Jump to content

Mathematics for Applied Sciences (Osnabrück 2023-2024)/Part I/Lecture 24



Base change

We know, due to Theorem 23.15 , that in a finite-dimensional vector space, any two bases have the same length, the same number of vectors. Every vector has, with respect to every basis, unique coordinates (the coefficient tuple). How do these coordinates behave when we change the bases? This is answered by the following statement.


Lemma

Let K be a field, and let V be a K-vector space of dimension n. Let 𝔳=v1,…,vn and 𝔴=w1,…,wn denote bases of V. Suppose that

vj=∑i=1ncijwi

with coefficients cij∈K, which we collect into the n×n-matrix

M𝔴𝔳=(cij)ij.
Then a vector u, which has the coordinates (s1⋮sn) with respect to the basis 𝔳, has the coordinates
(t1⋮tn)=M𝔴𝔳(s1⋮sn)=(c11c12…c1nc21c22…c2n⋮⋮⋱⋮cn1cn2…cnn)(s1⋮sn)
with respect to the basis 𝔴.

Proof  

This follows directly from

u=∑j=1nsjvj=∑j=1nsj(∑i=1ncijwi)=∑i=1n(∑j=1nsjcij)wi,

and the definition of matrix multiplication.

◻


The matrix M𝔴𝔳, which describes the base change from 𝔳 to 𝔴, is called the transformation matrix. In the j-th column of the transformation matrix, there are the coordinates of vj with respect to the basis 𝔴. When we denote, for a vector u∈V and a basis 𝔳, the corresponding coordinate tuple by Ψ𝔳(u), then the transformation can be quickly written as

Ψ𝔴(u)=M𝔴𝔳(Ψ𝔳(u)).

We consider in ℝ2 the standard basis,

𝔲=(10),(01),

and the basis

𝔳=(12),(−23).

The basis vectors of 𝔳 can be expressed directly with the standard basis, namely

v1=(12)=1(10)+2(01) and v2=(−23)=−2(10)+3(01).

Therefore, we get immediately

M𝔲𝔳=(1−223).

For example, the vector that has the coordinates (4,−3) with respect to 𝔳, has the coordinates

M𝔲𝔳(4−3)=(1−223)(4−3)=(10−1)

with respect to the standard basis 𝔲. The transformation matrix M𝔳𝔲 is more difficult to compute. We have to write the standard vectors as linear combinations of v1 and v2. A direct computation (solving two linear systems) yields

(10)=37(12)−27(−23)

and

(01)=27(12)+17(−23).

Hence,

M𝔳𝔲=(3727−2717).



Linear mappings

Let K be a field, and let V and W be K-vector spaces. A mapping

φ:V⟶W

is called a linear mapping if the following two properties are fulfilled.

  1. φ(u+v)=φ(u)+φ(v) for all u,v∈V.
  2. φ(sv)=sφ(v) for all s∈K and v∈V.


Here, the first property is called additivity, and the second property is called compatibility with scaling. When we want to stress the base field, then we say K-linear. The identity IdV:V→V, the null mapping V→0, and the inclusion U⊆V of a linear subspace are the simplest examples of a linear mapping. For a linear mapping, the compatibility with an arbitrary linear combination holds, that is,

φ(∑i=1nsivi)=∑i=1nsiφ(vi);

see exercise *****.


Let K denote a field, and let Kn be the n-dimensional standard space. Then the i-th projection, this is the mapping

Kn⟶K,(x1,…,xi−1,xi,xi+1,…,xn)⟼xi,

is a K-linear mapping. This follows immediately from componentwise addition and scalar multiplication on the standard space. The i-th projection is also called the i-th coordinate function.


Lemma

Let K denote a field, and let U,V,W denote vector spaces over K. Suppose that

φ:U⟶V and ψ:V⟶W

are linear mappings. Then also the composition

ψ∘φ:U⟶W
is a linear mapping.

Proof

◻



Lemma

Let K be a field, and let V and W be K-vector spaces. Let

φ:V⟶W

be a bijective linear map. Then also the inverse mapping

φ−1:W⟶V
is linear.

Proof

◻



Determination on a basis

Behind the following statement (the determination theorem), there is the important principle that in linear algebra (of finite-dimensional vector spaces), the objects are determined by finitely many data.


Theorem

Let K be a field, and let V and W be K-vector spaces. Let vi, i∈I, denote a basis of V, and let wi, i∈I, denote elements in W. Then there exists a unique linear mapping

f:V⟶W

with

f(vi)=wi for all i∈I.

Proof

This proof was not presented in the lecture.
◻


The graph of a linear mapping from ℝ to ℝ, the mapping is determined by the proportionality factor k alone.

The easiest linear mappings are (beside the null mapping) the linear maps from K to K. Such a linear mapping

φ:K⟶K,x⟼φ(x),

is determined (by Theorem 24.7 , but this is also directly clear) by φ(1), or by the value φ(t) for a single element t∈K, t≠0. In particular, φ(x)=ax, with a uniquely determined a∈K. In the context of physics, for K=ℝ, and if there is a linear relation between two measurable quantities, we talk about proportionality, and a is called the proportionality factor. In school, such a linear relation occurs as "rule of three“.



Linear mappings and matrices
The effect of several linear mappings from ℝ2 to itself, represented on a brain cell.

Due to Theorem 24.7 , a linear mapping

φ:Kn⟶Km

is determined by the images φ(ej), j=1,…,n, of the standard vectors. Every φ(ej) is a linear combination

φ(ej)=∑i=1maijei,

and therefore the linear mapping is determined by the elements aij. So, such a linear map is determined by the mn elements aij, 1≤i≤m, 1≤j≤n, from the field. We can write such a data set as a matrix. Because of the determination theorem, this holds for linear maps in general, as soon as in both vector spaces bases are fixed.


Let K denote a field, and let V be an n-dimensional vector space with a basis 𝔳=v1,…,vn, and let W be an m-dimensional vector space with a basis 𝔴=w1,…,wm.

For a linear mapping

φ:V⟶W,

the matrix

M=M𝔴𝔳(φ)=(aij)ij,

where aij is the i-th coordinate of φ(vj) with respect to the basis 𝔴, is called the describing matrix for φ with respect to the bases.

For a matrix M=(aij)ij∈Matm×n(K), the linear mapping φ𝔴𝔳(M) determined by

vj⟼∑i=1maijwi

in the sense of Theorem 24.7 ,

is called the linear mapping determined by the matrix M.

For a linear mapping φ:Kn→Km, we always assume that everything is with respect to the standard bases, unless otherwise stated. For a linear mapping φ:V→V from a vector space in itself (what is called an endomorphism), one usually takes the same bases on both sides. The identity on a vector space of dimension n is described by the identity matrix, with respect to every basis.


Theorem

Let K be a field, and let V be an n-dimensional vector space with a basis 𝔳=v1,…,vn, and let W be an m-dimensional vector space with a basis 𝔴=w1,…,wm. Then the mappings

φ⟼M𝔴𝔳(φ) and M⟼φ𝔴𝔳(M),

defined in definition, are inverse

to each other.

Proof

This proof was not presented in the lecture.
◻



A linear mapping

φ:Kn⟶Km

is usually described by the matrix M with respect to the standard bases on the left and on the right. The result of the matrix multiplication

(y1⋮ym)=M(x1⋮xn)

can be interpreted directly as a point in Km. The j-th column of M is the image of the j-th standard vector ej.



Rotations

A rotation of the real plane ℝ2 around the origin, given the angle α counterclockwise, maps (10) to (cos⁡αsin⁡α) and (01) to (−sin⁡αcos⁡α). Therefore, plane rotations are described in the following way.


A linear mapping

D(α):ℝ2⟶ℝ2,

which is given by a rotation matrix (cos⁡α−sin⁡αsin⁡αcos⁡α) (with some α∈ℝ)with respect to the standard basis is called

rotation.

A space rotation is a linear mapping of the space ℝ3 in itself around a rotation axis (a line through the origin) with an certain angle α. If the vector v1≠0 defines the axis, and u2 and u3 are orthogonal to v1 and to each other, and all have length 1, then the rotation is described by the matrix

(1000cos⁡α−sin⁡α0sin⁡αcos⁡α)

with respect to the basis v1,u2,u3.



The kernel of a linear mapping

Let K denote a field, let V and W denote K-vector spaces, and let

φ:V⟶W

denote a K-linear mapping. Then

kern⁡φ:=φ−1(0)={v∈V∣φ(v)=0}
is called the kernel of φ.

The kernel is a linear subspace of V.

The following criterion for injectivity is important.


Lemma

Let K denote a field, let V and W denote K-vector spaces, and let

φ:V⟶W

denote a K-linear mapping. Then φ is injective if and only if kern⁡φ=0

holds.

Proof  

If the mapping is injective, then there can exist, apart from 0∈V, no other vector v∈V with φ(v)=0. Hence, φ−1(0)={0}.
So suppose that kern⁡φ=0, and let v1,v2∈V be given with φ(v1)=φ(v2). Then, due to linearity,

φ(v1−v2)=φ(v1)−φ(v2)=0.

Therefore, v1−v2∈kern⁡φ, and so v1=v2.

◻


<< | Mathematics for Applied Sciences (Osnabrück 2023-2024)/Part I | >>
PDF-version of this lecture
Exercise sheet for this lecture (PDF)