lec21
eigen vector
find a function, in goes vector x , out comes vector Ax .
$Ax parallelto x i.e. Ax=\lambda x$
If A is singular, λ=0 is an eigenvalue
projection matrix
for any x in the plane, Px=x i.e. λ=1
for any x perpendicular to the plane, Px=0 , i.e. λ=0
All eigenvalues
$n\times n matrixhas n$ eigenvalues.
Sum of these eigenvalues = tr(A) (some if diagonal)
Product of eigenvalues = det(A)
Solve Ax=λx
(A−λI)x=0
⇔ A−λIis singular⇔ det(A−λI)=0
E.g.
A=[3113]
det(A−λI)=3−λ113−λ=(3−λ)2−1=λ2−6λ+8(6:trace, 8:det)
Example for rotation matrix
Q=[ 0 1−10]
tr(A)=0=λ1+λ2det(A)=1=λ1λ2
λ1=i,λ2=−i(conjugated)
if A is symmetric, λ will have no imaginary parts.
if A=−A⊤ , λ will have pure imaginary parts.
Example for triangular matrix
for
[ 3 013]
x1=[10]
no independent x2 .