Eigenvectors / Eigen Values (Cayley Hamilton) Flashcards
eigenvector/eigenvalue of a matrix
Is A is an nxn matrix over a field 𝔽 and v∈𝔽ⁿcol is non zero then v is an eigenvector with eigenvalue λ when
Av=λv
(where A is a matrix that represents a linear map)
An eigenvector has…
exactly one eigenvalue
An eigenvalue has…
many possible eigenvectors
if v is an eigenvector then…
every multiple of v is also an eigenvector
0 can and cant be…
0 can be an eigenvalue
0 can’t be an eigenvector
eigenspace
If λ is an eigenvalue, the corresponding eigenspace is given by key(f-λI) and is the set of all non zero eigenvectors corresponding to λ.
Eigenvalues are the roots of the equation…
det(f-xI) = 0
each eigenspace has dimension…
atleast 1
0 is an eigenvalue of f iff…
iff the nullity of f is non-zero
Lemma 9.3. eigenvalue of a square matrix.
A scalar λ is an eigenvalue of a square matrix A if and only if n(A-λI)>0
Corollary 9.5
A scalar λ∈ℂ is an eigenvalue of an nxn square matrix A (over ℂ or ℝ) iff det(A-λI)=0
If det(A-λI)=0 then…
there are infinitely many eigenvectors for λ
How to find eigenvalues
solve det(A-xI)=0
How to find eigenvectors
- find A-λI
- put into echelon form
- find ker(A-λI). Extract equations from this form.
- if equations each in terms of two variables only, put into a vector
- otherwise let x = equation in terms of other two variables. let a=y, b=z and write a () + b () where you put 1s in correct position for y and z in each and in x position put the coefficients of the equation.
if λ is an eigenvalue of a then:
- there is a v s.t. Av = λv
- dim ker (A - λI)>0
- nullity (A- λI) > 0
- rank (A - λI) < n
- det(A-λI)=0 (i.e. matrix is not invertible)
Characteristic polynomial of A
χₐ(x) = det(A - xI)
Proposition 10.1. eigenvalues in UT matrices.
The eigenvalues are the diagonal entries of fᵢᵢ.
The characteristic polynomial can also be written as…
χₐ(x) = Π(fᵢᵢ-x)
Powers of A…
commute
Polynomials in A….
commute
Theorem 10.3. Lemma for cayley Hamilton.
If A is upper triangular then the characteristic polynomial evaluated at A is 0.
proposition 10.5. similar matrices and polynomials.
if p(x) is a polynomial over 𝔽 and A,B are similar then p(A) and P(B) are similar
Proposition 10.6. similar matrices are their characteristic polynomials.
if B = P⁻¹AP then χₐ(x) = χb(x)
Theorem 10.7 (beore cayley hamilton)
Let A be a nxn square matrix over ℂ then A is similar to an upper triangular matrix.
Theorem 10.10 (the cayley hamilton theorem)
Let A be an nxn square matrix over ℂ or ℝ. Then χₐ(A)= det(A-A) = det(0) = 0
Theorem 10.11 (after cayley hamilton)
if χₐ(x) = (λ-x)ᵉp(x) where λ is not a root of p(x) then the nullity of the matrix (A-λI)ᵉ is e
How to find the determinant of a 2x2 matrix (a b , c d)
ad-bc
How to find the determinant of a 3x3 matrix (a b c , d e f , g h i)
expand which whichever column/row you want to e.g.
a | e f , h i | + b | f d , i g | + c | d e , g h |
how to find the inverse of a 3x3 matrix (a b c , d e f , g h i)
- find the determinant
- find all the cofactors of the matrix
- mirror across the diagonal
- divide by the determinant
how to find the inverse of a 2x2 matrix (a b , c d)
- find the determinant
- mirror the diagonal and switch non diagonal elements signs
- divide by the determinant
How to show a matrix can’t be put into upper triangular form (show A is not similar to an upper triangular matrix)
- find P, and its inverse
- calculate P⁻¹AP (or just calculate the bottom left element of P⁻¹AP)
- either its non zero so not upper triangular or, if you don’t know P show that the equation can’t be solved.
How to find the upper triangular form for a matrix
- find eigenvalues of A
- find the eigenspaces of A
- if the multiplicity of each element of χₐ(x) is not equal to the corresponding dimension of the eigenspace then can’t diagonalise. Find the eigenspaces for each eigenvalue.
- for the eigenvalue where multiplicty isnt equal to dimenision look at im(A-λI) (column space)
- find 3 spanning vectors for A (apply column operations to find possible vectors)
- Find A w.r.t these bases and write as a linear combination then put into the the matrix.