Sie sind auf Seite 1von 6

A Brief Introduction to Tensors

Jay R. Walton
Fall 2006
1 Preliminaries
In general, a tensor is a multilinear transformation dened over an underlying nite dimensional
vector space. In this brief introduction, tensor spaces of all integral orders will dened inductively.
Initially the underlying vector space, 1, will be assumed to be an inner product space in order to
simplify the discussion. Subsequently, the presentation will be generalized to vector spaces without
inner product. Usually, bold-face letters, a, will denote vectors in 1 and upper case letters, A, will
denote tensors. The inner product on 1 will be denoted by a b.
2 Tensors over an Inner Product Space
First tensor spaces are developed over an inner product vector space 1, . Implicit in the discussion
is use of an orthonormal basis in 1 for deriving component representations of vectors and tensors.
Zero-Order Tensors. The space of Zero-Order Tensors, T
0
, is isomorphic to the scalar eld,
T, corresponding to the underlying vector space 1, which in this course will be either the real or
complex numbers, 1 or (, respectively. A zero order tensor, T
0
, acts as a linear transformation
from T
0
to T
0
[] : T
0
T
0
via multiplication of scalars. That is, given T
0
,
[] := .
First-Order Tensors. The space of First-Order Tensors, T
1
, is isomorphic to the underlying
vector space, 1. A rst order tensor, a T
1
, acts as a linear transformation from T
0
T
1
and
from T
1
T
0
as follows. In the rst instance, for all T
0
,
a[] := a (1)
whereas in the second instance, for all b T
1
,
a[b] := a b. (2)
It should be noticed, that both (1) and (2) are bi-linear forms, i.e. they are linear forms in each of
their two independent variables separately.

Copyright c 2006 by J. R. Walton. All rights reserved.


1
Second-Order Tensors. The space of Second-Order Tensors, T
2
, is isomorphic to the space,
Lin[1], of linear transformations A : 1 1. Elements in T
2
act as linear transformations from
T
i
T
j
with i, j = 0, 1, 2 subject to i + j = 2. For i = 0, j = 2 one has for A T
2
and T
0
A[] := A. (3)
For i = 1, j = 1, one has for A T
2
and a T
1
A[a] := Aa (4)
where the expression of the right hand side of (4) denotes the action of the linear transformation
A Lin[1] on the vector a 1. For i = 2, j = 0, one has for A, B T
2
A[B] := A B (5)
where A B denotes the natural inner product on Lin[1] dened by
A B := tr[A
T
B]. (6)
In (6), tr[A] denotes the trace of A Lin[1] and A
T
denotes the transpose of A.
Finally, a second-order tensor A T
2
can be used to dene a bi-linear transformation on 1.
Specically, A[, ] : T
1
T
1
T
0
is dened by
a, b)
A
:= a (A[b]) (7)
for all a, b T
1
.
An important class of second order tensors is given by the Elementary Tensor Product of two
rst order tensors. Specically, given a, b T
1
the elementary tensor product a b of a and b is
the second order tensor whose action on a rst order tensor c T
1
is dened by
a b[c] := a(b c). (8)
From the denitions (6), (8), one sees that
[a b]
T
= b a
Tr[a b] = a b
a b c d = (a c)(b d)
a bc d = (b c)a d.
Coordinates with Respect to an Orthonormal Basis. Given an orthonormal basis, e
1
, . . . , e
n
,
for the underlying vector space 1, one can construct a natural orthonormal basis for the space
of second order tensors, T
2
, of the form
e
i
e
j
, i, j = 1, . . . , n. (9)
Consequently, given A T
2
, one has
A =

i,j
a
ij
e
i
e
j
2
with the coordinates of A relative to the natural basis given by
a
ij
= A e
i
e
j
= e
i
(Ae
j
).
Hence, if x 1 has coordinates x
k
relative to the basis e
1
, . . . , e
n
, then the action of A on x can
be computed using components
[Ax] = [a
ij
x
j
]
where summation over the index j is implied. It is useful to note that one easily shows that if
a, b 1 have coordinates [a] = [a
i
] and [b] = [b
i
], respectively, relative to the orthonormal basis
e
1
, . . . , e
n
, then the components of the second-order tensor a b relative to the natural basis on
T
2
are
[a b] = [a
i
b
j
].
Finally, the component form of the bi-linear form (7) is
a, b)
A
= A a b = a (Ab) = a
ij
a
i
b
j
where summation over i, j = 1, . . . , n is implied.
Third-Order Tensors. The space of third-order tensors, T
3
, is most easily constructed by rst
considering elementary tensor products of the form a b c for rst-order tensors (vectors in
1) a, b, c T
1
. A third-order tensor can be used to dene a linear transformation from T
p

T
3p
for p = 0, 1, 2, 3. The action of a third-order elementary tensor product as such a linear
transformation can be completely specied by dening its action on p
th
order elementary tensor
products. (Why?) For p = 0 and T
0
, one denes
a b c[] := a b c T
3
.
For p = 1 and d T
1
, one denes
a b c[d] := (c d)a b T
2
.
For p = 2 and d e T
2
, one denes
a b c[d e] := (b d)(c e)a T
1
.
It should be noted that in this expression, the scalar multiplying a can also be written as bcde,
where now denotes the dot-product on T
2
dened previously.
Finally, for d e f T
3
, one denes
a b c[d e f ] := (a d)(b e)(c f ). (10)
Note that this last expression is a multi-linear form on the six variables a, . . . , f . Moreover, this
last expression can be used to dene an inner product on the space T
3
. Indeed, as a natural basis
for T
3
one takes the set of elementary tensor products, e
i
e
j
e
k
, i, j, k = 1 . . . , n, and then
denes the dot-product on T
3
using (10) to dene
e
i
e
j
e
k
e
l
e
m
e
n
:= (e
i
e
l
)(e
j
e
m
)(e
k
e
n
) =
il

jm

kn
where
ij
denotes the Kronecker symbol

ij
=
_
1 when i = j,
0 when i ,= j
3
and extending the denition to all of T
3
by linearity. In particular, a general third order tensor
A T
3
has the component representation A = [a
ijk
] where the a
ijk
are dene through
A = a
ijk
e
i
e
j
e
k
where summation over i, j, k is implied.
Fourth and Higher-Order Tensors. The generalization to fourth-order tensors and higher
should now be clear. One rst denes the special class of N
th
-order elementary tensor products
of rst-order tensors, and then uses the dot product to dene their various actions as multi-linear
transformations. The vector space of all N
th
or tensors is then constructed by taking all nite
linear combinations of such N
th
order elementary tensor products.
For example, an N
th
order tensor elementary tensor product of the form
A = a
1
. . . a
Np
b
1
. . . b
p
denes a multilinear transformation A : T
p
T
Np
through
A[c
1
. . . c
p
] = a
1
. . . a
Np
(b
1
c
1
) . . . (b
p
c
p
).
An important fourth-order tensor in applications is the Elasticity Tensor of linear elasticity
theory. Specically, the elasticity tensor, D, is the fourth-order tensor by which the stress tensor,
T, is computed from the innitessimal strain tensor, E, as
T = D[E]
which in component form becomes
t
ij
= d
ijkl
e
kl
where summation over k, l = 1, 2, 3 is implied.
3 Tensors over a Vector Space without Inner Product
The construction of tensor spaces of all orders given below proceeds in somewhat the same fashion
as done previously, only now the underlying vector space, 1, is not assumed to have an inner
product, In particular, the term orthonormal basis has no meaning in this context. However, many
of the conveniences of an orthonormal basis can be realized through the introduction of the notion
of a Dual Space to 1.
3.1 Dual Space and Dual Basis
3.1.1 Dual Space
Given a nite dimensional vector space 1, one denes its Dual Space 1

to be Lin[1, R], the vector


space of all linear transformations from 1 to the real numbers (or more generally, to the associated
scalar eld T). Recall that if 1 has dimension N, then Lin[1, R] can be realized as all 1 N
matrices with real entries. The action of a linear transformation a

on a vector b 1 is
denoted by a

, b).
Example. Let 1 = R
N
(ignoring its natural inner product). Elements a 1 are n-tuples of real
numbers
a =
_
_
_
a
1
.
.
.
a
N
_
_
_
4
whereas elements b

are 1 N matrices
b

=
_
b
1
. . . b
N
_
.
The action b

, a) is then given by matrix multiplication


b

, a) =
_
b
1
. . . b
N
_
_
_
_
a
1
.
.
.
a
N
_
_
_ = b
1
a
1
+ . . . + b
N
a
N
.
3.1.2 Dual Basis
Given a basis B = e
1
, . . . , e
N
for 1, one denes its Dual Basis to be the unique basis B

=
e
1
, . . . , e
N
for the dual space satisfying
e
i
, e
j
) =
i
j
, for i, j = 1, . . . , N (11)
where
i
j
denotes the Kronecker symbol. Every vector a 1 has a representation a = a
1
e
1
+ . . . +
a
N
e
N
. The coecients a
i
, i = 1, . . . , N are called the Contravariant Coordinates of the vector
a 1. Correspondingly, every dual vector b

has a representation b

= b
1
e
1
+ . . . + b
N
e
N
,
with the coecients b
i
, i = 1, . . . , N being called the Covariant Coordinates of b

. It follows from
(11) that
a
i
= e
i
, a) and b
i
= b

, e
i
). (12)
3.2 The Tensor Spaces
The tensor space T
p
q
(1) is dened to be the vector space of all (p + q)-multilinear, real-valued
functions
A : 1

. . . 1

. .
ptimes
1 . . . 1
. .
qtimes
R. (13)
Thus, A is a function of p-variables from 1

and q-variables from 1 that is linear in each variable


separately. The Contravariant Order of A is p and the Covariant Order of A is q. A Pure
Contravariant Tensor has order (p, 0) while a Pure Covariant Tensor has order (0, q).
Example. Every transformation A Lin(1) denes a tensor

A T
1
1
through

A(v

, v) = v

, Av)
for every v

and v 1.
Example. Any p-vectors from 1 and q-dual vectors from 1

can be used to construct a tensor in


T
p
q
in the form of a tensor product. More specically, if v
1
, . . . , v
p
1 and v
1
, . . . , v
q
1

, then
one denes the tensor product v
1
. . . v
p
v
1
. . . v
q
T
p
q
through the action
v
1
. . . v
p
v
1
. . . v
q
(u
1
, . . . , u
p
, u
1
, . . . , u
q
) = u
1
, v
1
) u
p
, v
p
)
v
1
, u
1
) v
q
, u
q
).
Given a basis B = e
1
, . . . , e
N
for 1 with associated dual basis B

= e
1
, . . . , e
N
for 1

, one
constructs the natual product basis for T
p
q
as
e
i
1
. . . e
ip
e
j
1
. . . e
jq
, i
1
, . . . , i
p
, j
1
, . . . , j
q
= 1, . . . , N.
5
Thus, one sees the dim(T
p
q
) = N
(p+q)
where dim(1) = N. It is now straight forward to construct
the component form of a general tensor. Hence, for a tensor A T
p
q
, one denes its component
form relative to the natural product basis
[A] =
_
a
i
1
...ip
j
1
...jq
_
through the following argument. Let u
1
, . . . , u
p
1

have covariant coordinates [u


i
]
B
= [u
i
k
], i =
1, . . . , p, k = 1, . . . , N and let u
1
, . . . , u
q
1 have contravariant coordinates [u
j
]
B
= [u
k
j
], j =
1, . . . , q, k = 1, . . . , N. Then,
A(u
1
, . . . , u
p
, u
1
, . . . , u
q
) = a
i
1
...ip
j
1
...jq
e
i
1
. . . e
ip
e
j
1
. . . e
jq
(u
1
, . . . , u
p
, u
1
, . . . , u
q
)
= a
i
1
...ip
j
1
...jq
u
1
, e
i
1
, ) u
p
, e
ip
)e
j
1
, u
1
) e
jq
, u
q
)
= a
i
1
...ip
j
1
...jq
u
1
i
1
u
p
ip
u
j
1
1
u
jq
q
.
Generalized Tensor Product. There is a useful generalization of the elementary tensor product
to tensors of arbitrary order. Specically, given A T
p
q
and B T
r
s
, one denes the tensor product
AB T
p+r
q+s
through the action
AB(v
1
, . . . , v
p+r
, v
1
, . . . , v
q+s
) := A(v
1
, . . . , v
p
, v
1
, . . . , v
q
)
B(v
p+1
, . . . , v
p+r
, v
q+1
, . . . , v
q+s
).
Thus the orders of tensors add in forming tensor products.
Generalized Contraction. Similarly, it is useful to introduce a generalization of the dot product
(contraction operator) to higher order tensors. To that end, let A T
p
q
and B T
s
r
with r p
and s q. One then denes the dot product A B to be a tensor in T
pr
qs
given (in component
form) by
[A B] :=
_
a
i
1
...ip
j
1
...jq
b
j
qs+1
...jq
i
pr+1
...ip
_
.
Thus, the orders of tensors subtract in the generalized dot product.
6

Das könnte Ihnen auch gefallen