# What are eigenvalues and eigenvectors 4

**content**

“Preliminary remark

" The definition

»The characteristic polynomial and the algebraic multiplicity

»Eigenvectors and the geometric multiplicity

“Examples

#### Preliminary remark

Eigenvalues and eigenvectors are one of the most important topics in linear algebra. Their applications are very diverse. The idea is a simple one, we are looking for vectors \ (\ vec v_i \) which are mapped to a multiple of themselves by the matrix \ (A \), in formulas that means

\ begin {align *}

A \ vec {v} _i = \ lambda_i \ vec {v} _i.

\ end {align *}

#### The definition

Let us formalize the above definition even further. As so often, we use the letters \ (u, v, w \) for vectors instead of \ (\ vec {v}, \ vec {u}, \ vec {w} \). Matrices are named with capital letters \ (A, B, \ dots \) and scalars with \ (\ lambda, \ alpha \). In addition, we only deal with matrices over the real field \ (\ mathbb {R} \). We consider the \ ((n \ times n) \) - matrix \ (A \) as a linear mapping over the vector space \ (\ mathbb {R} ^ n \). If a vector \ (v \ in V \) satisfies the equation

\ begin {align *}

Av = \ lambda v

\ end {align *}

for a \ (\ lambda \ in \ mathbb {R} \), we call a \ (v \) eigenvector with associated eigenvalue \ (\ lambda \).

A trivial example would be the matrix

\ begin {align *}

A = \ begin {pmatrix}

1 & 0\\

0 & 1

\ end {pmatrix}

\ end {align *}

with the associated eigenvectors \ (v_1 = (1; 0) \) and \ (v_2 = (0; 1) \), both have the eigenvalue \ (\ lambda_1 = 1 \). The matrix \ (A \) with

\ begin {align *}

\ left (

\ begin {array} {ccc}

0 & \ frac {1} {2} & \ frac {1} {2} \

\ frac {1} {2} & 0 & \ frac {1} {2} \

\ frac {1} {2} & \ frac {1} {2} & 0 \

\ end {array}

\ right)

\ end {align *}

has the eigenvectors

\ begin {align *}

v_1 = \ begin {pmatrix}

1\\1\\1

\ end {pmatrix}, v_2 = \ begin {pmatrix}

-1\\0\\1

\ end {pmatrix},

v_3 = \ begin {pmatrix}

-1\\1\\0

\ end {pmatrix}

\ end {align *}

with the eigenvalues

\ begin {align *}

\ lambda_1 = 1, \ lambda_2 = \ lambda_3 = - \ frac {1} {2}.

\ end {align *}

One can easily recalculate this, but how did we find the vectors and the eigenvalues?

#### The characteristic polynomial and algebraic multiplicity

We derive the characteristic polynomial so that eigenvalues can be calculated. The train of thought is a simple one, fulfilling eigenvalue and eigenvector

\ begin {align *}

& Av = \ lambda v \

& Av- \ lambda v = 0 \

& (A- \ lambda I_n) v = 0,

\ end {align *}

where \ (I_n \) is the \ ((n \ times n) \) - identity matrix. Now the linear mapping / matrix \ ((A- \ lambda I_n) \) maps the vector \ (v \) to the zero vector. We therefore know that \ ((A- \ lambda I_n) \) cannot be injective, or that it is more important for us and does not have full rank. If the matrix \ ((A- \ lambda I_n) \) does not have a full rank, we know (LINK) that its determinant is 0. So we get

\ begin {align *}

\ det (A- \ lambda I_n) = 0.

\ end {align *}

For example, for the above matrix we calculate:

\ begin {align *}

\ det (

\ begin {pmatrix}

0 & \ frac {1} {2} & \ frac {1} {2} \

\ frac {1} {2} & 0 & \ frac {1} {2} \

\ frac {1} {2} & \ frac {1} {2} & 0 \

\ end {pmatrix}

- \ lambda

\ begin {pmatrix}

1 & 0 & 0 \\

0 & 1 & 0 \\

0 & 0 & 1 \\

\ end {pmatrix}) = 0 \

\ rightarrow \ det (\ begin {pmatrix}

- \ lambda & \ frac {1} {2} & \ frac {1} {2} \

\ frac {1} {2} & - \ lambda & \ frac {1} {2} \

\ frac {1} {2} & \ frac {1} {2} & - \ lambda \

\ end {pmatrix} = 0 \

- \ frac {1} {4} (\ lambda-1) (2 \ lambda + 1) ^ 2 = 0

\ end {align *}

This equation can be solved and has the solutions \ (\ lambda_1 = 1 \) and \ (\ lambda_ {2,3} = - \ frac {1} {2} \), these are our eigenvalues. We therefore get eigenvalues by solving the so-called characteristic polynomial under the equation \ (\ det (A- \ lambda I_n) = 0 \).

It is said that \ (\ lambda_1 = 1 \) has algebraic multiplicity 1 and \ (\ lambda_ {2,3} = - \ frac {1} {2} \) has algebraic multiplicity 2, because it is a double solution.

#### Eigenvectors and geometric multiplicity

We just learned how to calculate eigenvalues. If we have this, the loosening of the associated one leads to it

\ begin {align *}

(A- \ lambda I_n) v = 0

\ end {align *}

after \ (v \). We already know that this system must always have a solution, because rank (\ ((A- \ lambda I_n) \)) = rank (\ ((A- \ lambda I_n) \) | 0). Let's get started and insert \ (\ lambda_1 \) to calculate \ (v_1 \)

\ begin {align *}

\ begin {pmatrix}

-1 & \ frac {1} {2} & \ frac {1} {2} \

\ frac {1} {2} & -1 & \ frac {1} {2} \

\ frac {1} {2} & \ frac {1} {2} & -1 \

\ end {pmatrix} v = 0

\ end {align *}

This system of equations (here (LINK) more about how we solve it) has the solution

\ begin {align *}

L = \ {s \ begin {pmatrix}

1\\1\\1

\ end {pmatrix} | s \ in \ mathbb {R} \}

\ end {align *}

and thereby we get the eigenvector \ (v = (1; 1; 1) \). The dimension of the solution is 1, we call this dimension the geometric multiplicity. The eigenvector (s) for the eigenvalue \ (\ lambda_ {2,3} = - \ frac {1} {2} \) (with algebraic multiplicity 2) are obtained by solving

\ begin {align *}

\ begin {pmatrix}

\ frac {1} {2} & \ frac {1} {2} & \ frac {1} {2} \

\ frac {1} {2} & \ frac {1} {2} & \ frac {1} {2} \

\ frac {1} {2} & \ frac {1} {2} & \ frac {1} {2} \

\ end {pmatrix} v = 0.

\ end {align *}

The matrix of this system of equations obviously has rank 1, therefore the solution of the system of equations has rank 2. It is given by

\ begin {align *}

L = \ {s \ begin {pmatrix}

-1\\0\\1

\ end {pmatrix} + r \ begin {pmatrix}

-1\\1\\0

\ end {pmatrix} | s, r \ in \ mathbb {R} \}

\ end {align *}

and this gives us the two eigenvectors \ (v_2 = (- 1; 0; 1) \) and \ (v_3 = (- 1; 1; 0) \). The eigenvalue \ (\ lambda_ {2,3} = - \ frac {1} {2} \) therefore has geometric multiplicity 2. In the examples you will find matrices whose eigenvalues have a greater algebraic than geometric multiplicity.

#### Examples

Upper triangular matrix: Calculate eigenvalues and eigenvectors of the matrix \ (B \) with

\ begin {align *}

B = \ left (

\ begin {array} {ccc}

1 & 2 & 3 \\

0 & 4 & 5 \\

0 & 0 & 6 \\

\ end {array}

\ right).

\ end {align *}

What do you notice?

Solution: It is an upper triangular matrix. Thereby we very easily get the characteristic polynomial by expanding after the first column

\ begin {align *}

\ det (B- \ lambda I_3) = - (\ lambda-6) (\ lambda-4) (\ lambda-1).

\ end {align *}

We recognize that with an upper triangular matrix we can apparently read off the eigenvalues in the diagonal, because

\ begin {align *}

- (\ lambda-6) (\ lambda-4) (\ lambda-1) = 0

\ end {align *}

has, according to the product zero theorem, the solutions \ (\ lambda_1 = 6, \ lambda_2 = 4, \ lambda_3 = 1 \). We calculate the associated eigenvectors by solving the three systems of equations

\ begin {align *}

(B-6I_3) v = 0, (B-4I_3) v = 0, (B-1I_3) v = 0

\ end {align *}

and get the three eigenvectors \ (v_1 = (16; 25; 10) \), \ (v_2 = (2,3,0) \) and \ (v_3 = (1,0,0) \).

Geometric multiplicity: Upper triangular matrix: Calculate eigenvalues and eigenvectors of the matrix \ (B \) with

\ begin {align *}

B = \ left (

\ begin {array} {ccc}

4 & 1 & 0 \\

0 & 4 & 0 \\

0 & 0 & 1 \\

\ end {array}

\ right).

\ end {align *}

What do you notice?

Solution: As explained in the previous example, we can read off the eigenvalues, \ (\ lambda_ {1,2} = 4 \) and \ (\ lambda_3 = 1 \) apply. \ (V_3 \) can easily be calculated as \ (\ lambda_3 \), it gives \ (v_3 = (0; 0; 1) \). Interesting is the eigenvalue \ (\ lambda_ {1,2} = 4 \) with algebraic multiplicity 2 then

\ begin {align *}

& (B-4I_3) v = 0 \ \

& \ left (

\ begin {array} {ccc}

0 & 1 & 0 \\

0 & 0 & 0 \\

0 & 0 & -3 \\

\ end {array}

\ right) v = 0

\ end {align *}

has only a one-dimensional solution, because \ ((B-4I_3) \) has rank 2. This solution is given by

\ begin {align *}

L = \ {s \ begin {pmatrix}

1\\0\\0

\ end {pmatrix} | s \ in \ mathbb {R} \}

\ end {align *}

and therefore we have only one eigenvector, \ (v_1 = (1; 0; 0) \). The eigenvalue \ (\ lambda_ {1,2} = 4 \) therefore has algebraic multiplicity 2 but only geometric multiplicity 1.

\ (2 \ times 2 \): Compute eigenvalues and eigenvectors of the matrix \ (B \) with

\ begin {align *}

B = \ left (

\ begin {array} {cc}

1 & 2 \\

3 & 4 \\

\ end {array}

\ right).

\ end {align *}

The characteristic polynomial results in \ (\ lambda ^ 2-5 \ lambda -2 \), which has the rounded solutions \ (\ lambda_1 = 5.372 \) and \ (\ lambda_2 = -0.372 \). In general, eigenvalues are of course not integers, even if we did this in our previous examples for the sake of simplicity. The associated eigenvectors are \ (v_1 = (0.457; 1) \) and \ (v_2 = (-1.457, 1) \).

- Is graphite amorphous or crystalline
- Where are places in Florida underestimated
- Is aerospace engineering better than ocean engineering
- Why don't people like peppers
- Has anyone found a good virtual assistant
- What are the IBRD and IDA
- How do I find my lost headphones
- Are there apprenticeships in Liberia
- What is the purpose of corrections
- What is Rh negative and positive
- What are some examples of kosher laws
- Which airports do you hate
- How can I increase my brain chemicals
- What makes a good beer
- How to calculate a a b
- Why am I not getting my periods?
- Are glucose and fructose stereoisomers
- What is Activity_main XML in Android
- What is the Tamil word for busy
- How can I make eclairs
- How do you recognize anions and cations
- Dizziness is linked to trichotillomania
- Are protein vegan
- Wanted to be related to Barack Obama