Linear Algebra ReSupply Date VI. Matrix Represention.

前面我們討論了很多關於 vector space 與 linear transformation 的理論，接下來我們把前面抽

象的概念正式具象化。前面我們在講 basis 我們通常是把它視為一個 set，而不會去討論內部的
element(vector) 之間的關係，而在這裡我們會用到這些 vector 的順序 (也就是 order 的概念)。

Definition 1. Ordered Basis

Let V be a vector space over F with dim(V ) < ∞, then we defined the ordered basis β for V if β
has specific order.

Example. Ordered Basis

Let V = R3 , then we consider that

(i) β1 = {(1, 0, 0), (0, 1, 0), (0, 0, 1)}.

(ii) β2 = {(0, 1, 0), (1, 0, 0), (0, 0, 1)}.

We can see that β1 ̸= β2 as ordered basis and β1 = β2 as un-ordered basis.

用任意不同的 ordered basis 寫出來的 vector 或 matrix 是長得不一樣的，但是他們之間的性質

是不變的，接著我們去討論 ordered basis 所寫出來的 matrix representation。

Definition 2. Column Vector

Let β = {u1 , ..., un } be an ordered basis for V , if for all x ∈ V , x = a1 u1 + · · · + an un for some unique
ai ∈ F , then we defined by  
a1
 . 
[x]β =  . 
 . 
an
is called the column vector.

Notation. Matrix Representation

Let T ∈ L(V, W ) and dim(V ), dim(W ) < ∞, and let β = {v1 , ..., vn }, γ = {w1 , ..., wm } be the basis
for V, W , then the linear transformation is uniquely determined by {T (v1 ), ..., T (vn )}.
each T (vi ) has a unique expression in terms of {w1 , ..., wm } i.e.

∑
m
T (vj ) = aij wi
i=1

for some aij ∈ F .

1 | Linear Algebra ReSupply Date VI JunWei,Zhang Edit

Definition 3. Matrix Representation
Using above notation, we calll the matrix A, Aij = aij is called ij-componemt of A, then matrix
representation of T where β, γ fixed. denoted by A = [T ]γβ and A is a m × n matrix, denoted by
A ∈ Mm×n (F ). If T ∈ L(V, V ), then we denoted [T ]β .

Remark. Matrix Representation

Let T, U ∈ L(V, W ), if [T ]γβ = [U ]γβ , then T = U .

Example. Matrix Representation

Let T ∈ L(R2 , R3 ) and β = {(1, 0), (0, 1)} and γ = {(1, 0, 0), (0, 1, 0), (0, 0, 1)}.
If T (a1 , a2 ) = (a1 + 3a2 , a1 , 2a1 − 4a2 ), then we find [T ]γβ .

Solution.
Since T (1, 0) = (1, 1, 2) = 1(1, 0, 0) + 1(0, 1, 0) + 2(0, 0, 1) and T (0, 1) = (3, 0, −4) = 3(1, 0, 0) +
0(0, 1, 0) − 4(0, 0, 1), then we have  
1 3
[T ]γβ =  1 0 
2 −4

我們在中學有學過 matrix 是 linear，在這裡我們的所有 properties 都跟小時候一樣。

Definition 4. Linear transformation in Addition vs Scalar-Mulitiplication
Let T, U ∈ L(V, W ) ba arbitary function and V, W be vector space over F , then we defined the
following:
(a) (T + U )(x) = T (x) + U (x) for all x ∈ V .

(b) (cT )(x) = cT (x) for all x ∈ V, c ∈ F .

Theorem 1.
Let T, U ∈ L(V, W ) ba arbitrary function and V, W be vector space over F , then we have
(a) aT + U ∈ L(V, W ) for all a ∈ F .

(b) S = {T : V → W : T ∈ L(V, W )} is a vector space.

Proof.
We show that the following:

(a) for all x, y ∈ V , a, c ∈ F , then we have

(aT + U )(cx + y) = (aT + U )(cx) + (aT + U )(y) = c(aT (x) + U (x)) + (aT (y) + U (y))
= c(aT + U )(x) + (aT + U )(y)

Thus, we know that aT + U ∈ L(V, W ) for all a ∈ F .

(b) we choose zero function is the identity element of S. the proof is trivial.

2 | Linear Algebra ReSupply Date VI JunWei,Zhang Edit

Theorem 2.
Let T, U ∈ L(V, W ) and V, W be vector space over F with dim(V ), dim(W ) < ∞, then we have

(a) [T + U ]γβ = [T ]γβ + [U ]γβ .

(b) [aT ]γβ = a[T ]γβ

Proof.
Let β = {v1 , ..., vn } and γ = {w1 , ..., wm }, then we have

∑
m ∑
m
T (vj ) = aij wi ; U (vj ) = bij wi
i=1 i=1
∑m
(a) Since (T + U )(vj ) = T (vj ) + U (vj ) = i=1 (aij + bij )wi , then we have

([T + U ]γβ )ij = aij + bij = ([T ]γβ )ij + ([U ]γβ )ij

Thus, we get [T + U ]γβ = [T ]γβ + [U ]γβ .

∑m
(b) Since (aT )(vj ) = aT (vj ) = a i=1 (aij + bij )wi , then we have

([aT ]γβ )ij = a · aij = a([T ]γβ )ij

Thus, we get [aT ]γβ = a[T ]γβ + [U ]γβ .

Theorem 3.
Let T ∈ L(V, W ), U ∈ L(W, Z) and V, W, Z be vector space over F with dim(V ), dim(W ), dim(Z) <
∞, then U T ∈ L(V, Z).

Proof.
By definition, we have

U T (ax + y) = U (aT (x) + T (y)) = aU (T (x)) + U (T (y)) = aU T (x) + U T (y)

for all x, y ∈ V , a ∈ F .
Thus, we get U T ∈ L(V, Z).

3 | Linear Algebra ReSupply Date VI JunWei,Zhang Edit

Theorem 4.
Let T, U1 , U2 ∈ L(V, V ) and V be vector space over F , then we have

(a) T (U1 + U2 ) = T U1 + T U2

(b) T (U1 U2 ) = (T U1 )U2

(c) T IdV = T = IdV T

(d) a(U1 + U2 ) = aU1 + aU2 for all a ∈ F .

Proof.
We show that the following:

(a) (i) Since T (U1 + U2 )(x) = T ((U1 + U2 )(x)) = T (U1 (x) + U2 (x)) = T (U1 (x)) + T (U2 (x)),
then we have T (U1 + U2 )(x) = T U1 (x) + T U2 (x).
Thus, we get T (U1 + U2 ) = T U1 + T U2 .

(ii) Since (U1 + U2 )T (x) = (U1 + U2 )(T (x)) = U1 (T (x)) + U2 (T (x)), then we have

T (U1 + U2 )(x) = U1 T (x) + U2 T (x)

Thus, we get (U1 + U2 )T = U1 T + U2 T .

(b) Since T (U1 U2 )(x) = T (U1 U2 (x)) = T (U1 (U2 (x))) = T U1 (U2 (x)), then we have

T (U1 U2 ) = (T U1 )U2

(c) Since T IdV (x) = T (IdV (x)) = T (x) = IdV (T (x)) = IdV T (x), then we get

T IdV = T = IdV T

(d) Since a(U1 U2 )(x) = a(U1 (U2 (x))) = aU1 (U2 (x)) = U1 (aU2 (x)), then we have

a(U1 U2 ) = (aU1 )U2 = U1 (aU2 )

上面我們討論完了 matrix 的 addition 和 scalar-mulitiplication，接著我們去討論 matrix 間的

mulitiplication，這裡我們用的 definition 很以前一樣，但是我們在這裡用比較理論的 definition。

Definition 5. Matrix Mulitplication

Let A ∈ Mm×n (F ), B ∈ Mn×p (F ), then we defined the Matrix Mulitplication by

∑
n
(AB)ij = Aik Bkj
k=1

for all i ∈ [1, m], j ∈ [1, p].

4 | Linear Algebra ReSupply Date VI JunWei,Zhang Edit

 基本上，matrix 和 matrix 之間的關係與 linear transformation 和 linear transformation 之間的
關係是相似的，只是要注意的是 matrix representation 要去注意所在的 vector space 與其對應的
basis。然後這邊需要各位去記得當你在證明 matrix 之間的運算的時候要從 component(entry) 去
看，這樣可以把問題變得很簡單。

Theorem 5.
Let T ∈ L(V, W ), U ∈ L(W, Z) and V, W, Z be vector space over F with dim(V ), dim(W ), dim(Z) <
∞, and let α = {v1 , ..., vn }, β = {w1 , ..., wm }, γ = {z1 , ..., zp } be the basis for V, W, Z.
If B = [T ]βα , A = [U ]γβ , then AB = [U T ]γα .

Proof.
By definition, we have

∑
m ∑
m ∑
m ∑
p
(U T )(vj ) = U (T (vj )) = U ( Bkj wk ) = Bkj U (wk ) = Bkj Aik zi
k=1 k=1 k=1 i=1
∑
p
∑
m ∑
p
= ( Aik Bkj )zi = cij zi
i=1 k=1 i=1
∑m
Thus, we get ([U T ]γα )ij = cij = k=1 Aik Bkj = (AB)ij .

Theorem 6.
Let T ∈ L(V, W ), U ∈ L(W, Z) and V, W, Z be vector space over F , and let α = {v1 , ..., vn },
β = {w1 , ..., wm }, γ = {z1 , ..., zp } be the basis for V, W, Z, then

[U T ]γα = [U ]γβ [T ]βα

(the proof by above remark).

Remark. Commutative in Matrix Multiplication

The matrix multiplication is not commutative.

Example. Commutative in Matrix Multiplication

By definition, we have
( )( ) ( )
1 1 0 1 1 1
=
0 0 1 0 0 0
( )( ) ( )
0 1 1 1 0 0
=
1 0 0 0 1 1

Remark. Cancellation Law in Matris Multiplication

The matrix multiplication has no cancellation law.

5 | Linear Algebra ReSupply Date VI JunWei,Zhang Edit

Example. Cancellation
 Law in Matris Multiplication
0 0 0
 
Let A =  1 0 0 

, then we have
0 1 0
    
0 0 0 0 0 0 0 0 0
    
A2 = 
 1 0 0   1 0 0  =  0 0 0  , A 3 = O3
   
0 1 0 0 1 0 1 0 0

Definition 6. Transpose
We defined the transpose of A ∈ Mm×n (F ) by (AT )ij = Aji , denoted by AT .

Theorem 7. the Property of Transpose

Let A ∈ Mm×n (F ), B ∈ Mn×p (F ), then (AB)T = B T AT .

Proof.
By definition, we have
∑
n ∑
n
(AB)Tij = (AB)ji = Ajk Bki = (B T )ik (AT )kj = ((B T )(AT ))ij
k=1 k=1

Thus, we get (AB)T = B T AT .

最後在證明有關 transpose 的性質時也盡量從 component(entry) 去看，不要在那邊把整個矩陣

畫出來 (可以輔助，但是他不是最佳的證明方式)。

Notation. Reference Material

(1) Edition. NTNU Math113 Jun-Wei,Zhang - Date. 2022-01-10.

(2) Linear Algebra (4th Edition 2003) - Stephen H.Friedberg, Arnold J. Insel, Lawrence E. Spence.

(3) Linear Algebra (2nd Edition 1971) - Kenneth M Hoffman, Ray Kunze.

6 | Linear Algebra ReSupply Date VI JunWei,Zhang Edit