Upload
others
View
24
Download
0
Embed Size (px)
Citation preview
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9. General Vector Spaces9. General Vector Spaces
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.1. Vector Space Axioms9.1. Vector Space Axioms
Vector Space Axioms
The 10 properties in this definition are called the vector space axioms, and a set V with two operations that satisfy these 10 axioms is called a vector space.
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.1. Vector Space Axioms9.1. Vector Space Axioms
Function Spaces
If u=(u1, u2, …, un) is a vector in Rn, then we can regard the components of u to be the values of a function f whose independent variable varies over the integers from 1 to n; that is,
To continue the analogy between n-tuples and functions, let us denote the set of real-valued functions that are defined for all real values of x by F(-∞, ∞).f=g if and only if f(x)=g(x) for all real values of x
(cf)(x)=cf(x)
(f+g)(x)=f(x)+g(x)
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.1. Vector Space Axioms9.1. Vector Space Axioms
Subspaces
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.1. Vector Space Axioms9.1. Vector Space Axioms
Subspaces
Let n be a nonnegative integer, and let Pn be the set of all real-valued functions of the formExample 1Example 1
where a0, a1, …, an are real numbers; that is, Pn is the set of all polynomials of degree n or less. Show that Pn is a subspace of F(-∞, ∞).
Polynomials are defined for all real values of x, so Pn is a subset of F(-∞, ∞).
Let and
Thus,
and
which shows that cp and p+q are polynomials of degree n or less.
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.1. Vector Space Axioms9.1. Vector Space Axioms
Subspaces
Cm(-∞, ∞): the real-valued functions with continuous mth derivatives
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.1. Vector Space Axioms9.1. Vector Space Axioms
Subspaces
The subspace of F(-∞, ∞) that is spanned by the functions 1, x, x2, …, xn consists of all linear combinations of these functions and hence consists of all functions of the form
Example 11, 12Example 11, 12
These are the polynomials of degree n or less, so span{1, x, x2, …, xn}=Pn
These functions are linearly independent and hence form a basis for Pn.
This is called the standard basis for Pn.
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.1. Vector Space Axioms9.1. Vector Space Axioms
Wroński’s Test for linear Independence of Functions
Suppose that f1(x), f2(x), …, fn(x) are functions in C(n-1)(-∞, ∞).
If these functions are linearly dependent, then there exist scalars k1, k2, …, kn that are not all zero and such that
for all x in the interval (-∞, ∞).
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.1. Vector Space Axioms9.1. Vector Space Axioms
Wroński’s Test for linear Independence of Functions
which is called the Wronskian of f1(x), f2(x), …, fn(x), is zero for every x in the interval (-∞, ∞).
This implies that the linear system
has a nontrivial solution for every x in the interval (-∞, ∞).
And this, in turn, implies that the determinant
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.1. Vector Space Axioms9.1. Vector Space Axioms
Wroński’s Test for linear Independence of Functions
This function is nonzero for all real values of x and hence is not identically zero on (-∞, ∞). Thus, the functions are linearly independent.
Example 14Example 14f1(x)=1, f2(x)=ex, and f3(x)=e2x
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.1. Vector Space Axioms9.1. Vector Space Axioms
Dimension
We saw in Example 12 that the function 1, x, x2, …, xn form a basis for Pn. This implies that all bases for Pn have n+1 vectors and hence that dim(Pn)=n+1.
Also, we showed above that P∞ is infinite-dimensional, so it follows that F(-∞,∞), C(-∞,∞), C1(-∞,∞), Cm(-∞,∞), and C∞(-∞,∞) are also infinite-dimensional.
Example 15Example 15
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.2. Inner Product Spaces; Fourier Series9.2. Inner Product Spaces; Fourier Series
Inner Product Axioms
To give a vector space a geometric structure we need to have a third operation that will generalize the concept of dot product.
length, distance, angle, and orthogonality
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.2. Inner Product Spaces; Fourier Series9.2. Inner Product Spaces; Fourier Series
Inner Product Axioms
We make the following definition whose statements duplicate those in Theorem 1.2.6, but in different notation.
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.2. Inner Product Spaces; Fourier Series9.2. Inner Product Spaces; Fourier Series
Inner Product Axioms
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.2. Inner Product Spaces; Fourier Series9.2. Inner Product Spaces; Fourier Series
The Effect of Weighting on Geometry
If f and g are continuous functions on the interval [a, b], then the formulaExample 4 (The Integral Inner Product)Example 4 (The Integral Inner Product)
defines an inner product on the vector space C[a, b] that we will call the integral inner product.
The norm of a function f relative to this inner product is
(7)
(8)
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.2. Inner Product Spaces; Fourier Series9.2. Inner Product Spaces; Fourier Series
Axiom 1 – If f and g are continuous functions on [a, b], thenExample 4 (The Integral Inner Product)Example 4 (The Integral Inner Product)
Axiom 2 – If f, g, and h are continuous functions on [a, b], then
Axiom 3 – If f and g are continuous functions on [a, b] and k is a scalar, then
Axiom 4 – If f is a continuous function on [a, b], then
The Effect of Weighting on Geometry
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.2. Inner Product Spaces; Fourier Series9.2. Inner Product Spaces; Fourier Series
If p and q are distinct positive integers, then the functions cospx and cosqx are orthogonal with respect to the inner product
Example 5Example 5
The Effect of Weighting on Geometry
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.2. Inner Product Spaces; Fourier Series9.2. Inner Product Spaces; Fourier Series
Algebraic Properties of Inner Products
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.2. Inner Product Spaces; Fourier Series9.2. Inner Product Spaces; Fourier Series
Algebraic Properties of Inner Products
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.2. Inner Product Spaces; Fourier Series9.2. Inner Product Spaces; Fourier Series
Algebraic Properties of Inner Products
REMARK
From (30) of Section 1.2 for the angle θ between nonzero vectors u and v can be extended to general inner product spaces are
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.2. Inner Product Spaces; Fourier Series9.2. Inner Product Spaces; Fourier Series
Orthogonal Bases
Recall that Theorem 7.9.1 that an orthogonal set of nonzero vectors in Rn is linearly independent. The same is true in a general inner product space V.
Example 8Example 8
Trigonometric polynomial
If cn and dn are not both zero, the f(x) is said to have order n.
The set of all trigonometric polynomials of order n or less is the subspace of C(-∞,∞) that is spanned by the functions in the set
We will denote this subspace by Tn.
Show that S is an orthogonal basis for Tn with respect to the integral inner product.
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.2. Inner Product Spaces; Fourier Series9.2. Inner Product Spaces; Fourier Series
Orthogonal Bases
The set S spans Tn.
And, S is an orthogonal set:
Example 8Example 8
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.2. Inner Product Spaces; Fourier Series9.2. Inner Product Spaces; Fourier Series
Orthogonal Bases
Find an orthonormal basis for Tn relative to the inner product.Example 9Example 9
Thus, normalizing the vectors in S yields the orthonormal basis
(21)
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.2. Inner Product Spaces; Fourier Series9.2. Inner Product Spaces; Fourier Series
Best Approximation
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.2. Inner Product Spaces; Fourier Series9.2. Inner Product Spaces; Fourier Series
Fourier Series
Let f be a continuous function on the interval [0, 2π].
We can compute the best mean square approximation to f by a trigonometric polynomial of order n or less. We will use the orthonormal basis (21) in Example 9.
Let the orthogonal projection of f on Tn be
(26)
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.2. Inner Product Spaces; Fourier Series9.2. Inner Product Spaces; Fourier Series
Fourier Series
Thus,
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.2. Inner Product Spaces; Fourier Series9.2. Inner Product Spaces; Fourier Series
Fourier Series
or more briefly,
The numbers a0, a1, …, an, b1, …, bn are called the Fourier coefficients of f, and (26) is called the nth-order Fourier approximation of f.
It can be proved that if f is continuous on the interval [0, 2π], then the mean square error in the nth-order Fourier approximation of f approaches zero as n∞. This is denoted by writing
which is called the Fourier series for f.
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.2. Inner Product Spaces; Fourier Series9.2. Inner Product Spaces; Fourier Series
General Inner Products on Rn
Axiom 1 – Since uTAv is a 1×1 matrix, it is symmetric and hence
Axiom 2 – Using properties of the transpose we obtain
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.2. Inner Product Spaces; Fourier Series9.2. Inner Product Spaces; Fourier Series
General Inner Products on Rn
and
Axiom 3 – Again using properties of the transpose we obtain
Axiom 4 – Since A is positive definite and symmetric, the expression <v, v>=vTAv is a positive definite quadratic form, and hence Definition 8.4.2 implies that
if and only if
Weighted Euclidean inner product
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.3. General Linear Transformations; Isomorphism9.3. General Linear Transformations; Isomorphism
General Linear Transformations
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.3. General Linear Transformations; Isomorphism9.3. General Linear Transformations; Isomorphism
Kernel and Range
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.3. General Linear Transformations; Isomorphism9.3. General Linear Transformations; Isomorphism
Properties of The Kernel and Range
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.3. General Linear Transformations; Isomorphism9.3. General Linear Transformations; Isomorphism
Isomorphism
We will show that Rn is “the mother of all real finite-dimensional vector spaces” in the sense that every real n-dimensional vector space differs from Rn only in the notation used to represent vectors.
The fact that the transformation T between a vector space V and n-tuples in Rn is one-to-one, onto, and linear means that it matches up V with Rn in such a way that operations on vectors in either space can be performed using their counterparts in the other space.
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.3. General Linear Transformations; Isomorphism9.3. General Linear Transformations; Isomorphism
Thus, although a polynomial a0+a1x+a2x2 is obviously a different mathematical object from an ordered triple (a0, a1, a2), the vector spaces formed by these objects have the same algebraic structure.
Example 17Example 17The following table shows how the transformation
matches up vector operations in P2 and R3.
Isomorphism
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.3. General Linear Transformations; Isomorphism9.3. General Linear Transformations; Isomorphism
The fact that every real n-dimensional vector space is isomorphic to Rn makes it possible to apply theorems about vectors in Rn to general finite-dimensional vector spaces.
Isomorphism
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.3. General Linear Transformations; Isomorphism9.3. General Linear Transformations; Isomorphism
Consider the differentiation operatorExample 20Example 20
on the vector space of polynomials of degree 3 or less.
If we map P3 and P2 into R4 and R3, respectively, by the natural isomorphisms, then the transformation D produces a corresponding transformation
For example,
Isomorphism
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.3. General Linear Transformations; Isomorphism9.3. General Linear Transformations; Isomorphism
Since Example 20Example 20Example 20Example 20
and hence the standard matrix for T is
Isomorphism
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.3. General Linear Transformations; Isomorphism9.3. General Linear Transformations; Isomorphism
This matrix performs the differentiation Example 20Example 20Example 20Example 20
by operating on the images of the polynomials under the natural isomorphisms, as confirmed by the computation
Isomorphism
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.3. General Linear Transformations; Isomorphism9.3. General Linear Transformations; Isomorphism
Inner Product Space Isomorphisms
If V and W are inner product spaces, then we call an isomorphism T:VW an inner product space isomorphisom if
It can be proved that if V is any real n-dimensional inner product space and Rn has the Euclidean inner product (the dot product), then there exists an inner product space isomorphism from V to Rn.
Under such an isomorphism, the inner product space V has the same algebraic and geometric structure as Rn.
E-mail: [email protected]://web.yonsei.ac.kr/hgjung
9.3. General Linear Transformations; Isomorphism9.3. General Linear Transformations; Isomorphism
Let Rn be the vector space of real n-tuples in comma-delimited form, let Mn be the vector space of real n×1 matrices,
let Rn have the Euclidean inner product <u, v>=u∙v,
and let Mn have the inner product <u, v>=uTv in which u and v are expressed in column form.
The mapping T:RnMn defined by
Example 21Example 21
is an inner product space isomorphism, so the distinction between the inner product space Rn
and the inner product space Mn is essentially notational, a fact that we have used many times in this text.
Inner Product Space Isomorphisms