what is a matrix
a rectangular array of numbers/ symbols/ functions… arranged into rows + columns
mxn matrix means m rows n collumns
entry in M in row i and column j
Mij
transpose
if M is an mxn matrix
M^t is the matrix with entires
(M^t)ij = Mji
becomes an nxm matrix
is a reflection of M in the diagonal starting top left
(A^t)^t = A
symmetric matrix
M^t = M
skew symmetric matrix
M^t = -M
if skew symmetric
Aii = -Aii = 0
any matrix is the sum of a symmetric and skew symmetric matrix
zero matrix
all entries are 0
identity matrix
In is the square nxn matrix with entries
Iij = δij =
0 i=/= j or 1 i=j
Mmxn(R)
the set of all mxn matrices with real entries
matrix addition
(A+B)ij = Aij + Bij
scalar multiplication
(λA) ij = λ(Aij)
(-A) = (-1)A
properties of addition
1) A + 0 = A = 0 + A (additive unit)
2) A + B = B + A (commutativity)
3) A + (-A) = 0 (additive inverse)
4) A + (B+C) = (A+B) + C (associativity)
prove by writing out entry by entry
properties of scalar multiplication
1) 0A = 0
2) 1A = A
3) (λμ) A = λ(μA) (associativity)
4a) (λ + μ)A = λA + μA
b) λ(A + B) = λA + λB (distributivity)
prove by writing out entry by entry
matrix multiplication
Mmxn(R) x Mnxp(R) -> Mmxp(R)
(AB)ij = ai1b1j + ai2b2j+…+ ainbnj
= Σaik bkj from k=1 to n
properties of matrix multiplication
1) 0X = 0
2) IX = X
3) λ(XY) = (λX)Y = X(λY)
4) (XY)Z = X(YZ) (associativity)
5a) (X + X’)Y = XY + X’Y
X(Y + Y’) = XY + XY’
b) λ(X + X’) = λX + λX’
(λ + λ’)X = λX + λ’X
(distributivity)
proof of 0X = 0
0ij = 0
0Xij = 0 for all ij
proof of IX = X
(IX)ij = Σδik akj =aij
proof of associativity
X(YZ) = (XY)Z
X mxn Y nxp Z nxq
(X(YZ))ij = Σ xik(YZkj)
= Σ xik( Σykr zrj)
= Σxik ykr zrj
for k 1…n and r 1..p
((XY)Z)ij = Σ (XY)ir zrj
= Σ (Σxik ykr) xrj
= Σ xik ykr zrj
for k 1…n and r 1..p
= (X(YZ))ij
invertible
A is invertible or non-singular if there is an nxn matrix B with
AB = I = BA
B is the inverse of A = A^-1
no inverse
non invertible
singular
no B exists where AB = I
properties of inverse/ transpose
1) a matrix can have at most 1 inverse
2) if X and Y are invertible then XY is invertible and (XY)^-1 = Y^-1X^-1
3) (MN)^t = N^t M^t
and (λM)^t = λM^t
4) if A is invertible then A^t is invertible
(A^t)^-1 = (A^-1)^t
5) if A and B are square matrices and AB = I then BA = I
proof that a matrix can have at most 1 inverse
assume 2 inverses B,B’
AB = BA = I
AB’ = B’A = I
(BA)B’ = B(AB’)
IB’ = BI
B’ = B
prove
(A^t)^-1 = (A^-1)^t
(A^t)(A^-1)^t = (A^-1A)^t = I^t = I
so (A^-1)^t is inverse to A^t
so (A^-1)^t = (A^t)^-1
transpose distributivity
(X +Y)^t
= X^t + Y^t
solving equations using matrix
can reqrite as a matrix times (x,y,z) and use inverse to find x,y,z