek−1 γ L

R k−k k a
k−1 b k−1 β − ek−1 β L k b k−1 γ
a
b β γ

 kc k−1 L
+L
k
 kc k−1 L
 ka k−1 − L
k
 ka k−1 k k
b β c γ b γ c β

 ka k 
−C
k kc
k−1 γ k−1 β ,
b c


R
k−1 α
k−1 β k−1 γ k a

= ek a L
k−1 α
k−1 β k−1 γ −D 
 k−1 γ C k−1 α
k−1 β k a


+C
k−1 α
k−1 β k b T k−1
kb
ka ,
γ

k ck k−1
R ka
k ck k−1 − D
= ek a L k ck k + C
k−1 γ C kk c k Tk k−1
c
ka ,
b γ b γ b a b d γ

k−1
R α k−1 α
k−1 β k b k c = ek c C k−1 β k b − ek b C k−1 β k c
α

 k−1  k−1
k−1 μ k−1 μ
+C k−1 β k b C k−1 μ k c − C
α
k−1 β k c C α
k−1 μ k b ,

R k ak k − ek c C
k ak k k = ek d C k ak k
b c d b c b d

k ek k C
+C k ek k C
k ak k − C k ak k .
b c e d b d e c

Proof It follows from “shell by shell computations” as in Refs. [1–4, 6–9, 30–32, 34–36]. 

Appendix B: Proof of Theorem 3.1

Such a proof can be obtained by straightforward computations as in Parts I and II of mono-

graph [3], containing all developments from Refs. [30–32], see also summaries and some
important details and discussions in Refs. [1, 2]. In this section, we generalize some for-
mulas by considering “shell” labels for indices, when k = 0, 1, 2, . . . using data (15) for a
metric k g = {gk β k γ } (14).
We can perform a N-adapted differential calculus on a N-anholonomic manifold if in-
k
stead of partial derivatives ∂k α = ∂/∂u α there are considered operators (10) parametrized
ka
in the form ek−1 α = ∂k−1 α − Nk−1 ∂k = ∂k−1 α − wk−1 α ∂k v − nk−1 α ∂k y , for y 4+2k = k v and
α a
y 5+2k
= y. For instance, for data (15), the coefficients of N-connection curvature (13) are
k

k−1 α k−1 β = ∂k−1 α wk−1 β − ∂k−1 β wk−1 α

4+2k
− wk−1 α ∂k v wk−1 β + wk−1 β ∂k v wk−1 α ; (52)
5+2k
k−1 α k−1 β = ∂k−1 α nk−1 β − ∂k−1 β nk−1 α
− wk−1 α ∂k v nk−1 β + wk−1 β ∂k v nk−1 α .

In a similar form we compute all coefficients of the canonical d-connection (17) and its
Ricci and Einstein tensors.
Int J Theor Phys (2010) 49: 884–913 909

B.1 Coefficients of the Canonical d-connection

kγ
For data (15), we get such nontrivial coefficients of 
 kα kβ :

222 = ∂2 g2 ,
L 223 = ∂3 g2 ,
L 233 = − ∂2 g3 ,
L 322 = − ∂3 g2 ,
L
2g2 2g2 2g2 2g3

323 = ∂2 g3 ,
L 333 = ∂3 g3 ,
L 44i = 1 (∂i h4 − wi ∂v h4 );
L 54j = 1 ∂v nj ,
L
2g3 2g3 2h4 2

55j = 1 (∂j h5 − wj ∂v h5 );
L 44
C 4
=
∂ v h4
, 55
C 4
=−
∂ v h5
, 45
C 5
=
∂ v h5
;
2h5 2h4 2h4 2h5
...
1 (53)
4+2k k−1 =
L (∂k−1 α h4+2k − wk−1 α ∂k v h4+2k ),
4+2k α 2h4+2k

5+2k k−1 = 1 ∂k v nk−1 α ,

L 5+2k k−1 =
L
1
(∂k−1 α h5+2k − wk−1 α ∂k v h5+2k );
4+2k α 2 5+2k α 2h5+2k

4+2k ∂k v h4+2k 5+2k ∂k v h5+2k

4+2k = 5+2k = −
4+2k 4+2k
C , C ,
2h4+2k 2h4+2k

4+2k ∂ k v h5+2k
5+2k =
5+2k
C .
2h5+2k
We note that

ji c = 1 g ik ∂gj k = 0, k−1

k−1 α 1 k−1 k−1 ∂gk−1 β k−1 τ
C ..., C β kc
= g α τ = 0, (54)
2 ∂y c 2 ∂y k c
which is an important condition for generating exact solutions of the Einstein equations for
the Levi-Civita connection, see formulas (30).

B.2 Calculation of Torsion Coefficients

The nontrivial coefficients of torsions (20) for data (15) are given by formulas (52) and,
respectively, (53) resulting in

Tk−1
4+2k
α k−1 β
= ∂k−1 β wk−1 α − ∂k−1 α wk−1 β

− wk−1 β ∂k v wk−1 α + wk−1 α ∂k v wk−1 β ;

Tk−1
5+2k
α k−1 β
= ∂k−1 β nk−1 α − ∂k−1 α nk−1 β

wk−1 β ∂k v nk−1 α − wk−1 α ∂k v nk−1 β ,

1 (55)
T4+2k
4+2k
k−1 α = ∂k v wk−1 α − (∂k−1 α h4+2k − wk−1 α ∂k v h4+2k ),
2h4+2k
h5+2k 1
T5+2k
4+2k
k−1 α = ∂k nk−1 α , T4+2k
5+2k
k−1 α = ∂k nk−1 α ,
2h4+2k v 2 v
1
T5+2k
5+2k
k−1 α = − (∂k−1 α h5+2k − wk−1 α ∂k v h5+2k ).
2h5+2k
910 Int J Theor Phys (2010) 49: 884–913

B.3 Calculation of the Ricci Tensor

ij = R
For instance, let us compute the values R kij k from (26),

R i.hj − ej L
ihj k = ek L m
i.hk + L i m i i
.hj Lmk − L.hk Lmj − C.ha .j k ,
a

using (51) and C .ha

i i.hj = ∂k L
= 0 (54). We have ek L i.hj + Nka ∂a L
i.hj = ∂k L i.hj )∗ =
i.hj + wk (L
i.hj because L
∂k L i.hj do not depend on variable y 4 = v. We use, in brief, denotations of type

∂2 g = g , ∂3 g = g , ∂4 g = g ∗ .
•

Deriving (53), we obtain

 
222 = g2•• (g • )2 223 = g2• g• g
∂2 L − 2 2, ∂2 L − 2 22 ,
2g2 2(g2 ) 2g2 2(g2 )
 
233 = − g3•• g• g• 322 = − g2• g• g
∂2 L + 2 32 , ∂2 L + 2 32 ,
2g2 2(g2 ) 2g3 2(g3 )
•• • 2 •  • 
323 = g3 − (g3 ) ,
∂2 L 333 = g3 − g3 g3 ,
∂2 L
2g3 2(g3 )2 2g3 2(g3 )2
•  •  ll l 2
222 = g2 − g2 g2 ,
∂3 L 223 = g2 − (g2 ) ,
∂3 L
2g2 2(g2 )2 2g2 2(g2 )2
   
233 = − g3• g• g 322 = − g2 g• g
∂3 L + 3 22 , ∂3 L + 2 22 ,
2g2 2(g2 ) 2g3 2(g3 )
 
323 = g3• g• g 333 = g3ll (g l )2
∂3 L − 3 32 , ∂3 L − 3 2.
2g3 2(g3 ) 2g3 2(g3 )

For these values, there are only 2 nontrivial components,

•• • • • 2 ll l l l 2
2323 = g3 − g2 g3 − (g3 ) + g2 − g2 g3 − (g2 ) ,
R
2g2 4(g2 )2 4g2 g3 2g2 4g2 g3 4(g2 )2
•• • • • 2 ll l l l 2
3223 = − g3 + g2 g3 + (g3 ) − g2 + g2 g3 + (g2 )
R
2g3 4g2 g3 4(g3 ) 2 2g3 4(g3 )2 4g2 g3

22 = −R
with R 3223 and R
33 = R
2323 , or


33 = −
22 = R 1 g • g • (g • )2 gl gl (g l )2
R g3•• − 2 3 − 3 + g2 − 2 3 − 2
2g2 g3 2g2 2g3 2g3 2g2

as in (32).
Now, we consider

c  c 
cbka = ∂ L.bk − ∂ C.ba + L
R c.dk C
.ba
d
− d.bk C
L .da
c
− d.ak C
L .bd
c
+C c d
.bd T.ka
∂y a ∂x k
c.bk
∂L
= .ba|k
−C c
+C c d
.bd T.ka
∂y a
Int J Theor Phys (2010) 49: 884–913 911

a
from (51). Contracting indices, we get Rbk = R
abka = ∂L.bk .ba|k
−C a
+C.bd
a d
T.ka . Let us denote
∂y a
     d   e   d 
Cb = C.ba and write C.b|k = ek Cb − L bk Cd = ∂k Cb − Nk ∂e Cb − L bk Cd = ∂k C
c b − wk Cb∗ −
 d  
L bk Cd . We express Rbk = [1] Rbk + [2] Rbk + [3] Rbk , where

[1] Rbk
4bk )∗ ,
= (L [2] Rbk
b + wk C
= −∂k C b∗ + L d ,
dbk C

[3] Rbk =C a d
.bd  4 T.k4
T.ka = C.b4 4
+C 4 5
.b5 T.k4 + C 5 4
.b4 .b5
T.k5 + C 5 5
T.k5
∗ ∗
4 = C
for C 44
4 45
+C 5 h h
5 = C
= 2h44 + 2h55 , C 54
4
+C55
5
= 0.
4k = [1] R4k + [2] R4k + [3] R4k with
We compute R

[1] R4k
44k )∗ ,
= (L [2] R4k
4 + wk C
= −∂k C 4∗ + L 4 ,
44k C

[3] R4k =C 4 4
.44 T.k4 + C 4 5
.45 T.k4 + C 5 4
.44 .45
T.k5 + C 5 5
T.k5 .

Summarizing, we get

  
4k = wk h∗∗ (h∗5 )2 h∗4 h∗5 h∗5 ∂k h4 ∂k h5
2h5 R 5 − − + + − ∂k h∗5
2h5 2h4 2 h4 h5
which is equivalent to (34).
5k =
In a similar way, we compute R [1] R5k + [2] R5k + [3] R5k , where

[1] R5k
45k )∗ ,
= (L [2] R5k
5 + wk C
= −∂k C 5∗ + L 4 ,
45k C

[3] R5k =C 4 4
.54 T.k4 + C 4 5
.55 T.k4 + C 5 4
.54 .55
T.k5 + C 5 5
T.k5 .
∗ h∗5
We have R 45k )∗ + L
5k = (L 4 + C
45k C 4 5
.55 T.k4 + C 5 4
.54 T.k5 = (− hh54 n∗k )∗ − h5 ∗ h4
n ( + ) +
h4 k 2h4 2h5
h∗5 h5 ∗ h∗5 1 ∗
n
2h5 2h4 k
− n ,
2h4 2 k
which can be written
 
5k = h5 n∗∗ h5 ∗ 3 ∗ ∗
2h4 R k + h4 − h5 n k ,
h4 2
i.e. we prove (35).
For
i.j k  i 
∂L ∂ C.j a
ij ka =
R − i l l i c i .ji b T.ka
+ L.lk C.j a − L.j k C.la − L.ak C.j c + C b
∂y k ∂x k

from (51), we obtain zeros because C .ji b = 0 and L i.j k do not depend on y k . So, R
j a =

R j ia = 0.
i
a a
Taking Rabcd = ∂ C.bc
d −
∂C
e a e a
c + C.bc C.ed − C.bd C.ec from (51) and contracting the indices in
.bd
∂y ∂y
d d
bc =
order to obtain the Ricci coefficients,R
∂C .bc
−
∂C .bd
+C e d
.bc C.ed − C e d
.bd C.ec , we compute
∂y d ∂y c

bc = (C
R .bc
4 ∗ b + C
) − ∂c C 4 
.bc C4 − C 4 4
.b4 C.4c − C 4 5
.b5 .b4
C.4c − C 5 4
C.5c − C 5 5
.b5 C.5c .

There are nontrivial values, R 44 = (C.44

4 ∗
) −C 4∗ + C444  44
(C4 − C 4 .45
) − (C 5 2 55 =
) and R
(C 4 ∗ .55 (−C
.55 ) − C 4 4 + 2C
.45 ) resulting in
5


44 = R 55 = 1 (h∗5 )2 h∗4 h∗5
R −h∗∗
5 + +
2h4 h5 2h5 2h4
912 Int J Theor Phys (2010) 49: 884–913

which is just (33).

Computations for higher shells, with k = 1, 2, . . . are similar.
Theorem 3.1 is proven.

References

1. Vacaru, S.: Parametric nonholonomic frame transforms and exact solutions in gravity. Int. J. Geom.
Methods Mod. Phys. (IJGMMP) 4, 1285–1334 (2007)
2. Vacaru, S.: Exact solutions with noncommutative symmetries in Einstein and gauge gravity. J. Math.
Phys. 46, 042503 (2005)
3. Vacaru, S., Stavrinos, P., Gaburov, E., Gonţa, D.: Clifford and Riemann-Finsler Structures in Geometric
Mechanics and Gravity. Selected Works, Differential Geometry—Dynamical Systems, vol. 7. Geometry
Balkan Press, Bucharest (2006). www.mathem.pub.ro/dgds/mono/va-t.pdf and arXiv:gr-qc/0508023
4. Vacaru, S.: Finsler and Lagrange geometries in Einstein and string gravity. Int. J. Geom. Methods Mod.
Phys. (IJGMMP) 5, 473–511 (2008)
5. Vacaru, S.: On general solutions in Einstein gravity. arXiv:0909.3949 [gr-qc], version 1
6. Vacaru, S.: Superstrings in higher order extensions of Finsler superspaces. Nucl. Phys. B 434, 590–656
(1997)
7. Vacaru, S.: Spinors and field interactions in higher order anisotropic spaces. J. High Energy Phys. 09,
011 (1998)
8. Vacaru, S., Vicol, N.: Nonlinear connections and spinor geometry. Int. J. Math. Math. Sci. (IJMMS) 23,
1189–1237 (2004)
9. Vacaru, S., Stavrinos, P.: Spinors and Space-Time Anisotropy. Athens University Press, Athens (2002)
10. Miron, R.: The Geometry of Higher-Order Lagrange Spaces, Application to Mechanics and Physics.
FTPH, vol. 82. Kluwer Academic, Boston (1997)
11. Miron, R., Hrimiuc, D., Shimada, H., Sabau, V.S.: The Geometry of Hamilton and Lagrange Spaces.
Kluwer Academic, Dordrecht (2000)
12. Miron, R.: The Geometry of Higher-Order Hamilton Spaces. Kluwer Academic, Dordrecht (2003)
13. Anastasiei, M., Vacaru, S.: Fedosov quantization of Lagrange-Finsler and Hamilton-Cartan spaces and
Einstein gravity lifts on (co) tangent bundles. J. Math. Phys. 50, 013510 (2009)
14. Miron, R., Anastasiei, M.: Vector Bundles and Lagrange Spaces with Applications to Relativity. Geom-
etry Balkan Press, Bukharest (1997). Translation from Romanian of Editura Academiei Romane (1987)
15. Miron, R., Anastasiei, M.: The Geometry of Lagrange Spaces: Theory and Applications. FTPH, vol. 59.
Kluwer Academic, Dordrecht (1994)
16. Vacaru, S.: Curve flows and solitonic hierarchies generated by Einstein metrics. Acta Appl. Math. (2009).
arXiv:0810.0707 [math-ph]
17. Anco, S., Vacaru, S.: Curve flows in Lagrange-Finsler geometry, bi-Hamiltonian structures and solitons.
J. Geom. Phys. 59, 79–103 (2009)
18. Vacaru, S., Singleton, D.: Ellipsoidal, cylindrical, bipolar and toroidal wormholes in 5D gravity. J. Math.
Phys. 43, 2486–2504 (2002)
19. Vacaru, S., Singleton, D.: Warped, anisotropic wormhole/soliton configurations in vacuum 5D gravity.
Class. Quantum Gravity 19, 2793–2811 (2002)
20. Vacaru, S.: Ricci flows and solitonic pp-waves. Int. J. Mod. Phys. A 21, 4899–4912 (2006)
21. Vacaru, S., Visinescu, M.: Nonholonomic Ricci flows and running cosmological constant: I. 4D Taub-
NUT metrics. Int. J. Mod. Phys. A 22, 1135–1159 (2007)
22. Vacaru, S.: Nonholonomic Ricci flows: Exact solutions and gravity. Electron. J. Theor. Phys. 6(20), 27–
58 (2009)
23. Vacaru, S.: Nonholonomic Ricci flows and parametric deformations of the solitonic pp-waves and
Schwarzschild solutions. Electron. J. Theor. Phys. 6(21), 63–93 (2009)
24. Vacaru, S.: Nonholonomic Ricci flows, exact solutions in gravity, and symmetric and nonsymmetric
metrics. Int. J. Theor. Phys. 48, 579–606 (2009)
25. Vacaru, S.: Exact solutions in locally anisotropic gravity and strings. In: Rembielinski, J. (ed.) Particile,
Fields and Gravitations. AIP Conference Proceedings, vol. 453, pp. 528–537. American Institute of
Physics, Woodbury (1998). arXiv:gr-qc/9806080
26. Vacaru, S.: Locally anisotropic kinetic processes and thermodynamics in curved spaces. Ann. Phys.
(N.Y.) 290, 83–123 (2001)
27. Vacaru, S.: Anholonomic soliton–dilaton and black hole solutions in general relativity. J. High Energy
Phys. 04, 009 (2001)
Int J Theor Phys (2010) 49: 884–913 913

28. Kramer, D., Stephani, H., Herdlt, E., MacCallum, M.A.H.: Exact Solutions of Einstein’s Field Equations.
Cambridge University Press, Cambridge (1980). 2nd edition (2003)
29. Bicak, J.: Selected solutions of Einstein’s field equations: Their role in general relativity and astrophysics.
Lect. Notes Phys. 540, 1–126 (2000)
30. Vacaru, S.: Nonholonomic deformations of disk solutions and algebroid symmetries in Einstein and extra
dimension gravity. gr-qc/0504095
31. Vacaru, S.: A new method of constructing black hole solutions in Einstein and 5D dimen-
sion gravity. hep-th/0110250. Chapter 10 in Clifford and Riemann-Finsler Structures in Geo-
metric Mechanics and Gravity, Selected Works, by S. Vacaru, P. Stavrinos, E. Gaburov, D.
Gonta. Differential Geometry—Dynamical Systems, Monograph 7 (Geometry Balkan Press, 2006).
www.mathem.pub.ro/dgds/mono/va-t.pdf and gr-qc/0508023
32. Vacaru, S., Gaburov, E., Gontsa, D.: A method of constructing off-diagonal solutions in metric-
affine and string gravity. hep-th/0310133. Chapter 2 in Clifford and Riemann-Finsler Structures in
Geometric Mechanics and Gravity, Selected Works, by S. Vacaru, P. Stavrinos, E. Gaburov and D.
Gonta. Differential Geometry—Dynamical Systems, Monograph 7 (Geometry Balkan Press, 2006).
www.mathem.pub.ro/dgds/mono/va-t.pdf and gr-qc/0508023
33. Vacaru, S.: Gauge gravity and conservation laws in higher order anisotropic spaces. hep-th/9810229
34. Vacaru, S.I., Chiosa, I.A., Vicol, N.A.: Locally anisotropic supergravity and gauge gravity on noncom-
mutative spaces. In: Duplij, S., Wess, J. (eds.) NATO Advanced Research Workshop Proceedings Non-
commutative Structures in Mathematics and Physics, Kyiv, Ukraine, September 23–27, pp. 229–243.
Kluwer Academic, Dordrecht (2001). hep-th/0011221
35. Dehnen, H., Vacaru, S.: Nonlinear connections and nearly autoparallel maps in general relativity. Gen.
Relativ. Gravit. 35, 807–850 (2003)
36. Vacaru, S., Dehnen, H.: Locally anisotropic structures and nonlinear connections in Einstein and gauge
gravity. Gen. Relativ. Gravit. 35, 209–250 (2003)
37. Vacaru, S.: Branes and quantization of an A-model complexification for Einstein gravity in almost Kähler
variables. Int. J. Geom. Methods Mod. Phys. (IJGMMP) 6, 873–909 (2009)
38. Vacaru, S.: Einstein gravity in almost Kähler variables and stability of gravity with nonholonomic distri-
butions and nonsymmetric metrics. Int. J. Theor. Phys. 48, 1973–2000 (2009)
39. Vacaru, S.: Deformation quantization of nonholonomic almost Kähler models and Einstein gravity. Phys.
Lett. A 372, 2949–2955 (2008)
40. Wald, R.M.: General Relativity. University of Chicago Press, Chicago (1984)
41. Vacaru, S.: Black holes, ellipsoids, and nonlinear waves in pseudo-Finsler spaces and Einstein gravity.
arXiv:0905.4401 [gr-qc]
42. Vacaru, S.: Finsler black holes induced by noncommutative anholonomic distributions in Einstein grav-
ity. arXiv:0907.4278 [math-ph]
43. Anastasiei, M., Vacaru, S.: Nonholonomic black ring and solitonic solutions in Finsler and extra dimen-
sion gravity theories. arXiv:0906.3811 [hep-th]