Lecture 27: 8.2
Isomorphism
Wei-Ta Chu
Theorem 8.2.2
Theorem 8.2.2: If V is a finite-dimensional vector space, and if T: V → V is a linear operator, then the following statements re equivalent
T is one-to-one
ker(T) = {0} ker(T) = {0}
Example
Show that if V is a finite-dimensional vector space and c is any nonzero scalar, then the linear operator T:V →V defined by T(v)=cv is one-to-one and onto
Solution: The operator T is onto (and hence one-to-one) for if v is any vector in V then that vector is the image of for if v is any vector in V then that vector is the image of the vector (1/c)v
Example
Example
Let be the sequence space, and consider the linear “shifting operators” on V defined by
T1(u1, u2, …, un, …) = (0, u1, u2, …, un, …)
T2(u1, u2, …, un, …) = (u2, u3, …, un, …)
(a) Show that T is one-to-one but not onto. (a) Show that T1 is one-to-one but not onto. (b) Show that T2 is onto but not one-to-one.
Solution (a): The operator T1 is one-to-one because
distinct sequences in obviously have distinct images. This operator is not onto because no vector in maps into the sequence (1, 0, 0, …, 0, …), for example.
Example
Solution (b): The operator T2 is not one-to-one because, for example, the vectors (1, 0, 0, …, 0, …) and (2, 0, 0, …, 0, …) both map to (0, 0, 0, …, 0, …). This operator is onto because every possible sequence of real numbers can be obtained with an appropriate choice of the numbers u , be obtained with an appropriate choice of the numbers u2,
Example
The linear transformations T1: P3 →R4 and T2: M22 →R4 defined by
are both one-to-one and onto.
Verify by showing that their kernels contain only the zero vector.
Example
Let T:Pn →Pn+1 be the linear transformation
T(p)=T(p(x))=xp(x)
If p=p(x)=c0+c1x+…+cnxn and q=q(x)=d0+d1x+…+dnxn
are distinct polynomials, then they are different in at least one coefficient.
one coefficient.
Thus, T(p)=c0x+c1x2+…+c
nxn+1 and
T(q)=d0x+d1x2+…+dnxn+1 also differ in at least one coefficient.
If follows that T is one-to-one since it maps distinct
polynomials p and q into distinct polynomials T(p) and
Dimension and Linear
Transformations
A linear transformation T:V →W in the case where V and
W are finite-dimensional:
If dim(W) < dim(V), then T cannot be one-to-one If dim(V) < dim(W), then T cannot be onto.
Isomorphism
Definition: If a linear transformation T: V→ W is both
one-to-one and onto, then T is said to be an isomorphism, and the vector spaces V and W are said to be isomorphic. The word isomorphic is derived from the Greek words
iso, meaning “identical” and morphe, meaning “form”. iso, meaning “identical” and morphe, meaning “form”.
Isomorphic vector spaces have the same “algebraic form,” even though they may consist of different kinds of
Isomorphism
The isomorphism matches
up vector operations in P2 and R3.
Operation in P2 Operation in R3
Theorem 8.2.3
Theorem 8.2.3: Every real n-dimensional vector space is isomorphic to Rn.
Proof: Let V be a real n-dimensional vector space. To prove that V is isomorphic to Rn we must find a linear
transformation T:V→Rn that is one-to-one and onto. transformation T:V→Rn that is one-to-one and onto. Let v1, v2, …, vn be any basis for V, let u=k1v1 + k2v2 +
…+knvn be the representation of a vector u in V as a linear combination of the basis vectors
Define the transformation T:V→Rn by T(u)=(k1, k2, …,
Theorem 8.2.3
To prove the linearity, let u and v be vectors in V, let c be a scalar, and let u=k1v1+k2v2+…+knvn and
v=d1v1+d2v2+…+dnvn be the representations of u and v as linear combinations of the basis vectors.
Then, Then,
13
Theorem 8.2.3
To show T is one-to-one, we must show that if u and v are distinct vectors in V, then so are their images in Rn. But if
u is not equal to v, and if the representations of these
vector in terms of the basis vectors are described above, then we must have for at least one i. Thus,
then we must have for at least one i. Thus,
which shows that u and v have distinct images under T. The transformation T is onto, for if w=(k1, k2, …kn) is any vector in Rn, then it follows that w is the image under T of
Example
The mappingfrom Pn-1 to Rn is one-to-one, onto, and linear. This is
called the natural isomorphism from Pn-1 to Rn because, as the following computations show, it maps the natural
basis {1, x, x2, …, xn-1} for P into the standard basis for basis {1, x, x2, …, xn-1} for Pn-1 into the standard basis for
Rn:
15
Example
The matrices form a basis for the vector space M22 of 2 by 2 matrices. An isomorphism T: M22 → R4 can be
constructed by first writing a matrix A in M22 in terms of the basis vectors as
and then defining T as T(A) = (a1, a2, a3, a4)
Inner Product Space Isomorphisms
In the case where V is a real n-dimensional inner product space, both V and Rn have a geometric structure arising
from their respective inner products.
It is reasonable to inquire if there exists an isomorphism form V to Rn that preserves the geometric structure as well form V to Rn that preserves the geometric structure as well as the algebraic structure.
For example, we would want orthogonal vectors in V to have orthogonal counterparts in Rn.
Inner Product Space Isomorphisms
In order for an isomorphism to preserve geometric structure, it obviously has to preserve inner products, since notions of length, angle, and orthogonality are all based on the inner product.Thus, if V and W are inner product spaces, then we call an Thus, if V and W are inner product spaces, then we call an isomorphism T:V→W an inner product space
Inner Product Space Isomorphisms
It can be proved that if V is any real n-dimensional inner product space and Rn has the Euclidean inner product,
then there exists an inner product space isomorphism form V to Rn.
Under such an isomorphism, the inner product space V Under such an isomorphism, the inner product space V has the same algebraic and geometric structure as Rn. Every n-dimensional inner product space is a “carbon copy” of Rn with the Euclidean inner product that differs only in the notation used to represent vectors.
Example
Let Rn be the vector space of real n-tuples in
comma-delimited form, let Mn be the vector space of real n by 1 matrices, let Rn have the Euclidean inner product
, and let Mn have the inner product in which u and v are expressed in column form. in which u and v are expressed in column form. The mapping T: Rn → Mn defined by
Lecture 27: 8.3
Compositions and Inverse
Transformations
Transformations
Wei-Ta Chu
Composition of Linear
Transformations
Definition: If T1: U → V and T2: V → W are linear transformations, then the composition of T2 with T1, denoted by T2。T1 (which is read “T2 circle T1”), is the function defined by the formula
Theorem 8.3.1
Theorem 8.3.1: If T1: U → V and T2: V → W are linear transformations, then (T2。T1): U → W is also a linear transformation
Proof: If u and v are vectors in U and c is a scalar, then (T 。T )(u+v) = T (T (u+v)) = T (T (u) + T (v))
(T2。T1)(u+v) = T2(T1(u+v)) = T2(T1(u) + T1(v)) = T2(T1(u)) + T2(T1(v))
= (T2。T1)(u) + (T2。T1)(v) and
(T2。T1)(cu) = T2(T1(cu)) = T2(cT1(u)) =cT2(T1(u)) = c(T2。T1)(u)
Example
Let T1:P1→P2 and T2: P2→P2 be the linear
transformations given by the formulas T1(p(x))=xp(x) and
T2(p(x))=p(2x+4)
Example
If T:V→V is any linear operator, and if I:V →V is the identity operator, then for all vectors v in V, we have (T。I)(v)=T(I(v))=T(v)
(I。T)(v)=I(T(v))=T(v)
It follows that T。I and I。T are the same as T; It follows that T。I and I。T are the same as T; that is T。I =T and I。T = T
More compositions: T1:U→V, T2:V→W, T3:W→Y. The composition (T3。T2。T1)(u) = T3(T2(T1(u)))
Inverse Linear Transformations
A matrix operator TA:Rn→Rn is one-to-one if and only if the matrix A is invertible. If w is the image of a vector x under the operator TA, then x is the image under TA-1 of the vector w.
If T:V→W is a linear transformation, then the range of T, If T:V→W is a linear transformation, then the range of T, denoted by R(T), is the subspace of W consisting of all images under T of vectors in V. If T is one-to-one, then each vector w in R(T) is the image of a unique vector v in
V.
Inverse Linear Transformations
T-1: R(T)→V is a linear transformation. Moreover, T-1(T(v)) = T-1(w) = v
T(T-1(w))=T(v)=w
Example
The linear transformation T:Pn→Pn+1 is given by
T(p)=T(p(x))=xp(x), which is one-to-one
In this case the range of T is not all of Pn+1 but rather the subspace of Pn+1 consisting of polynomials with a zero constant term.
constant term.
T(c0+c1x+…+cnxn) = c0x+c1x2+…+cnxn+1
It follows that T-1:R(T)→Pn is given by
T-1(c0x+c1x2+…+cnxn+1) = c0+c1x+…+cnxn
Example
Let T:R3→R3 be the linear operator defined by
T(x1,x2,x3) = (3x1+x2, -2x1-4x2+3x3, 5x1+4x2-2x3)
Determine whether T is one-to-one; if so, find T-1(x1,x2,x3) Solution: The standard matrix for T is
This matrix is invertible.
Example
Therefore, T-1(x
1,x2,x3)
Composition of One-to-One Linear
Transformations
Theorem 8.3.2: If T1:U→V and T2:V→W are one-to-one linear transformations, then
T2。T1 is one-to-one (T2。T1)-1 = T1-1 。T2-1
Proof(a): We want to show that T 。T maps distinct Proof(a): We want to show that T2。T1 maps distinct vectors in U into distinct vectors in W. If u and v are distinct vectors in U, then T1(u) and T1(v) are distinct vectors in V since T1 is one-to-one. The fact that T2 is one-to-one imply that T2(T1(u)) and T2(T1(v)) are also
distinct vectors. These can also expressed as (T2。T1 )(u) and (T2。T1 )(v) so T2。T1 maps u and v into distinct
vectors in W.
Composition of One-to-One Linear
Transformations
Proof(b): We want to show (T2。T1)-1 (w) = (T
1-1。T2-1)
(w).
Let u= (T2。T1)-1 (w). Then, (T
2。T1)(u) = w, or
equivalently, T2(T1(u))=w. Taking T2-1 of each side of this
equation, then taking T -1 of each side of the result, then equation, then taking T1-1 of each side of the result, then
u=T1-1(T2-1(w)), or equivalently, u= (T1-1。T2-1)(w).