1. Foundations of Numerics from Advanced Mathematics ... · another is a Cube minipulation, again....

Preview:

Citation preview

TU Munchen

1. Foundations of Numerics from Advanced Mathematics

Linear Algebra

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 1

TU Munchen

1.1. Linear Algebra

Mathematical Structures

• a mathematical structure consists of one or several sets and one or severaloperations defined on the set(s)

• special elements:

– neutral element (of an operation)– inverse element (of some element x)

• a group: a structure to add and subtract

• a field: a structure to add, subtract, multiply, and divide

• a vector space: a set of vectors over a field with two operations: scalarmultiplication, addition of vectors, obeying certain axioms (which?)

• note: sometimes, the association with classical (geometric) vectors is helpful,sometimes it is more harmful

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 2

TU Munchen

Exercise Mathematical Structures

Show that the possible manipulations ofthe Rubik’s Cube with the operation ’ex-ecute after’ are a group.

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 3

TU Munchen

Exercise Mathematical Structures – Solution

Show that the possible manipulations ofthe Rubik’s Cube with the operation ’ex-ecute after’ are a group.

• Closure: executing any two manipulations after oneanother is a Cube minipulation, again.

• Associativity: the result of a sequence of threemanipulations is obviously always the same no matter howyou group them (the first two or the last two together).

• Identity: obviously included (just ’do nothing’).• Invertibility: execute a manipulation in backward direction.

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 4

TU Munchen

Exercise Mathematical Structures

Show that the rational numbers with the operations + (add) and∗ (multiply) are a field.

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 5

TU Munchen

Exercise Mathematical Structures – Solution

Show that the rational numbers with the operations + (add) and∗ (multiply) are a field.

• Closure: obviously closed under + and ∗.• Identity: 0 for +, 1 for ∗.• Invertibility: each element q has an inverse −q under +

and 1q under ∗. The latter holds for all elements except from

the neutral element of +, i.e., 0.• Associativity: well-known for both + and ∗.• Commutativity: also known from school (a + b = b + a,

a ∗ b = b ∗ a).• Distributivity: dito (a ∗ (b + c) = a ∗ b + a ∗ c).

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 6

TU Munchen

Exercise Mathematical Structures

Is the set of N ×N matrices (N ∈ N) matrices with real numbersas entries over the field of real numbers a vector space?

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 7

TU Munchen

Exercise Mathematical Structures – Solution

Is the set of N ×N matrices (N ∈ N) matrices with real numbersas entries over the field of real numbers a vector space?

The answer is yes. Look up the axioms and show that they holdfor the xample on your own.

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 8

TU Munchen

Vector Spaces

• a linear combination of vectors

• linear (in)dependence of a set of vectors

• the span of a set of vectors

• a basis of a vector space

– definition?– why do we need a basis?– is a vector’s basis representation unique?– is there only one basis for a vector space?

• the dimension of a vector space

• does infinite dimensionality exist?

• important applications:

– (analytic) geometry– numerical and functional analysis: function spaces are vector spaces

(frequently named after mathematicians: Banach spaces, Hilbert spaces,Sobolev spaces, ...)

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 9

TU Munchen

Exercise Vector Spaces

Is the set of vectors{(

10

),

(01

),

(13

)}linearly

independent?

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 10

TU Munchen

Exercise Vector Spaces – Solution

Is the set of vectors{(

10

),

(01

),

(13

)}linearly

independent?

The set of vectors is not linearly independent, since the thirdelement can easily be written as a linear combination of the firsttwo:(

13

)= 1 ·

(10

)+ 3 ·

(01

).

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 11

TU Munchen

Exercise Vector Spaces

span

{( 100

),

( 001

)}= ?

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 12

TU Munchen

Exercise Vector Spaces – Solution

span

{( 100

),

( 001

)}=

{( a0b

);a,b ∈ R

}.

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 13

TU Munchen

Exercise Vector Spaces

Consider the set of all possible polynomials with realcoefficients as a vector space over the field of real numbers.What’s the dimension of this space? Give a basis.

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 14

TU Munchen

Exercise Vector Spaces – Solution

Consider the set of all possible polynomials with realcoefficients as a vector space over the field of real numbers.What’s the dimension of this space? Give a basis.

The space is infinite dimensional, a basis is for example{1, x , x2, x3, . . .

}.

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 15

TU Munchen

Linear Mappings

• definition in the vector space context; notion of a homomorphism

• image and kernel of a homomorphism

• matrices, transposed and Hermitian of a matrix

• relations of matrices and homomorphisms

• meaning of injective, surjective, and bijective for a matrix; rank of a matrix

• meaning of the matrix columns for the underlying mapping

• matrices and systems of linear equations

• basis transformation and coordinate transformation

• mono-, epi-, iso-, endo-, and automorphisms

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 16

TU Munchen

Exercise Linear Mappings

Is the mapping f : R3 → R3, ~x 7→ 5 · ~x +

( 123

)linear?

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 17

TU Munchen

Exercise Linear Mappings – Solution

Is the mapping f : R3 → R3, ~x 7→ 5 · ~x +

( 123

)linear?

f is not linear, since f (α~x) 6= αf (~x).

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 18

TU Munchen

Exercise Linear Mappings

What’s the linear mapping f : R2 → R2 corresponding to the

matrix(

4 03 2

)?

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 19

TU Munchen

Exercise Linear Mappings – Solution

What’s the linear mapping f : R2 → R2 corresponding to the

matrix(

4 03 2

)?

f((

xy

))=

(4x

3x + 2y

).

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 20

TU Munchen

Exercise Linear Mappings

Give the rank of the matrix

1 0 0 00 1 0 00 0 0 00 0 0 1

.

Is the corresponding linear mapping injective, surjective,bijective?

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 21

TU Munchen

Exercise Linear Mappings – Solution

Give the rank of the matrix

1 0 0 00 1 0 00 0 0 00 0 0 1

.

Is the corresponding linear mapping injective, surjective,bijective?

The rank is three. Thus, the corresponding linear mapping isneither injective, nor surjective or bijective.

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 22

TU Munchen

Examples Linear Mappings

Monomorphism:

( 0 10 01 0

)

Epimorphism:

( 1 0 00 1 00 0 0

)

Iso-/Automorphism:(

0 11 0

)

Endomorphism:

( 2 1 00 1 21 0 1

).

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 23

TU Munchen

Determinants

• definition

• properties

• meaning

• occurrences

• Cramer’s rule

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 24

TU Munchen

Determinants – Definition

det(A) =

∣∣∣∣∣∣∣∣∣a1,1 a1,2 · · · a1,Na2,1 a2,2 · · · a2,N

.... . .

...aN,1 aN,N

∣∣∣∣∣∣∣∣∣ =

a1,1

∣∣∣∣∣∣∣∣∣∣∣

a2,2 · · · · · · a2,N...

......

...aN,2 · · · · · · aN,N

∣∣∣∣∣∣∣∣∣∣∣− a1,2

∣∣∣∣∣∣∣∣∣a2,1 a2,3 · · · a2,Na3,1 a3,N

......

aN,1 aN,3 · · · aN,N

∣∣∣∣∣∣∣∣∣+ . . .

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 25

TU Munchen

Exercise Determinants

det(A) = 0⇒ A defines a . . .morphism.

det(A) 6= 0⇒ A defines a . . .morphism.

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 26

TU Munchen

Exercise Determinants – Solution

det(A) = 0⇒ A defines an Endomorphism.

det(A) 6= 0⇒ A defines an Automorphism.

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 27

TU Munchen

Exercise Determinants

det(A · B) =?

det(A−1) =?

det(AT ) =?

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 28

TU Munchen

Exercise Determinants – Solution

det(A · B) = det(A) · det(B).

det(A−1) = det(A)−1.

det(AT ) = det(A).

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 29

TU Munchen

Exercise Determinants

Determine the solution of the linear system

2x1 + x2 = 42x2 + x3 = 0

x1 + x2 + x3 = 3

with the help of determinants.

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 30

TU Munchen

Exercise Determinants – Solution

Determine the solution of the linear system

2x1 + x2 = 42x2 + x3 = 0

x1 + x2 + x3 = 3

with the help of determinants.

x1 =

∣∣∣∣∣∣∣∣4 1 00 2 13 1 1

∣∣∣∣∣∣∣∣∣∣∣∣∣∣∣∣2 1 00 2 11 1 1

∣∣∣∣∣∣∣∣= 7

3 ; x2 =

∣∣∣∣∣∣∣∣2 4 00 0 11 3 1

∣∣∣∣∣∣∣∣∣∣∣∣∣∣∣∣2 1 00 2 11 1 1

∣∣∣∣∣∣∣∣= −7

3 ; x3 =

∣∣∣∣∣∣∣∣2 1 40 2 01 1 3

∣∣∣∣∣∣∣∣∣∣∣∣∣∣∣∣2 1 00 2 11 1 1

∣∣∣∣∣∣∣∣= 4

3 .

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 31

TU Munchen

Eigenvalues

• notions of eigenvalue, eigenvector, and spectrum

• similar matrices A,B:∃S : B = SAS−1

(i.e.: A and B as two basis representations of the same endomorphism)

• resulting objective: look for the best / cheapest representation (diagonal form)

• important: matrix A is diagonalizable iff there is a basis consisting ofeigenvectors only

• characteristic polynomial, its roots are the eigenvalues

• Jordan normal form

• important:

– spectrum characterizes a matrix– many situations / applications where eigenvalues are crucial

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 32

TU Munchen

Exercise Eigenvalues

Diagonalize the matrix(

3 22 3

). Give both eigenvalues and

eigenvectors and the basis transformation matrix transformingthe given matrix in diagonal form.

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 33

TU Munchen

Exercise Eigenvalues – Solution

Diagonalize the matrix(

3 22 3

). Give both eigenvalues and

eigenvectors and the basis transformation matrix transformingthe given matrix in diagonal form.

Eigenvalues:∣∣∣∣ 3− λ 22 3− λ

∣∣∣∣ = 9− 6λ+ λ2 − 4 = 5− 6λ+ λ2

⇒ λ1,2 = 6±√

36−202 = 3± 2⇒ λ1 = 5, λ2 = 1.

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 34

TU Munchen

Exercise Eigenvalues – Solution

Diagonalize the matrix(

3 22 3

). Give both eigenvalues and

eigenvectors and the basis transformation matrix transformingthe given matrix in diagonal form.

Eigenvector for λ1 = 5:(−2 22 −2

)(xy

)=

(00

)⇔ x = y ⇒ ~x1 =

(11

)Eigenvector for λ2 = 1:

(2 22 2

)(xy

)=

(00

)⇔ x = −y ⇒ ~x2 =

(1−1

)

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 35

TU Munchen

Exercise Eigenvalues – Solution

Diagonalize the matrix(

3 22 3

). Give both eigenvalues and

eigenvectors and the basis transformation matrix transformingthe given matrix in diagonal form.

The basis transformation matrix thus is(

1 11 −1

)and results in the diagonal matrix(

5 00 1

).

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 36

TU Munchen

Scalar Products and Vector Norms

• notions of a linear form and a bilinear form

• scalar product: a positive-definite symmetric bilinear form

• examples of vector spaces and scalar products

• vector norms:

– definition: positivity, homogeneity, triangle inequality– meaning of triangle inequality– examples: Euclidean, maximum, and sum norm

• normed vector spaces

• Cauchy-Schwarz inequality

• notions of orthogonality and orthonormality

• turning a basis into an orthonormal one: Gram-Schmidt orthogonalization

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 37

TU Munchen

Exercise Scalar Products and Vector Norms

Are the following operators scalar products in the vector spaceof continuous functions on the interval [a;b]?

〈f ,g〉1 :=∫ b

a f (x) · g(x)dx

〈f ,g〉2 :=∫ b

a f (x) · g(x)2dx

〈f ,g〉3 :=∫ b

a f +(x) · g(x)dx

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 38

TU Munchen

Exercise Scalar Products and Vector Norms – Solution

Are the following operators scalar products in the vector spaceof continuous functions on the interval [a;b]?

〈f ,g〉1 :=∫ b

a f (x) · g(x)dx Yes!

〈f ,g〉2 :=∫ b

a f (x) · g(x)2dx No! (not linear in g)

〈f ,g〉3 :=∫ b

a f +(x) · g(x)dx No! (not positive definite)

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 39

TU Munchen

Exercise Scalar Products and Vector Norms

Proof that a set {~x1, ~x2, . . . , ~xN} of non-zero orthogonal vectorsin a vector space with scalar product (·, ·) always is a basis ofits span.

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 40

TU Munchen

Exercise Scalar Products and Vector Norms – Solution

Proof that a set {~x1, ~x2, . . . , ~xN} of non-zero orthogonal vectorsin a vector space with scalar product (·, ·) always is a basis ofits span.

Proof by contradiction:

Assume that the set is not linearly independent. Then, there is a element ~xi taht can bewritten as a linear combination ~xi =

∑k∈I αk~xk of other elements, where the index set

I ⊂ {1, 2, . . . ,N} does not contain i . With this, we get

0 6= (~xi , ~xi ) =(~xi ,∑

k∈I αk~xk)

=∑

k∈I αk (~xi , ~xk ) = 0.

Contradiction. Thus, the vector set is linearly independent and, thus, is a basis of itsspan. �

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 41

TU Munchen

Exercise Scalar Products and Vector Norms

Transform

{( 111

),

( 110

),

( 100

)}into an orthogonal basis

of R3.

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 42

TU Munchen

Exercise Scalar Products and Vector Norms – Solution

Transform

{( 111

),

( 110

),

( 100

)}into an orthogonal basis

of R3.

Gram-Schmidt orthogonalization:

~x1 =

111

, ~x2 =

110

−~x1,

110

(~x1,~x1)~x1 =

110

− 23~x1 =

1313− 2

3

,

~x3 =

100

−~x1,

100

(~x1,~x1)~x1 −

~x2,

100

(~x2,~x2)~x2 =

715− 8

151

15

.

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 43

TU Munchen

Matrix Norms

• definition:

– properties corresponding to those of vector norms– plus sub-multiplicativity:

‖AB‖ ≤ ‖A‖ · ‖B‖

– plus consistency‖Ax‖ ≤ ‖A‖ · ‖x‖

• matrix norms can be induced from corresponding vector norms: Euclidean,maximum, sum

‖A‖ := max‖x‖=1

‖Ax‖

• alternative: completely new definition, for example Frobenius norm (considermatrix as a vector, then take Euclidean norm)

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 44

TU Munchen

Classes of Matrices

• symmetric: A = AT

• skew-symmetric: A = −AT

• Hermitian: A = AH = AT

• s.p.d. (symmetric positive definite): xT Ax > 0 ∀x 6= 0

• orthogonal: A−1 = AT (the whole spectrum has modulus 1)

• unitary: A−1 = AH (the whole spectrum has modulus 1)

• normal: AAT = AT A or AAH = AHA, resp. (for those and only those matricesthere exists an orthonormal basis of eigenvectors)

Miriam Mehl: 1. Foundations of Numerics from Advanced Mathematics

Linear Algebra, October 23, 2012 45

Recommended