Matrix Algebra Cheat Sheet Edexcel FP2

Eigenvectors and Eigenvalues Most square matrices can be transformed into a diagonal matrix. This process is called 3 2
Example: Diagonalise 𝐀𝐀 = � � by finding a matrix 𝐏𝐏 and diagonal matrix 𝐃𝐃 such that 𝐃𝐃 =
diagonalisation. The diagonal entries of this matrix are the eigenvalues of the original matrix. 2 3
If a matrix 𝐀𝐀 transforms a non-zero vector v such that the output is a scalar multiple of v, 𝐏𝐏 −1 𝐀𝐀𝐀𝐀.
1 2 0
𝐀𝐀𝐀𝐀 = 𝜆𝜆𝐯𝐯, Example: Diagonalise 𝐀𝐀 = �0 3 0� into a diagonal matrix 𝐃𝐃, such that 𝐃𝐃 = 𝐏𝐏 −1 𝐀𝐀𝐀𝐀, where Find and solve the 𝟑𝟑 − λ 2
det(𝐀𝐀 − λ𝐈𝐈) = � �
2 −4 2 characteristic 2 3−λ
2 − 22 = −λ2 − 6λ + 5 = 0 = (λ − 1)(5 − λ) = 0 ⟹ λ = 1, 5
then v is called an eigenvector of matrix 𝐀𝐀, with 𝜆𝜆 its corresponding eigenvalue. Eigenvectors are 𝐏𝐏 is to be found. equation to obtain = (3 − λ)
special to a matrix transformation, as their direction is unchanged during a linear transformation the eigenvalues.
– they get stretched or squashed. The equation above is called the eigenvalue equation. Invariant Find and solve the 1−λ 2 0 Since 𝐀𝐀 is When λ = 1,
characteristic det(𝐀𝐀 − λ𝐈𝐈) = � 0 3−λ 0 � symmetric, use 3 2 𝑥𝑥 𝑥𝑥
lines of a transformation are parallel to eigenvectors. To find the eigenvectors and corresponding 𝐀𝐀𝐀𝐀 = 𝜆𝜆𝐯𝐯 ⟹ � �� � = � �
equation to obtain 2 −4 2 − λ 2 3 𝑦𝑦 𝑦𝑦
eigenvalues of a given matrix 𝐀𝐀, orthogonal
the eigenvalues. = (𝟏𝟏 − λ)�(𝟑𝟑 − λ)(2 − λ) − (0)(−4)� − (2)(0(2 − λ) − (0)(2))
diagonalisation. Equating elements from either side,
𝐀𝐀𝐀𝐀 = 𝜆𝜆𝐯𝐯 ⟹ 𝐀𝐀𝐀𝐀 = 𝜆𝜆𝐈𝐈𝐈𝐈 +(0)�(0)(−4) − (2)(𝟑𝟑 − 𝜆𝜆)� 3𝑥𝑥 + 2𝑦𝑦 = 𝑥𝑥 𝑥𝑥 + 𝑦𝑦 = 0
Find the normalised ⟹
= (1 − λ)(𝜆𝜆2 − 5λ + 6) = −(λ − 1)(λ − 2)(λ − 3) ⟹ λ = 1, 2, 3 eigenvector for 𝜆𝜆1 . 2𝑥𝑥 + 3𝑦𝑦 = 𝑦𝑦 𝑥𝑥 + 𝑦𝑦 = 0
⟹ (𝐀𝐀 − λ𝐈𝐈)𝐯𝐯 = 𝟎𝟎 Find the eigenvector When λ = 1, These equations accept any solution if 𝑦𝑦 = −𝑥𝑥, ∴ 𝐯𝐯 = �−1 1
�
for 𝜆𝜆1 . 1 2 0 𝑥𝑥 𝑥𝑥 1�
where I is the identity matrix. For this to be true for a non-zero v, (𝐀𝐀 − λ𝐈𝐈) must be singular, i.e. 𝐀𝐀𝐀𝐀 = 𝜆𝜆𝐯𝐯 ⟹ �0 3 0� �𝑦𝑦� = �𝑦𝑦� 1 1
∴ 𝐯𝐯� = � � = � √2 �
2 −4 2 𝑧𝑧 𝑧𝑧 2
�1 + (−1) 2 −1 −1�
𝐝𝐝𝐝𝐝𝐝𝐝(𝑨𝑨 − 𝝀𝝀𝝀𝝀) = 𝟎𝟎. √2
𝑥𝑥 + 2𝑦𝑦 = 𝑥𝑥 𝑦𝑦 = 0 Find the normalised When λ = 5,
This is the characteristic equation of 𝐀𝐀 and is a polynomial in 𝜆𝜆. The roots are the eigenvalues of 3𝑦𝑦 = 𝑦𝑦 ⟹ 𝑦𝑦 = 0 eigenvector for 𝜆𝜆2 . 3 2 𝑥𝑥 𝑥𝑥
A. The respective eigenvectors can then be found by plugging each root 𝜆𝜆 back into the equation. 𝐀𝐀𝐀𝐀 = 𝜆𝜆𝐯𝐯 ⟹ � � � � = 𝟓𝟓 � �
2𝑥𝑥 − 4𝑦𝑦 + 2𝑧𝑧 = 𝑧𝑧 𝑧𝑧 = 𝑦𝑦 − 2𝑥𝑥 2 3 𝑦𝑦 𝑦𝑦
1 Equating elements from either side,
0 1 These equations accept any solution if 𝑧𝑧 = −2𝑥𝑥 ∴ 𝐯𝐯 = � 0 �
Example: Find the eigenvalues and eigenvectors of 𝐀𝐀 = � �. 3𝑥𝑥 + 2𝑦𝑦 = 5𝑥𝑥 −𝑥𝑥 + 𝑦𝑦 = 0
−2 −3 −2 ⟹
Find the eigenvector When λ = 2, 2𝑥𝑥 + 3𝑦𝑦 = 5𝑦𝑦 𝑥𝑥 − 𝑦𝑦 = 0
Find and solve the 0−λ 1 for 𝜆𝜆2 . 1 2 0 𝑥𝑥 𝑥𝑥 These equations accept any solution if 𝑦𝑦 = 𝑥𝑥, ∴ 𝐯𝐯 = �11�
det(𝐀𝐀 − λ𝐈𝐈) = � �
characteristic equation to −2 −3 − λ 𝐀𝐀𝐀𝐀 = 𝜆𝜆𝐯𝐯 ⟹ �0 3 0� �𝑦𝑦� = 𝟐𝟐 �𝑦𝑦� 1�
2
= (−λ)(−3 − λ) − (−2)(1) = λ + 3λ + 2 = 0 ⟹ λ = −1, −2 1 1
obtain the eigenvalues 2 −4 2 𝑧𝑧 𝑧𝑧 ∴ 𝐯𝐯� = � � = � √2�
Find the eigenvector for When λ = −1, √12 + 12 1 1�
𝑥𝑥 𝑥𝑥 𝑥𝑥 + 2𝑦𝑦 = 2𝑥𝑥 𝑥𝑥 = 2𝑦𝑦 √2
𝜆𝜆1 . Any multiple of this 𝐀𝐀𝐀𝐀 = 𝜆𝜆𝐯𝐯 ⟹ �
0 1
�� � = −� � 3𝑦𝑦 = 2𝑦𝑦 ⟹ 𝑦𝑦 = 0 𝐏𝐏 is the matrix of
vector is also an −2 −3 𝑦𝑦 𝑦𝑦
2𝑥𝑥 − 4𝑦𝑦 + 2𝑧𝑧 = 2𝑧𝑧 𝑥𝑥 = 2𝑦𝑦 normalised
eigenvector. Equating the top and bottom elements from either side
0𝑥𝑥 + 𝑦𝑦 = −𝑥𝑥 𝑥𝑥 + 𝑦𝑦 = 0 0 eigenvectors, and 𝐃𝐃 1� 1�
These equations accept any solution if 𝑥𝑥 = 𝑦𝑦 = 0 ∴ 𝐯𝐯 = �0� 1 0
−2𝑥𝑥 − 3𝑦𝑦 = −𝑦𝑦
⟹
2𝑥𝑥 + 2𝑦𝑦 = 0 1 is the diagonal 𝐏𝐏 = � √2 √2� , 𝐃𝐃 = � �
matrix of −1� 1� 0 5
Both equations give the same information and accept any solution if Find the eigenvector When λ = 3, √2 √2
𝑦𝑦 = −𝑥𝑥 ∴ 𝐯𝐯 = �−1 1
� is an acceptable eigenvector. for 𝜆𝜆3 . 1 2 0 𝑥𝑥 𝑥𝑥 eigenvalues, as
𝐀𝐀𝐀𝐀 = 𝜆𝜆𝐯𝐯 ⟹ �0 3 0� �𝑦𝑦� = 𝟑𝟑 �𝑦𝑦� before.
Find the eigenvector for When λ = −2, 2 −4 2 𝑧𝑧 𝑧𝑧
𝜆𝜆2 . Any multiple of this 0 1 𝑥𝑥 𝑥𝑥
𝐀𝐀𝐀𝐀 = 𝜆𝜆𝐯𝐯 ⟹ � � � � = −2 � � 𝑥𝑥 + 2𝑦𝑦 = 3𝑥𝑥 𝑥𝑥 = 𝑦𝑦 Diagonal matrices are easier to perform long calculations with. For any diagonal 𝑘𝑘 × 𝑘𝑘 matrix 𝐃𝐃,
vector is also an −2 −3 𝑦𝑦 𝑦𝑦
3𝑦𝑦 = 3𝑦𝑦 ⟹ 𝑦𝑦 = 𝑦𝑦
eigenvector. Equating the top and bottom elements from either side
0𝑥𝑥 + 𝑦𝑦 = −2𝑥𝑥 2𝑥𝑥 + 𝑦𝑦 = 0 2𝑥𝑥 − 4𝑦𝑦 + 2𝑧𝑧 = 3𝑧𝑧 𝑧𝑧 = 2𝑥𝑥 − 4𝑦𝑦 𝑎𝑎1 0 0 𝑎𝑎1𝑛𝑛 0 0
⟹ These equations accept any solution if 𝑥𝑥 = 𝑦𝑦, 𝑧𝑧 = 2𝑥𝑥 − 4𝑦𝑦. If 𝐃𝐃 = � 0 ⋱ 0 �, 𝐃𝐃𝑛𝑛 = �0 ⋱ 0 �.
−2𝑥𝑥 − 3𝑦𝑦 = −2𝑦𝑦 2𝑥𝑥 + 𝑦𝑦 = 0
Both equations give the same information and accept any solution if 1 0 0 𝑎𝑎𝑘𝑘 0 0 𝑎𝑎𝑘𝑘𝑛𝑛
1
Taking 𝑥𝑥 = 𝑦𝑦 = 1, 𝐯𝐯 = � 1 �
𝑦𝑦 = −2𝑥𝑥 ∴ 𝐯𝐯 = �−2 � is an acceptable eigenvector. −2
𝐏𝐏 is the matrix of Cayley-Hamilton Theorem
eigenvectors. 𝐃𝐃 is the
diagonal matrix of 1 0 1 1 0 0 This theorem states that every square matrix A is a solution to its own characteristic equation.
If 𝐀𝐀 were a 3 × 3 matrix, the characteristic equation would be a cubic equation and there would 𝐏𝐏 = � 0 0 1 �, 𝐃𝐃 = �0 2 0� Plugging a matrix into a polynomial feels unnatural, but it works by considering 𝐈𝐈 as 1.
eigenvalues, in the
be up to 3 eigenvalues and eigenvectors. A normalised eigenvector is an eigenvector of length 1. −2 1 −2 0 0 3
𝐕𝐕
same order as their 0 1
If 𝐯𝐯 is an eigenvector, its normalised vector is given by 𝐯𝐯� = , where |𝐯𝐯| is the length of 𝐯𝐯. eigenvectors in 𝐏𝐏. Example: Verify the Cayley-Hamilton Theorem on 𝐀𝐀 = � � and use it to find its inverse.
|𝐯𝐯| −2 −3
Diagonalisation Find the det(𝐀𝐀 − λ𝐈𝐈) = �
0−λ 1
� = 0 = 𝜆𝜆(𝜆𝜆 + 3) + 2 = 𝜆𝜆2 + 3𝜆𝜆 + 2 = 0
There is a relationship between any 𝐀𝐀, 𝐏𝐏 and 𝐃𝐃 in general: characteristic −2 −3 − λ
A square matrix is diagonal if all elements that are not on its leading diagonal are zero. The equation of 𝐀𝐀.
leading diagonal is the line of elements from the top left to bottom right of a square matrix. 𝐃𝐃 = 𝐏𝐏−1 𝐀𝐀𝐀𝐀. Use the Cayley- 𝑴𝑴𝒖𝒖𝒖𝒖𝒖𝒖𝒖𝒖𝒖𝒖𝒖𝒖𝒖𝒖 𝒃𝒃𝒃𝒃 𝐀𝐀−𝟏𝟏
𝐀𝐀2 + 3𝐀𝐀 + 2𝐈𝐈 = 𝟎𝟎 ������������ 𝐀𝐀 + 3𝐈𝐈 + 2𝐀𝐀−𝟏𝟏 = 𝟎𝟎
Hamilton
2 0 −7 0 0 0 0 9 0 0 0 If 𝐀𝐀 = 𝐀𝐀T , its entries are a reflection along the leading diagonal. Such a matrix is called
� � Theorem.
0 3 � 0 𝑒𝑒 𝜋𝜋𝜋𝜋 0� �0 4 0 � �2 0 0 � symmetric, and its normalised eigenvectors are always orthogonal to each other. A matrix
Rearrange for 1 3 3 1 0 −3 −1�
� = � �2
0 0 1 1 0 0 0 6 0 1 0
made of orthogonal vectors is called an orthogonal matrix. Orthogonal matrices have the 𝐀𝐀−1 = − 𝐀𝐀 − 𝐈𝐈 = − �
1
�− � 2�
𝐀𝐀−𝟏𝟏 . 2 2 2 −2 −3 2 0 1 1 0
Diagonal Diagonal Not Diagonal (The leading Not Diagonal (The leading property that their transpose is their inverse and vice versa. In this case, orthogonal
diagonal is not the only diagonal is not filled in) diagonalisation is used.
collection of non-zero elements)
If 𝐀𝐀 is a symmetric matrix, 𝐃𝐃 = 𝐏𝐏𝑇𝑇 𝐀𝐀𝐀𝐀 since if 𝐀𝐀 is symmetric, 𝐏𝐏 is orthogonal. Hence, 𝐏𝐏T = 𝐏𝐏−𝟏𝟏 .

https://bit.ly/pmt-cc
https://bit.ly/pmt-edu https://bit.ly/pmt-cc