Linear Algebra cheat sheet

1 Vector spaces Definition. 1. We say that the vectors (b1 , . . . , bn ) are Note that A is the matrix whose columns are the coordinates of
In the following k = Q, R or C. linearly independent when b is injective, that is, when the f (b j ) in the basis (b10 , . . . bm
0 ).
only null linear combination x1 b1 + · · · + xn bn = 0 is the
Definition. A k-vector space is an abelian group V together with trivial one, x1 = · · · = xn = 0. If v ∈ V has coordinates X = ( x1 , . . . , xn ) T (column vector) in the
· basis (b1 , . . . bn ), then the coordinates X 0 of f (v) in the basis
a map k × V −→ V, (λ, v) 7→ λv, satisfying 2. We say that the vectors (b1 , . . . , bn ) span V when b is (b10 , . . . bm
0 ) are given by X 0 = AX.
1. λ(v1 + v2 ) = λv1 + λv2 , surjective, that is, when every vector v ∈ V can be written
2. (λ1 + λ2 )v = λ1 v + λ2 v,
as a linear combination of b1 , . . . , bn . Proposition 3.1 Let A be the matrix of a linear map f : V −→ V 0 in
3. We say that (b1 , . . . , bn ) form a basis for V when some basis. Then
3. (λµ)v = λ(µv), ∼ dim Im f = rank A
=
b : kn −→ V is a linear isomorphism. In such a case n is
4. 1 · v = v, and therefore
called the dimension of V and every v ∈ V can be written
for all λ, µ, λ1 , λ2 ∈ k and all v, v1 , v2 ∈ V. in a unique way as a linear combination of the base
A subgroup U ⊆ V is a linear subspace if λu ∈ U for all elements, v = x1 b1 + · · · + xn bn and we call dim Ker f = #(columns of A) − rank A
λ ∈ k, u ∈ U. b−1 (v) = ( x1 , . . . , xn ) the coordinates of v in such a basis.
Corollary 3.2 Let v1 , . . . , vr ∈ V. Let B be the matrix whose columns
Definition. A group homomorphism f : V −→ W between Lemma 2.1 Every sequence of linearly independent vectors v1 , . . . , vk are the coordinates of v1 , . . . , vr in some basis.
vector spaces is k-linear if can be extended to a basis for V. 1. These vectors are linearly independent if and only if rank B = r.
f (λv) = λ f (v) In particular, if rank B = r = dim V, they form a basis for V;
Lemma 2.2 Every spanning sequence of vectors v1 , . . . , vk contains a and if r > dim V, then they are linearly dependent.
for all λ ∈ k, v ∈ V. basis for V.
If f is bijective, it is a linear isomorphism, and in such a case so if 2. These vectors span V if and only if rank B = dim V.
f −1 . Lemma 2.3 If a linear map sends a basis to a basis, then it is an
The composite of linear maps is also linear. isomorphism. Corollary 3.3 Let A be the matrix of a linear map f : V −→ W
between vector spaces of the same finite dimension. Then f is an
Definition. Let f : V −→ W be a linear map. The kernel and the isomorphism if and only if det A 6= 0.
Proposition 2.4 Let U ⊆ V be a linear subspace of a finite dimensional
image of f are vector space V.
Ker f := {v ∈ V : f (v) = 0},
1. dim U ≤ dim V, and the equality holds only when U = V.
4 Dual vector space
Im f := { f (v) ∈ W : v ∈ V }, If V, W are vector spaces, the set of linear maps f : V −→ W is
2. dim(V/U ) = dim V − dim U. again a vector space denoted by Hom(V, W ), of dimension
and both are linear subspaces (of V and W, respectively).
dim Hom(V, W ) = (dim V )(dim W ).
Lemma 1.1 A linear map f : V −→ W is injective if and only if Corollary 2.5 Let f : V −→ W be a linear map. Then
Ker f = 0. Definition. Let V be a vector space. The dual vector space of V is
dim V = dim Ker f + dim Im f V ∗ := Hom(V, k).
Recall that if G is a group and H ⊆ G is a normal subgroup, the
factor group G/H (also known as the quotient group) is the set of Proposition 4.1 Let (b1 , . . . , bn ) be a basis for V. There exists a unique
Corollary 2.6 Let f : V −→ W be a linear map between vector spaces basis ( β1 , . . . , βn ) for V ∗ with the property that βi (b j ) = δji . This basis
subsets of the form g + H = { g + h : h ∈ H }, and this set is a of the same finite dimension. Then the following are equivalent:
group with group law ( a + H ) + (b + H ) := ( a + b) + H. Observe is called the dual basis of (b1 , . . . , bn ).
that here additive (and not multiplicative) notation for the group (a) f is isomorphism,
law has been used. If φ ∈ V ∗ , then its coordinates in the dual basis are
(b) f is injective,
Proposition 1.2 Let V be a vector space and U ⊆ V a linear subspace. (c) f is surjective. φ = φ(b1 ) β1 + . . . + φ(bn ) βn .
The quotient group V/U has a structure of k-vector space defined by
λ(v + U ) := (λv) + U, where λ ∈ k, v ∈ V. 3 Matrix of a linear map Definition. Let f : V −→ W be a linear map. The pull-back or
dual of f is the linear map
Theorem 1.3 Let f : V −→ W be a linear map. Then there is an Let f : V −→ V0
be a linear map and let (b1 , . . . bn ) and (b10 , . . . bm
0 )

isomorphism of vector spaces be basis for V and V 0 , so that f ∗ : W ∗ −→ V ∗ , f ∗ (φ) := φ ◦ f .

∼
=
f¯ : V/ Ker f −→ Im f , f¯(v + Ker f ) := f (v). f (b j ) = a1j b10 + · · · + amj bm
0
If (b1 , . . . , bn ) is a basis for V and ( β1 , . . . , βn ) is its dual basis,
given ω ∈ W ∗ the coordinates of f ∗ (ω ) are
for some unique scalars aij ∈ k, 1 ≤ i ≤ m, 1 ≤ j ≤ n.
2 Dimension theory
f ∗ (ω ) = ω ( f (b1 )) β1 + . . . + ω ( f (bn )) βn .
Any ordered sequence of vectors (b1 , . . . , bn ) in V defines a linear Definition. The matrix A = ( aij ) is called the matrix of f with
map respect to the basis (b1 , . . . bn ) and (b10 , . . . bm
0 ). That is, if
b : kn −→ V and b0 : km −→ V 0 are the basis of V and V 0 , then A
b : kn −→ V , b( x1 , . . . , xn ) = x1 b1 + · · · + xn bn . is the matrix of the linear map (b0 )−1 ◦ f ◦ b : kn −→ km . Jorge Becerra, [email protected]