EIGENVALUES AND

EIGENVECTORS
GEOPHYSICS UI

- fatimah
Definition of Eigenvalues x λx

and Eigenvectors
x
0≤𝜆≤1
𝜆≥1
λx x
If 𝐴 is an 𝑛 × 𝑛 matrix, and 𝐴x is a scalar multiple of
x which is a nonzero vector in 𝑅𝑛 . x
eigenvalue of 𝐴
𝑦 −1 ≤ 𝜆 ≤ 0

𝐴x = 𝜆x 𝜆 ≤ −1

eigenvector corresponding to 𝜆 λx
3x λx
6

4 1 3 0
The vector x = is an eigenvector of 𝐴 = corresponding to 𝜆 = 3
2 8 −1
eigenvalue of 𝐴
2
x 3 0 1 3 1
𝐴x = = =3 = 𝟑x
8 −1 2 6 2
𝑥 eigenvector corresponding to 𝜆 = 3
2 4
Geometrically, multiplication by 𝐴 has stretched the vector x by a factor of 3

LINEAR ALGEBRA 2
 Computing Eigenvalues
and Eigenvectors
𝐴x = 𝜆x
𝐴x = 𝜆Ix
𝜆I − 𝐴 x = 0

det 𝜆I − 𝐴 = 0 Characteristic equation of A 𝜆 0 ⋯ 0 𝑎11 𝑎12 ⋯ 𝑎1𝑛

0 𝜆 ⋯ 0 𝑎21 𝑎22 ⋯ 𝑎2𝑛
𝜆𝐼 = 𝐴= ⋮ ⋮ ⋯ ⋮
⋮ ⋮ ⋯ ⋮
If 𝐴 is an 𝑛 × 𝑛 matrix, then 𝜆 is an eigenvalue of 𝐴 if 𝑎𝑛1 𝑎𝑛2 ⋯ 𝑎𝑛𝑛
0 0 ⋯ 𝜆
and only if it satisfies the equation
𝜆 − 𝑎11 −𝑎12 ⋯ −𝑎1𝑛
det 𝜆I − 𝐴 = 0 −𝑎21 𝜆 − 𝑎22 ⋯ −𝑎2𝑛
𝜆𝐼 − 𝐴 =
⋮ ⋮ ⋯ ⋮
−𝑎𝑛1 −𝑎𝑛2 ⋯ 𝜆 − 𝑎𝑛𝑛
𝜆𝑛 + 𝑐1 𝜆𝑛−1 + ⋯ + 𝑐𝑛 = 0
Characteristic 𝜆 − 𝑎11 −𝑎12 ⋯ −𝑎1𝑛
𝑝 𝜆 = 𝜆𝑛 + 𝑐1 𝜆𝑛−1 + ⋯ + 𝑐𝑛 polynomial of A −𝑎21 𝜆 − 𝑎22 ⋯ −𝑎2𝑛
det 𝜆𝐼 − 𝐴 =
⋮ ⋮ ⋯ ⋮
−𝑎𝑛1 −𝑎𝑛2 ⋯ 𝜆 − 𝑎𝑛𝑛
The matrix has at most of 𝑛 distinct eigenvalues

LINEAR ALGEBRA 3
For upper triangular matrix Finding Eigenvalues
𝜆 0 ⋯ 0 𝑎11 𝑎12 ⋯ 𝑎1𝑛
0 𝜆 ⋯ 0 𝑎21 𝑎22 ⋯ 𝑎2𝑛 3 0
𝜆𝐼 = 𝐴= ⋮ We found that 𝜆 = 3 is an eigenvalue of 𝐴 =
⋮ ⋮ ⋯ ⋮ ⋮ ⋯ ⋮ 8 −1
0 0 ⋯ 𝜆 𝑎𝑛1 𝑎𝑛2 ⋯ 𝑎𝑛𝑛 How to find all the eigenvalues of 𝐴?

𝜆 − 𝑎11 −𝑎12 ⋯ −𝑎1𝑛 det 𝜆I − 𝐴 = 0

0 𝜆 − 𝑎22 ⋯ −𝑎2𝑛
det 𝜆𝐼 − 𝐴 = 𝜆−3 0
⋮ ⋮ ⋯ ⋮ =0
0 0 ⋯ 𝜆 − 𝑎𝑛𝑛 −8 𝜆+1

𝜆 − 𝑎11 𝜆 − 𝑎22 … 𝜆 − 𝑎𝑛𝑛 = 0 𝜆−3 𝜆+1 = 0

The eigenvalues are 𝜆=3 𝜆 = −1

𝜆 = 𝑎11 , 𝜆 = 𝑎22 , …, 𝜆 = 𝑎𝑛𝑛 𝜆2 − 2𝜆 − 3 = 0

If 𝐴 is a 𝑛 × 𝑛 triangular matrix, then the eigenvalues The eigenvalues of 𝐴 are 𝜆 = 3 and 𝜆 = −1.
of 𝐴 are the entries on the main diagonal of 𝐴

LINEAR ALGEBRA
The eigenvectors of 𝐴 corresponding to an 𝜆I − 𝐴 x = 0
eigenvalue 𝜆 are the nonzero vectors that
satisfy 𝜆−3 0 𝑥1 0
=
−8 𝜆 + 1 𝑥2 0
𝜆𝐼 − 𝐴 x = 0
In the case where 𝜆 = 3
The eigenspace of 𝐴 is the solution space (null
space) of the matrix 𝜆𝐼 − 𝐴 0 0 𝑥1 0
=
−8 4 𝑥2 0
0 0

Eigenspaces
=
−8𝑥1 + 4𝑥2 0

The general solution is 𝑥1 = 𝑡, 𝑥2 = 2𝑡

𝑥1 𝑡 1
Find bases for the eigenspaces of the matrix 𝑥2 = = 𝑡
2𝑡 2

3 0 1
𝐴= Then it follows that is a basis for the
8 −1 2
eigenspace corresponding to 𝜆 = 3

LINEAR ALGEBRA 5
Eigenvalues and Invertibility
A square matrix 𝐴 is invertible if and only
if 𝜆 = 0 is not an eigenvalue of 𝐴 The inverse of 𝐴 is
1
1 −1 0
0 3
𝐴−1 = =
3 0 −3 −8 3 8
Matrix 𝐴 = is invertible since the −1
8 −1 3
eigenvalues are 3 and −1, neither is 0. Here is the proof

1
We can show that det 𝐴 ≠ 0 0
3 0 3
𝐴𝐴−1 =𝐴=
8 −1 8
−1
3 0 3
det 𝐴 = = −3
8 −1
1 0
det 𝐴 ≠ 0 𝐴𝐴−1 = =𝐼
0 1

LINEAR ALGEBRA 6
References
Anton, H., Rorres, C., 2014, Elementary Linear Algebra
Application Version, Wiley, USA

LINEAR ALGEBRA 7