Sie sind auf Seite 1von 6

Topic 4.

0 - Core: Matrices
4.1. Basics of Matrices
Matrix Definitions
𝑐𝑜𝑙𝑢𝑚𝑛 (𝑛)
2 −9 8
𝑨 = 𝑟𝑜𝑤𝑠 (𝑚) [ 0 4 −6]
−4 2 7

𝑨 𝑖𝑠 𝑎 𝑀𝑎𝑡𝑟𝑖𝑥 𝑜𝑓 𝑜𝑟𝑑𝑒𝑟 3 × 3

𝐴𝑛𝑦 𝑠𝑖𝑛𝑔𝑙𝑒 𝑣𝑎𝑙𝑢𝑒 𝑖𝑛 𝑡ℎ𝑒 𝑚𝑎𝑡𝑟𝑖𝑥 𝑖𝑠 𝑎𝑛 𝐸𝑙𝑒𝑚𝑒𝑛𝑡.

Note
𝑇ℎ𝑒 𝑜𝑟𝑑𝑒𝑟 𝑖𝑠 𝑤𝑟𝑖𝑡𝑡𝑒𝑛 𝑚 × 𝑛, 𝑡ℎ𝑎𝑡 𝑖𝑠 𝑟𝑜𝑤𝑠 𝑡ℎ𝑒𝑛 𝑐𝑜𝑙𝑢𝑚𝑛𝑠

𝑚(𝑟𝑜𝑤𝑠) × 𝑛(𝑐𝑜𝑙𝑢𝑚𝑛𝑠)

Use to store Data


Matrices can be used to store data such as tables.

Furniture
Chairs Tables Beds 𝟔 𝟏 𝟐
Flat 6 1 2 𝒄𝒂𝒏 𝒃𝒆 𝒆𝒙𝒑𝒓𝒆𝒔𝒔𝒆𝒅 𝒂𝒔 [ 𝟗 𝟐 𝟑]
Unit 9 2 3 𝟏𝟎 𝟑 𝟒
House 10 3 4

4.2. Core Matrix Knowledge


Core Operations
The following require two matrices of the same order, such that...

𝑚1 = 𝑚2 , 𝑛1 = 𝑛2

Equality

𝑎 𝑏 𝑤 𝑥
𝑖𝑓 [ ] = [𝑦 𝑧 ] ⇒ 𝑎 = 𝑤, 𝑏 = 𝑥, 𝑐 = 𝑦, 𝑑 = 𝑧
𝑐 𝑑
Addition

𝑎 𝑏 𝑤 𝑥 𝑎+𝑤 𝑏+𝑥
[ ] + [𝑦 𝑧] = [ 𝑐 + 𝑦 ]
𝑐 𝑑 𝑑+𝑧

Subtraction

𝑎 𝑏 𝑤 𝑥 𝑎−𝑤 𝑏−𝑥
[ ] − [𝑦 𝑧] = [ 𝑐 − 𝑦 ]
𝑐 𝑑 𝑑−𝑧
Multiplications
Scalar Multiplication

𝑎 𝑏 𝑘𝑎 𝑘𝑏
𝑘[ ]=[ ]
𝑐 𝑑 𝑘𝑐 𝑘𝑑
Matrix Multiplication

 𝑨𝑩 will only exist if 𝑨𝑛 = 𝑩𝑚


 𝑨𝑩's result will be of the order 𝑨𝑚 × 𝑩𝑛
 𝑨𝑩 ≠ 𝑩𝑨
 The terms pre-multiply and post-multiply can be used to denote multiplying by the multiplication or the
multiplication by the existing value.

Example:
𝑨 𝑖𝑠 𝑎 𝑚𝑎𝑡𝑟𝑖𝑥 𝑜𝑓 𝑜𝑟𝑑𝑒𝑟 4 × 3, 𝑚𝑒𝑎𝑛𝑖𝑛𝑔 𝑖𝑡 ℎ𝑎𝑠 3 𝑟𝑜𝑤𝑠 𝑎𝑛𝑑 4 𝑐𝑜𝑙𝑢𝑚𝑛𝑠

𝑩 𝑖𝑠 𝑎 𝑚𝑎𝑡𝑟𝑖𝑥 𝑜𝑓 𝑜𝑟𝑑𝑒𝑟 3 × 2, 𝑚𝑒𝑎𝑛𝑖𝑛𝑔 𝑖𝑡 ℎ𝑎𝑠 2 𝑟𝑜𝑤𝑠 𝑎𝑛𝑑 3 𝑐𝑜𝑙𝑢𝑚𝑛𝑠

𝑆𝑜 𝑎𝑠 3 = 3 𝑨𝑩 𝑤𝑖𝑙𝑙 𝑏𝑒 𝑎 𝑚𝑎𝑡𝑟𝑖𝑥 𝑜𝑓 𝑜𝑟𝑑𝑒𝑟 4 × 2, 𝑚𝑒𝑎𝑛𝑖𝑛𝑔 𝑖𝑡 ℎ𝑎𝑠 4 𝑟𝑜𝑤𝑠 𝑎𝑛𝑑 2 𝑐𝑜𝑙𝑢𝑚𝑛𝑠

𝐻𝑜𝑤𝑒𝑣𝑒𝑟, 𝑩𝑨 𝑑𝑜𝑒𝑠 𝑛𝑜𝑡 𝑒𝑥𝑖𝑠𝑡 𝑎𝑠 2 ≠ 4

Result:
In matrix multiplication, each value in the result is calculated by the sum all the multiplication of the values in the
same row as the element in the first matrix, by the values in the same column as the element in the second matrix.
Such that...

𝑎 𝑏 𝑤 𝑥 𝑎𝑤 + 𝑏𝑦 𝑎𝑥 + 𝑏𝑧
[ ] × [𝑦 𝑧 ] = [𝑐𝑤 + 𝑑𝑦 ]
𝑐 𝑑 𝑐𝑥 + 𝑑𝑧

So therefore, for each element, a general formula can be applied.

𝑨1,1 ⋯ 𝑨1,𝑛 𝑩1,1 ⋯ 𝑩1,𝑛


𝑓𝑜𝑟 [ ⋮ ⋱ ⋮ ]×[ ⋮ ⋱ ⋮ ]
𝑨𝑚,1 ⋯ 𝑨𝑚,𝑛 𝑩𝑚,1 ⋯ 𝑩𝑚,𝑛

𝐼𝑛 𝑡ℎ𝑒 𝑟𝑒𝑠𝑢𝑙𝑡 𝐸𝑙𝑒𝑚𝑒𝑛𝑡𝑦,𝑥 = 𝑨𝑦,1 𝑩1,𝑥 + ⋯ + 𝑨𝑦,𝑛 𝑩𝑚,𝑥

Result in full:
𝑨1,1 ⋯ 𝑨1,𝑛 𝑩1,1 ⋯ 𝑩1,𝑛 𝑨1,1 𝑩1,1 + ⋯ + 𝑨1,𝑛 𝑩𝑚,1 ⋯ 𝑨1,1 𝑩1,𝑛 + ⋯ + 𝑨1,𝑛 𝑩𝑚,𝑛
[ ⋮ ⋱ ⋮ ]×[ ⋮ ⋱ ⋮ ]=[ ⋮ ⋱ ⋮ ]
𝑨𝑚,1 ⋯ 𝑨𝑚,𝑛 𝑩𝑚,1 ⋯ 𝑩𝑚,𝑛 𝑨𝑚,1 𝑩1,1 + ⋯ + 𝑨𝑚,𝑛 𝑩𝑚,1 ⋯ 𝑨𝑚,1 𝑩1,𝑛 + ⋯ + 𝑨𝑚,𝑛 𝑩𝑚,𝑛

Identity and Zero Matrices


1 ⋯ 0 0 ⋯ 0
𝐼𝑑𝑒𝑛𝑡𝑖𝑡𝑦 (𝑰) = [ ⋮ ⋱ ⋮ ] 𝑍𝑒𝑟𝑜 (𝑶) = [ ⋮ ⋱ ⋮ ]
0 ⋯ 1 0 ⋯ 0
The identity matrix can only exist as a square matrix, The zero matrix can be of any order, and must contain
such that in the case of 𝑰𝑛 , 𝑰 is of order 𝑛 × 𝑛 only zeros.

Properties
𝑨𝑰 = 𝑰𝑨 = 𝑨 Properties
𝑰𝟐 = 𝑰 𝑨𝑶 = 𝑶𝑨 = 𝑶
𝑶𝟐 = 𝑶
It should also be noted that Information can be put into a matrix and then processed using matrix operations.
4.3. Determinant and Inverse of a square matrix
Determinant of a Square Matrix
The determinant of a square matrix, determines whether or not the matrix has an inverse.

Notation

det 𝑨 𝑜𝑟 |𝑨|

For a 𝟐 × 𝟐 Matrix
𝑎 𝑏
| | = 𝑎𝑑 − 𝑏𝑐
𝑐 𝑑
For a 𝟑 × 𝟑 Matrix
𝑎 𝑏 𝑐
| 𝑑 𝑒 𝑓 | = 𝑎𝑒𝑖 + 𝑏𝑓𝑔 + 𝑐𝑑ℎ − 𝑐𝑒𝑔 − 𝑏𝑑𝑖 − 𝑎𝑓𝑔
𝑔 ℎ 𝑖

This can also be remembered using the following method: Through adding the values multiplied of the downward
arrows, and subtracting the upward arrows.

𝑎 𝑏 𝑐 𝑎 𝑏
(𝑑 𝑒 𝑓 ) 𝑑 𝑒 = 𝑎𝑒𝑖 + 𝑏𝑓𝑔 + 𝑐𝑑ℎ − 𝑐𝑒𝑔 − 𝑏𝑑𝑖 − 𝑎𝑓𝑔
𝑔 ℎ 𝑖 𝑔 ℎ

Properties
It should be known that det 𝐴𝐵 = det 𝐴 + det 𝐵, this can be proved relatively simply for two 2 × 2 matrices.

𝑎 𝑏 𝑤 𝑥 𝑎𝑤 + 𝑏𝑦 𝑎𝑥 + 𝑏𝑧
|
𝑐
||
𝑑 𝑦 𝑧 | = | 𝑐𝑤 + 𝑑𝑦 𝑐𝑥 + 𝑑𝑧
|

(𝑎𝑑 − 𝑏𝑐)(𝑤𝑧 − 𝑥𝑦) = 𝑎𝑑𝑤𝑧 + 𝑏𝑐𝑥𝑦 − 𝑎𝑑𝑥𝑦 − 𝑏𝑐𝑤𝑧

((𝑎𝑤 + 𝑏𝑦)(𝑐𝑥 + 𝑑𝑧) − (𝑎𝑥 + 𝑏𝑧)(𝑐𝑤 + 𝑑𝑦)) = 𝑎𝑤𝑐𝑥 + 𝑏𝑦𝑑𝑧 + 𝑎𝑤𝑑𝑧 + 𝑏𝑦𝑐𝑥 − 𝑎𝑥𝑐𝑤 − 𝑏𝑧𝑑𝑦 − 𝑎𝑥𝑑𝑦 − 𝑏𝑧𝑐𝑤

𝑎𝑤𝑐𝑥 + 𝑏𝑦𝑑𝑧 + 𝑎𝑤𝑑𝑧 + 𝑏𝑦𝑐𝑥 − 𝑎𝑥𝑐𝑤 − 𝑏𝑧𝑑𝑦 − 𝑎𝑥𝑑𝑦 − 𝑏𝑧𝑐𝑤 = 𝑎𝑤𝑑𝑧 + 𝑏𝑦𝑐𝑥 − 𝑎𝑥𝑑𝑦 − 𝑏𝑧𝑐𝑤

𝑎𝑑𝑤𝑧 + 𝑏𝑐𝑥𝑦 − 𝑎𝑑𝑥𝑦 − 𝑏𝑐𝑤𝑧 = 𝑎𝑤𝑑𝑧 + 𝑏𝑦𝑐𝑥 − 𝑎𝑥𝑑𝑦 − 𝑏𝑧𝑐𝑤

Singular and Non-Singular Matrices


The determinant defines whether or not a square matrix is called singular or non-singular.

Singular Non-Singular
 det 𝑨 = 0  det 𝑨 ≠ 0
 No inverse matrix  Has an inverse matrix
Inverse of a Matrix
𝑓𝑜𝑟 𝑚𝑎𝑡𝑟𝑖𝑥 𝑨: 𝑇ℎ𝑒 𝑖𝑛𝑣𝑒𝑟𝑠𝑒 𝑖𝑠 𝑨−1

𝑨𝑨−1 = 𝑨−1 𝑨 = 𝑰

Calculation of the inverse


𝐹𝑜𝑟 𝑎 2 × 2 𝑀𝑎𝑡𝑟𝑖𝑥: 𝑨

1 𝑑 −𝑏
𝑨−1 = ×[ ]
det 𝑨 −𝑐 𝑎
For that and higher orders the GDC can be used to calculate inverse matrices. These can be found using...

𝑚𝑎𝑡 𝐴−1 𝑜𝑟 𝑚𝑎𝑡 𝐴^ − 1


1
It should also be noted that all inverse matrices take this form of det 𝑨 multiplied by a matrix. This is why when
1
det 𝑨 = 0, there is no inverse as det 𝑨 is undefined.

4.4. Solution of Systems of Linear Equations (Simultaneous Equations)


Solving Simultaneous Equations using Matrix Algebra
Simultaneous equations can be solved when there is a unique solution using matrix algebra. In the following case of
three simultaneous equations:

𝑎𝑥 + 𝑏𝑦 + 𝑐𝑧 = 𝑢
𝑑𝑥 + 𝑒𝑦 + 𝑓𝑧 = 𝑣
𝑔𝑥 + ℎ𝑦 + 𝑖𝑧 = 𝑤

We can therefore express these in a matrix form.

𝑥 𝑎 𝑏 𝑐 𝑢 𝑥 𝑎 𝑏 𝑐 𝑎𝑥 + 𝑏𝑦 + 𝑐𝑧
(𝑦) (𝑑 𝑒 𝑓 ) = ( 𝑣 ) 𝑎𝑠 (𝑦) (𝑑 𝑒 𝑓 ) = (𝑑𝑥 + 𝑒𝑦 + 𝑓𝑧)
𝑧 𝑔 ℎ 𝑖 𝑤 𝑧 𝑔 ℎ 𝑖 𝑔𝑥 + ℎ𝑦 + 𝑖𝑧

Using matrix algebra we can post-multiply the answer matrix by the inverse of the coefficient matrix, like so.

𝑥 𝑎 𝑏 𝑐 𝑎 𝑏 𝑐 −1 𝑢 𝑎 𝑏 𝑐 −1
(𝑦) (𝑑 𝑒 𝑓 ) (𝑑 𝑒 𝑓 ) = ( 𝑣 ) (𝑑 𝑒 𝑓)
𝑧 𝑔 ℎ 𝑖 𝑔 ℎ 𝑖 𝑤 𝑔 ℎ 𝑖

𝑥 𝑢 𝑎 𝑏 𝑐 −1
( ) = ( 𝑣 ) (𝑑
𝑦 𝑒 𝑓)
𝑧 𝑤 𝑔 ℎ 𝑖
Solving Simultaneous Equations using Row Reduction
Augmented Matrices
Matrices can be augmented, by simply putting two together. Going back to our simultaneous equations, we can
augment the answer and coefficient matrices together. This then allows us to perform elementary row operations
on the equations.

𝑎𝑥 + 𝑏𝑦 + 𝑐𝑧 = 𝑢
𝑑𝑥 + 𝑒𝑦 + 𝑓𝑧 = 𝑣
𝑔𝑥 + ℎ𝑦 + 𝑖𝑧 = 𝑤

𝑥 𝑎 𝑏 𝑐 𝑢 𝑎 𝑏 𝑐 𝑢
(𝑦) (𝑑 𝑒 𝑓 ) = ( 𝑣 ) 𝑠𝑜 𝑡ℎ𝑒 𝑠𝑦𝑠𝑡𝑒𝑚 𝑐𝑎𝑛 𝑏𝑒 𝑟𝑒𝑝𝑟𝑒𝑠𝑒𝑛𝑡𝑒𝑑 𝑏𝑦: (𝑑 𝑒 𝑓 | 𝑣)
𝑧 𝑔 ℎ 𝑖 𝑤 𝑔 ℎ 𝑖 𝑤

This augmented matrix represents the above system of linear equations.

Elementary Row Operations


Row Switching

Two rows can be interchanged in their position in the matrix.

𝑎 𝑏 𝑐 𝑢 𝑑 𝑒 𝑓 𝑣
(𝑑 𝑒 𝑓 | 𝑣 ) ⟹ (𝑎 𝑏 𝑐 | 𝑢)
𝑔 ℎ 𝑖 𝑤 𝑔 ℎ 𝑖 𝑤

Row Multiplication

A row can be multiplied by a non-zero scalar.

𝑎 𝑏 𝑐 𝑢 𝑎 𝑏 𝑐 𝑢
(𝑑 𝑒 𝑓 | 𝑣) ⟹ ( 𝑑 𝑒 𝑓 | 𝑣 ) 𝑘≠0
𝑔 ℎ 𝑖 𝑤 𝑔𝑘 ℎ𝑘 𝑖𝑘 𝑤𝑘

Row Addition

A row can be added to another as so.

𝑎 𝑏 𝑐 𝑢 𝑎+𝑔 𝑏+ℎ 𝑐+𝑖 𝑢+𝑤


(𝑑 𝑒 𝑓 | 𝑣) ⟹ ( 𝑑 𝑒 𝑓 | 𝑣 )
𝑔 ℎ 𝑖 𝑤 𝑔 ℎ 𝑖 𝑤

Combinations

Any number of elementary Row operations can be performed, and simple things such as subtraction can be derived
from multiplying by scalar −1 and then adding.

Row addition with scalar multiplication can also be performed like so.
𝑎 𝑏 𝑐 𝑢 𝑎 + 𝑘𝑔 𝑏 + 𝑘ℎ 𝑐 + 𝑘𝑖 𝑢 + 𝑘𝑤
( 𝑑 𝑒 𝑓 𝑣
| )⟹( 𝑑 𝑒 𝑓 | 𝑣 )
𝑔 ℎ 𝑖 𝑤 𝑔 ℎ 𝑖 𝑤
Row Echelon Form
A matrix is in row-echelon form when the following conditions are met:
 If there is a row of all zeros, then it is at the bottom of the matrix.
 The first non-zero element of any row is a one. That element is called the leading one.
 The leading one of any row is to the right of the leading one of the previous row.

Elementary Row Operations are used on any augmented matrix to put it into Row Echelon Form.

Reduced Row Echelon Form


A matrix is in reduced row-echelon form when all of the conditions of row-echelon form are met and all elements
above, as well as below, the leading ones are zero.

Case 1: Unique Solution


In our augmented matrix of these equations after bringing it into row echelon form, we return...

1 𝑜 𝑝 𝑟
(0 1 𝑞 | 𝑠) 𝑤ℎ𝑖𝑐ℎ 𝑟𝑒𝑝𝑟𝑒𝑠𝑒𝑛𝑡𝑠 …
0 0 1 𝑡
𝑥 + 𝑜𝑦 + 𝑝𝑧 = 𝑟 𝑦 + 𝑞𝑧 = 𝑠 𝑧=𝑡

𝑤ℎ𝑒𝑟𝑒 𝑜 − 𝑡 𝑎𝑟𝑒 𝑘𝑛𝑜𝑤𝑛 𝑛𝑢𝑚𝑏𝑒𝑟𝑠 𝑑𝑒𝑟𝑖𝑣𝑒𝑑 𝑡ℎ𝑟𝑜𝑢𝑔ℎ 𝑟𝑜𝑤 𝑜𝑝𝑒𝑟𝑎𝑡𝑖𝑜𝑛𝑠 𝑓𝑟𝑜𝑚 𝑡ℎ𝑒 𝑜𝑟𝑖𝑔𝑖𝑛𝑎𝑙 𝑎𝑢𝑔𝑚𝑒𝑛𝑡𝑒𝑑 𝑚𝑎𝑡𝑟𝑖𝑥

This makes our equations solvable as we can deduce that...

𝑦 + 𝑞𝑡 = 𝑠 ∴ 𝑦 = 𝑠 − 𝑞𝑡
𝑥 + 𝑜(𝑠 − 𝑞𝑡) + 𝑝𝑡 = 𝑟 ∴ 𝑥 = 𝑟 − 𝑝𝑡 − 𝑜(𝑠 − 𝑞𝑡)

Case 2: Unique Solution with Reduced Row Echelon Format


In the above case we can always bring it into the reduced form, this can be done from Case 1 by performing…

𝑅2 − 𝑞𝑅3 , 𝑅1 − 𝑝𝑅3 , 𝑅1 − 𝑜𝑅2 𝑤ℎ𝑖𝑐ℎ 𝑔𝑖𝑣𝑒𝑠 …

1 0 0 𝑟 − 𝑝𝑡 − 𝑜(𝑠 − 𝑞𝑡)
(0 1 0| 𝑠 − 𝑞𝑡 ) 𝑤ℎ𝑖𝑐ℎ 𝑟𝑒𝑝𝑟𝑒𝑠𝑒𝑛𝑡𝑠 …
0 0 1 𝑡
𝑥 = 𝑟 − 𝑝𝑡 − 𝑜(𝑠 − 𝑞𝑡) 𝑦 = 𝑠 − 𝑞𝑡 𝑧=𝑡

In this case we have an augmented matrix in reduced row echelon format, making our equations extremely simple to
solve. The values can then be read off. The decision whether or not to reduce the system of equations in the matrix
or not is completely up to the individual, they both come to the same conclusion

Case 3: Infinite number of Solutions


Take for example the augmented matrix representing a system of linear equations of, attempting to put it into row
echelon format creates...

1 𝑜 𝑝 𝑟
(0 1 𝑞 | 𝑠 )
0 0 0 0
Row 3 now represents the equation 0 = 0 which is always true, therefore supplies no new information. This means
as there are 3 unknowns that there are an infinite number of solutions, the equations for these infinite number of
solutions can therefore be expressed in terms of 𝑧.

Case 4: No solution
1 𝑜 𝑝 𝑟
(0 1 𝑞 | 𝑠) , 𝑡 ≠ 0
0 0 0 𝑡
As this case is inherently false (As 𝑅3 represents 0𝑥 + 0𝑦 + 0𝑧 = 𝑡), the system has no solutions.

Das könnte Ihnen auch gefallen