1. Introduction
Differentiable manifolds
with a non-symmetric metric tensor,
, and non-symmetric affine connection and their mappings are of interest to many scientists [
1,
2,
3,
4,
5,
6,
7,
8,
9,
10]. Although the notion of non-symmetric affine connection was used in several works before A. Einstein, for example in [
11], the use of non-symmetric connection became especially evident after the appearance of the works of A. Einstein, relating to create his Unified Field Theory [
12,
13], where the symmetric part of a basic tensor is related to gravitation and the anti-symmetric one to electromagnetism. Since 1951, Eisenhart was occupied with problems of spaces with non-symmetric basic tensors and non-symmetric connection in several works [
14]. He defined a generalized Riemannian space as “a space of coordinates with which ia associated a non-symmetric tensor
”, and the connection coefficients are defined by
In 1952, Eisenhart obtained two curvature tensors in generalized Riemannian space, using the fact that the connection coefficients are non-symmetric.
Geometric mappings are interesting and significant, both theoretically and practically. Geodesic and almost geodesic lines are important for geometry and physics. Geodesic mappings and their generalizations, like
F-planar and holomorphically projective mappings, have been considered in many papers [
1,
4,
6,
9,
10,
15,
16,
17]. As a kind of generalization of holomorphically projective mappings, biholomorphically projective mappings were defined and considered in the paper [
18].
In the sense of Eisenhart’s definition, an
N-dimensional manifold
equipped with a non-symmetric metric tensor
, whose components are
such that
for at least one pair of indices
, is a generalized (pseudo-)Riemannian space
. The symmetric and anti-symmetric parts of the metric
are
We also assume that
, which we use to define a contravariant symmetric metric tensor by components such as
. The affine connection coefficients of the space
are generalized Christoffell symbols
given by [
18],
where
where, for example,
. Generally,
, and the symmetric and anti-symmetric parts of
, respectively, are given by the formulas
The magnitude
is the torsion tensor of the space
.
In the case of generalized Riemannian space
, the following curvature tensors are linearly independent [
19]:
where “|” denotes a covariant derivative with respect to the symmetric affine connection
, whose curvature tensor is
.
The family of curvature tensors of the space
is
for scalars
u,
,
v,
, and
w.
The linearly independent curvature tensors (
2)–(7) are extracted from the family (
8) by selecting the following constants:
:
Remark 1. Let us obtain invariants for a geodesic mapping in the manner proposed by H. Weyl [20]. The basic equation of this mapping is From this basic equation, one finds the transformation rule of curvature tensor to , If the exchange is involved in Equation (11), we will transform it to After contracting the relation (12) by i and j, and by i and m, and with respect to the symmetry in Riemannian space, such as , we obtain After substituting the expression (13) into (12), we obtain From Equation (14), we obtain that the Weyl projective tensoris an invariant for the geodesic mapping f. The traces of are , , and . For this reason, there are two problems:
- 1.
We do not have an invariant for the geodesic mapping that does not neglect the Ricci tensor after contraction.
- 2.
If there are not Kronecker delta-symbols in the transformation rule of the curvature tensor under an analyzed mapping, we are not able to apply Weyl’s method to obtain an invariant for the studied mapping.
These problems were solved in the paper [21] by obtaining general formulas for basic and derived invariants of a mapping between affine connection spaces. We will apply this methodology in this research and obtain invariants for a special mapping of generalized Riemannian space. Because the corresponding deformation tensor will not be a function of a Kronecker delta-symbol, we will be able to obtain just one invariant for the studied mapping, which is a function of the curvature tensor . Unlike the Weyl projective tensor, the traces of the invariant that will be obtained in this paper will not vanish. 2. Review of Basic Invariants
One of the most important current problems in differential geometry is finding new invariant geometric objects. T. Y. Thomas [
22] and H. Weyl [
20] started the research about invariants of special diffeomorphisms between spaces with symmetric affine connection. Many authors have continued Thomas’s and Weyl’s works, for example V. Berezovski, J. Mikeš, E. Stepanova, etc. [
1,
2,
5,
15]. Also, some invariant geometrical objects for diffeomorphisms of spaces with non-symmetric affine connection are obtained in many papers, for example [
9,
16,
17,
23].
In this section, we will refer to the results presented in [
21]. Namely, if
is a mapping of Riemannian space
whose deformation tensor is
for geometric objects
and
symmetric by indices
j and
k, it is obtained that
and
for
and the corresponding
and
. The geometrical objects
and
are named the associated basic invariants of Thomas and Weyl type for the mapping
f.
For an equitorsion mapping
, the families curvature tensors (
8) and
for the covariant derivatives with respect to torsion-free affine connections
and
, respectively, denoted by “|” and “‖”, satisfy the relation
Because the mapping
f is equitorsion, the equality
holds. With this in mind, and the invariance of geometrical objects
and
, the equality
is satisfied for
After substituting the expressions of
and
given by (
16) and (17) and the corresponding expressions of
and
into the equalities (
19) and (20), respectively, one obtains
The difference of geometrical objects
and
given by Formulas (
21) and (22) is
where
is an anti-symmetrization without division by indices
m and
n.
The first seven summands in Equation (23) coincide with the right side of equality (
18). Hence, the forthcoming equality holds,
for the families
and
of the curvature tensors of the spaces
and
.
From the equality (24), we obtain the expressions of the geometrical objects
and
as follows:
The next theorem holds.
Theorem 1. Let be an equitorsion mapping of a generalized Riemannian space . The geometrical objects given by Equation (22) are invariants for the mapping f expressed in terms of the curvature tensor of the associated space .
The geometrical objects given by Equation (26) are invariants for the mapping f expressed in terms of the curvature tensors of the space .
The geometrical objects are equitorsion invariants for the mapping f.
Because there are six linearly independent curvature tensors [
19], the next corollaries are satisfied.
Corollary 1. Let be an equitorsion mapping of Riemannian space . Three of the invariants for this mapping, given by Equation (22), are linearly independent.
Corollary 2. Let be an equitorsion mapping of Riemannian space . Six of the invariants for this mapping, given by Equation (26), are linearly independent.
Corollary 3. Let be an equitorsion mapping of Riemannian space . The geometrical objects , whose components areare invariants for the mapping f. Remark 2. When we express the invariant in terms of curvature tensors of the space , we involve summands like . These summands are invariants for the equitorsion mapping f, and they make six linearly independent invariants. These summands are extracted from the invariant expressed in terms of the curvature tensor of the associated space and other non-invariant objects. For this reason, we will obtain invariants without including other invariants as separate summands.
3. Invariants for Equitorsion Canonical Biholomorphically
Projective Mappings
Let
and
be two generalized Riemannian spaces. We will observe these spaces in the common system of coordinates defined by the mapping
If
and
are connection coefficients of the spaces
and
, respectively, then
is the deformation tensor of the connection for a mapping
f [
18].
In the paper [
18], we defined biholomorphically projective mappings between two generalized Riemannian spaces
and
with almost complex structures that are equal in a common system of coordinates defined by the mapping
. We considered a generalized Riemannian space
with a non-symmetric metric tensor
and almost complex structure
such that
, where
a is a scalar invariant, and we defined a biholomorphically projective curve of the kind
and a biholomorphically projective mapping of the kind
.
Definition 1 ([
18])
. In the space , a curve l given in parametric formis said to be biholomorphically projective of the kind θ if it satisfies the following equation:where are functions of parameter t, where Definition 2 ([
18])
. A difeomorphism is a biholomorphically projective mapping of the kind if biholomorphically projective curves of the kind of the space are mapped to the biholomorphically projective curves of the kind θ of the space . After defining, we concluded that the biholomorphically projective curves of the first kind and the biholomorphically projective curves of the second kind match, so we simply named them the biholomorphically projective curves, and the corresponding mappings are named biholomorphically projective mappings.
Also, the mapping
is a biholomorphically projective mapping if in the common coordinate system the connection coefficients
and
satisfy the relation [
18]
and the deformation tensor has the form [
18]
where
is a symmetrization without division by indices
,
,
, and
are vectors,
and
is an anti-symmetric tensor. Inspired by the form of the deformation tensor (28), we have defined another type of mapping.
Let
and
be almost complex structures of the spaces
and
, respectively, where
in the common system of coordinates defined by the mapping
, and assume that it holds
where
a is a scalar invariant. The mapping
is a canonical biholomorphically projective mapping if in the common coordinate system the connection coefficients
and
satisfy the relation [
24]
where
is a symmetrization without division by indices
j and
k,
and
are vectors,
and
is an anti-symmetric tensor. If
is a deformation tensor with respect to the canonical biholomorphically projective mapping
, then we have [
24]
The mapping
is an equitorsion canonical biholomorphically projective mapping if the torsion tensors of the spaces
and
are equal in a common coordinate system after the mapping
f. Then, from Ref. [
24],
In this case, the relation (30) becomes [
24]
As a special case of equitorsion canonical biholomorphically projective mappings, the equitorsion mapping
, whose deformation tensor is
was studied. In this case, an invariant of Thomas type [
24]
is obtained for
, assuming that
,
, and
are linearly independent.
The invariant
is the associated basic invariant for mapping
f. From this invariant, one reads
We will express the previous
in some other form, better for further computing
for
After substituting the expression (36) into Equations (17) and (22), assuming that
F is covariantly constant, we obtain the following results:
Accordingly, we conclude that the following theorem is valid.
Theorem 2. Let be an equitorsion canonical biholomorphically projective mapping that preserves the structure F, which is covariantly constant according to the covariant derivative with respect to torsion-free affine connection. The geometrical objects and , given by (38) and (40), respectively, are the associated basic invariant and equitorsion invariant for the mapping f.
Corollary 4. The invariants and satisfy the equation The invariants and satisfy the equation The invariants , , , and are tensors.
Linearly Independent Invariants
The family of invariants
given by (42) depends
u and
, and none of the scalars present in the family of curvature tensors (
8) feature. After changing the corresponding scalars
u and
from linearly independent curvature tensors into the Equation (41), we will obtain the following invariants:
Theorem 3. Three of the invariants (44)–(49) are linearly independent.
Proof. The family of invariants
given by (40) is isomorphic to the vector space, whose elements are
. The fifths
, which generate six linearly independent curvature tensors, are given by (
9). After taking the values of
u and
from these fifths and substituting them into the vectors
v, we obtain the following vectors:
The rank of matrix is 3, which completes the proof of this theorem. □
Corollary 5. Three linearly independent invariants are , , and .
Corollary 6. Three of the invariants are linearly independent. Invariants , , are linearly independent.