1/56
Looks like no tags are added yet.
QR factorization: columns of Q are…
orthonormal (QTQ=Ir)
QR factorization: projection onto Col(A) is…
P=QQT
QR factorization: least square equation…
Rx=QTb
QR factorization: A and R have common…
Gramians and rank
How to find a QR factorization
Apply gram-schmidt to A to get Q, then R=QTA
Determinant of a triangular matrix is the…
product of its diagonal
det(AT)=
det(A)
det(AB)=
det(A)det(B)
For PA=LU factorization, det(A)=
det(P)det(L)det(U)
Minor (Mij)
det(Aij)
Cofactor (Cij)
(-1)I+jdet(Aij)
Adjugate matrix is
the transpose of the cofactor matrix
A is nonsingular is det(A)
does not equal zero
if A is invertible, solve Ax=b by
dividing a vector with the determinants of A with the ith column replaced by b by the determinant of A (Cramer’s rule)
Positively oriented
counter clockwise, right hand rule
Negatively oriented
clockwise, left hand rule
coplanar
linearly dependent vectors
Complex multiplication
(a1a2-b1b2)+(a1b2+a2b2)i
complex absolute value
root(a2+b2)
complex inner product <v,w>=
conjugate v (w)
magnitude of complex vector
square and add the real and imaginary parts of each entry in a vector, then take square root
conjugate transpose (A*)
conjugate(AT)
The eigenvalues of A* are the
conjugates of the eigenvalues of A
Hermitian matrix
H*=H (ST=S)
Unitary matrix
U*=U-1 (U*U=In, QT=Q-1)
Fundamental Theorem of Algebra: every monic f in C[t] factors as
f(t)=(t-r1)m1(t-r2)m2…
deg(f) =
m1+m2+mk (mj is the multiplicity of rj as a root of f)
Vieta’s formulas for monic polynomials
co=(-1)n (r1m1) (rkmk)
cn-1= -(m1r1+mkrk)
Characteristic Polynomial XA(t)=
det(tIn- A)
If A is nxn, then XA(t) is
monic with degree n
Characteristic polynomial of the transpose of A is
the same as the polynomial for A
The roots of XA(t) are the
eigenvalues of A
Characteristic polynomial of every nxn matrix A is of the form…
XA(t) = tn - trace(A)tn-1 + … + (-1)ndet(A)
algebraic multiplicity
multiplicity of each eigenvalue in the characteristic polynomial (amA(lamda))
sum of algebraic multiplies is
n
multiplicity inequality for characteristic matrix
1 <= gm(lamda) <= am(lamda)
A is diagonizable if…
A=XDX-1
AK=
XBKX-1
an nxn matrix is diagonalizable if and only if:
geometric multiplicities of all lamdas add up to n
geometric multiplicities = algebraic multiplicities
Similar matrices have the same…
determinants, traces, characteristic polynomials, eigenvalues, algebraic multiplicities, ranks, geometric multiplicities,
Eigenspace of similar matrix A
is every eigenvector of B multiplied by X.
The eigenvalues of every Hermitian matrix are
real
The eigenspaces of a Hermitian matrix are
orthogonal
Every Hermitian matrix is
diagonizable
Hermitian matrices have the spectral factorization
H=UDU*
A quadratic form is given by
q(x)=<x,Sx>
quadratic forms do not have
constants
eigenvalues determine what for quadratic forms
definitness
Completing the square
y=UTx
Gramians are always
positive semidefinite
Gramians are only positive definite if
A has full column rank
exp(D)
take e to each number on the diagonal
exp(A)=
Xexp(D)X-1
If A is diagonizable, u(t)=
Xexp(Dt)X-1uo
det(eA) =
etrace(A)
ea+bi
ea(cos(b) + isin(b))
eipi=
-1