Objectives

 Subspaces and Spanning sets

 Independence and Dimension

 Orthogonality

 Rank of a Matrix
 Subspace of Rn
Definition of subspace of Rn. • n
• U
 Let Ø≠U be a subset of Rn ••
 U is called a subspace of Rn if: • •
 S1. The zero vector 0 is in U
vector zero vector
 S2. If X,Y are in U then X+Y is in U
 S3. If X is in U then aX is in U for all real number a.
 Ex1. U={(a,a,0)|aR} is a subspace of R3 n
aX
 the zero vector of R , (0,0,0)U
3
• • •
 (a,a,0), (b,b,0)U(a,a,0)+(b,b,0)=(a+b,a+b,0)U • •• X+Y
 If (a,a,0) U and k R, then k(a,a,0)=(ka,ka,0)U
U• •
Y
 Ex2. U={(a,b,1): a,b R} is not a subspace of R 3
X
 (0,0,0)U  U is not a subspace
 Ex3. U={(a,|a|,0)|a R} is not a subspace of R3
 (-1,|-1|,0), (1,|1|,0)U but (0,2,0) U  U is not a subspace
 Subspaces or not - Do yourself
 V={[0 a 0]T in 3: a Z}
 U={[a 0 a+1]T in 3: aR}
 W={[a b a-b]T in 3: a,bR}
 Q={[a b |a+b|]T: a }
 H={[a b ab]T: a }
 P={(x,y,z)| x-2y+z=0 and 2x-y+3z=0}. P is called the solution
space of the system x-2y+z=0 and 2x-y+3z=0.
 Null space and image space of a matrix
 A is an mxn matrix, if X is nx1 matrix then AX is mx1 matrix
 nullA = {X in Rn: AX=0}
m
 imA = {AX: X is in R }
n 
A
nullA • imA
n

zero vector

nullA ={X Rn:AX=0} is a subspace of Rn: imA ={AX:X  Rn}is a subspace of Rm:
 A.0=00nullA  0=A.00nullA
 X,Y nullA AX=0, AY=0  AX,AY imA AX+AY=A(X+Y)=AZ
A(X+Y)=AX+AY=0 (X+Y) nullA AX+AY imA
 X nullA, a R  AX=0   AX imA, a R  a(AX)=A(aX)=AZ
A(aX)=a(AX)=0  aXnullA  a(AX)imA
 Null space nullA={X:AX=0}

1 1 0 
 For example, A   
 2 3 1  23
 x   x    x   x 
     0      1 1 0    0  
nullA    y  : A  y         y  :    y   
  z   z  0     z   2 3 1   z  0  
         
 x    t  
   x  y  0    
  y  :     t  : t  
  z  2 x  3 y  z  0    5t  
       
 Eigenspaces (không gian riêng)
 Suppose A is an nxn matrix and λ is an eigenvalue of A
 Eλ(A)={X: AX=λX} is an subspace of Rn
 For example,
 3 1 x3 1
A   c A 
x  det  xI  A     x  3 x  2 
0 2 0 x2
c A  x   0  x  3  x  2
0 1 0 0 1 0 t 
x  3 :     X    (or X= t ,0 )
0 5 0  0 0 0  0 
5 1 0  t 
x  2:  X  
0 0 0  5t 
E3   X : AX  3 X   t ,0  : t  
Các không gian riêng
E2   X : AX  2 X   t , 5t  : t   ứng với GTR
 Spanning sets
 Y=k1X1+k2X2+…+knXn is called a linear combination of the vectors X1,X2,…,Xn
 The set of all linear combinations of the the vectors X1,X2,…,Xn is called the
span of these vectors, denoted by span{X1,X2,…,Xn }.
 This means, span{X1,X2,…,Xn} = {k1X1+k2X2+…+knXn :kiR is arbitrary}
 span{X1,X2,…,Xn} is a subspace of Rn.
 For example, span{(1,0,1),(0,1,1)}={a(1,0,1)+b(0,1,1) :a,bR}.
 And we have (1,2,3)span{(1,0,1),(0,1,1)} because (2,-3,-1)= 2(1,0,1)+ -
3(0,1,1).
 (2,3,2)span{(1,0,1),(0,1,1)} because (2,3,2)≠a(1,0,1)+b(0,1,1) for all a,b.
 Examples
 If x=(1,3,-5) is expressed as a linear combination of the vectors v1 = (1, 1,
1); v2 =(1,1,-1); v3 = (1, 0, 2); then the coefficient of v3 is:
A. 2 B. 3 C. -2  D. 1 E. 0
 x is expressed as a linear combination of v1, v2, v3 means x=av1+bv2+cv3 for
some a,b,c and c is called the coefficient of v3.
 the system is 1 1 1 1 1 1 1 1 1 1 1 1
a+b+c = 1 1 1 0 3 0 0 -1 2 0 -2 1 -6  a =1
a+b = 3 1 -1 2 -5 0 -2 1 -6 0 0 -1 2  b =2
a – b +2c =-5  c =-2
 Which of the vectors below is a linear combination of u=(1,1,2); v=(2,3,5)?
A. (0,1,1)  B. (1,1,0) C. (1,1,1) D. (1,0,1)  E. (0,0,1)
 Có thể giải bằng biến đổi sơ cấp trên ma trận chứa các vector cột như sau:

u v A B C D E u v A B C D E u v A B C D E
 2 0 1 1 1 0  -2 0 1 1 1 0  -2 0 1 1 1 0
1 3 1 1 1 0 0 0  1 0 0 -1 0 0  1 0 0 -1 0
2 5 1 0 1 1 1 0 1 1 -2 -1 -1 1 0 0 0 -2 -1 0 1
 Theorem

 U=span{X1,X2,…,Xn} is in Rn and U is a subspace of

Rn
 If W is a subspace of Rn such that Xi are in W
then U  W
W
• 4
• ••
U  span x1 , x2 , x3 , x4 , x5  •
 Linear Independence

A set of vectors in Rm {X1,X2,…,Xn } is called linearly independent

if t1X1+t2X2+…+tnXn=0  t1=t2=…=tn=0 only

numbers in R vectors in Rm
Ex1. The set {[1 -1]T, [2 3]T}R2 is called linearly independent since t1[1 -]T +
t2[2 3]T = [0 0]T follows t1=t2=0.
Ex2. A set of vectors that containing zero vector never linearly independent.
Ex3. The set {(0,1,1), (1,-1,0), (1,0,1)} is not linearly independent because the
system t1(0,1,1)+t2(1,-1,0)+t3(1,0,1)=(0,0,0) has one solution t1=1, t2=1, t3=1
 Examples
 Show that {(1,1,0);(0,1,1);(1,0,1)} is linearly independent in R 3
t1 1,1,0   t2  0,1,1  t3 1,0,1  0,0,0 
 ...  t1  t2  t3  0
t1 1,1,0   t2  0,1,1  t3 1,0,1   0,0,0 
1t1  0t2  1t3  0 1 0 1 0  1 0 1 0  1 0 1 0 
      
 1t1  1t2  0t3  0  1 1 0 0   0 1 1 0   0 1 1 0 
0t  1t  1t  0 0 1 1 0  0 0 2 0  0 0 1 0 
 1 2 3

 t1  t2  t3  0  independent

More ex. {(1,0,-2), (2,1,0), 1 2 0 -1 1 2 0 -1 1 2 0 -1

(0,1,5), (-1,1,0)} is not linearly 0 1 1 1 0 1 1 1 0 1 1 1
independent (number of
leading 1s = number of vectors) -2 0 5 0 0 4 5 2 0 0 1 -2
 Examples – do yourself
 Determine whether each the following sets is linearly
independent or linearly dependent.
 {(-1,2,0)}
 {(0,0,0); (1,2,3); (-3,0,1)}
 {(1,1,-1); (-1,1,1); (1,-1,1)}
 {(-2,3,4,1); (4,-1,5,0); (-2,1,0,3)}
 {(1,1,0); (-2,3,1); (5,0,1); (-1,0,2)}
 {X-Y+Z,3X+Z,X+Y-Z}, where {X,Y,Z} is an independent set of
vectors. (see below)
1 3 1  3 1  3 1  3 1  3 1
-1 0 1 0 3 2 0  1 0  1 0  1 independent
1 1 -1 0 -2 -2 0 3 2 0 0 -1 0 0 
 Fundamental Theorem
 Theorem. Let U be a subspace of Rn is spanned by m vectors,
if U contains k linearly independent vectors, then k≤m
 This implies if k>m, then the set of k vectors is always linear
dependece.
 For example, Let U be the space spanned by {(1,0,1), (0,-
1,1)} and S={(1,0,1), (0,-1,1), (2,-1,3)} U. Then, S is not
linearly independent (m=2, k=3).
 Basis and dimension
 Definition of basis: Suppose U is a subspace of Rn, a set {X1,X2,
…,Xk} is called a basis of U if U=span{X1,X2,…,Xk} and {X1,X2,
…,Xk} is linear independence
 Ex1. Let U={(a,-a)|aR}. Then U is a subspace of R2. Consider the set
B={(1,-1)}. B is linearly independent and U={(a,-a):aR}={a(1,-1): aR }
=span{(1,-1)}. So, B is a basis of U.
 Note that B’={(-1,1)} is also a basis of U.
 But {(1,1)} is not a basis of U because U can not be spanned by {(1,1)}
 Ex2. Given that V=span{(1,1,1), (1,-1,0), (0,2,1)}. Then, B={(1,1,1), (1,-1,0),
(0,2,1)} is not linearly independent, because (0,2,1)=(1,1,1) – (1,-1,0)B is
not a basis of U.
 Consider B’={(1,1,1), (1,-1,0)}. B’ is linearly independent and all vectors in
V are spanned by B’ because a(1,1,1)+ b(1,-1,0)+ c(0,2,1) =a(1,1,1)+ b(1,-
1,0)+ c(1,1,1) –c(1,-1,0) = (a+c)(1,1,1)+(b-c)(1,-1,0). So, B’ is a basis of V.
 Some important theorems
 Theorem 1. The following are equivalence for an nxn matrix A.
 A is invertible.
 the columns of A are linearly independent.
 the columns of A span Rn.
 the rows of A are linearly independent.
 the rows of A span the set of all 1xn rows.
 Theorem 2. (Invariance theorem). If {a1,a2,..,am} and {b1,b2,…,bk} are bases of a
subspace U of Rn, then m=k. In this case, m=k is called dimension of U and we
write dimU=m.
 Ex1. Let U={(a,-a)|aR} be a subspace of R2. Then, B={(1,-1)} is a basis of U and
B’={(-1,1)} is also a basis of U. In this case, dimU=1.
 Ex2. {(1,0), (0,1)} is a basis of R2 and {(1,-2), (2,0)} is also a basis of R 2. But {(1,0),
(-1,1), (1,1)} is not a basis of R2. We have dimR2=2.
 The basis {(1,0), (0,1)} is called standard basis of R2.
 Ex3. Which of the following is a basis of R3?
 {(1,0,1), (0,0,1)}
 {(2,1,0), (-1,0,1), (1,0,1), (0,-1,1)}
 {(0,1), (1,0)}
 None of the others
 Some important theorems
 Theorem 3. Let U≠0 be a subspace of Rn. Then:
 U has a basis and dimUn.
 Any independent set of U can be enlarged (by adding vectors) to a basis of U.
 If B spans U, then B can be cut down (by deleting vectors) to a basis of U.
Ex1. Let U=span{(1,1,1), (1,0,1), (1,-2,1)} be a subspace of R 3. This means, B=
{(1,1,1), (1,0,1), (1,-2,1)} spans U.
 U has a basis and dimU3,
  B can be cut down to a basis of U: {(1,0,1), (1,1,1)} is a basis of U, dimU=2
 construct a basis for U: {(1,0,1)} {(1,0,1), (1,1,1)}.
1 1 1 1 1 1 1 1 1
1 0 -2 0 -1 -3 0 1 3
1 1 1 0 0 0 0 0 0
 Theorem 4. Let U be a subspace of Rn and B={X1,X2,…,Xm}U, where dimU=m.
Then B is independent if and only if B spans U.
 Theorem 5. Let UV be subspaces of Rn. Then:
 dimU  dimV.
 If dimU=dimV, then U=V.
 Examples
Determine whether U is a subspace of R3.
U={[0 a b]T: a,b  R} 
U={[0 1 s]T:s  R}
U={[a b a+1] T:a,b  R}
U={[a b a2]T: a, b  R}

1 0 2  0 2 Find all m such that the set {(2,m,1),(1,0,1),(0,1,1)} is linearly

0 1 m 0 1 m independent.
1 1 1 0 1 -1 m≠-1 
m=-1 only
 0 2 m=0 only
0  m m≠0
None of the others
0 0 -1-m

A basis for the subspace U={[a b a-b]T: a,b R} is…

a. {[1 0 1]T, [0 1 -1]T} 

b. {[1 1 0]T}
c. {[1 0 1]T, [-1 0 -1]T, [0 1 -1]T}
d. None of the others.
 Exercises
The dimension of the subspace U=span{(-2, 0, 3),  không thể là b vì dimUdimR3=3
(1, 2, -1),(-2, 8, 5),(-1, 2, 2)} is…  kiểm tra bằng biến đổi sơ cấp
a. 2 
1 -2 -1 -2  -2 -1 -2  -2 -1 -2
b. 4
c. 3 2 8 2 0 0 12 4 4 0  1/3 1/3 }only 2 
d. 1 -1 5 2 3 0 3 1 1 0 0 0 0

Let u and v be vectors in R3 and w  span{u,v}. Then …

a. {u,v,w} is linearly dependent. 
b. {u,v,w} is linearly independent.
c. {u,v,w} is a basis of R3
d. the subspace is spanned by {u,v,w} has the dimension 3.
Let {u,v,w,z} be independent. Then …. is also independent.
a. {u,v+w,z}  1 1 0 0  1 0 0  1 0 0
b. {u,v,v-z-u,z}
c. {u+v,u-w,z, v+z+w} 1 0 0 1 0 -1 0 1 0  0 -1
d. {u,v,w,u-v+w} 0 -1 0 1 0 -1 0 1 0 0  1
0 0 1 1 0 0 1 1 0 0 0 0
 Exercises
 Let U=span{(1,-1,1), (0,2,1)}. Find all value(s) of m for which (3,-1,m)U.
 (3,-1,m) U (3,-1,m)= a(1,-1,1) + b(0,2,1) for some a,b. Solve for a,bm=4

 Find all values of m so that {(2,-1,3); (0,1,2); (-4,0; m)} spans R 3.

 Theorem 4. Let U be a subspace of Rn and B={X1,X2,…,Xm}U, where dimU=m.
Then B is independent if and only if B spans U.
 So, {(2,-1,3); (0,1,2); (3,1; m)} spans R3  it is linearly independent  m≠10
2 0 -4  -1 0  -1 0  -1 0  -1 0
-1 1 0 2 0 -4 0 2 -4 0  2 0  2
3 2 m 3 2 m 0 5 m 0 5 m 0 0 m-10

 Find a basis for the solution space to the homogeneous system

x - y + 2z=0
2x + y + z=0
Solution: z=t, y=t, x=-t
 -1 2 0  -1 2 0  -1 2 0 Solution space: U={(-t,t,t)| tR}
2 1 1 0 0 3 -3 0 0  -1 0 ={t(-1,1,1)| tR}=span{(-1,1,1)}
A basis for U: {(-1,1,1)}
dimU=1
 Definitions
 Dot product:
Suppose X= [x1 x2 … xn]T, Y= [y1 y2 … yn]T are vectors in
Rn . The dot product of two vectors X and Y, denoted
as X•Y, is a number defined by X•Y=x1y1+x2y2+…+xnyn
 Length:
The length of the vector X=[x1 x2 … xm]T (or (x1,x2,
…,xn)) is

X  X  X  x12  x2 2  ...  xn 2
 Unit vector: a vector with length 1
 Distance: d(X,Y)= X-Y
 Theorem
• Let X,Y, and Z denote vectros in Rn. Then:
 XY=YX
 X(Y+Z)=XY+XZ
 (aX)Y=a(XY)=X(aY)
 XX=X2
 X0, X=0  X=0
 aX=|a|.X
• Ex. Suppose that Rn=span{F1,F2,…,Fn}. If XFi=0 for each I,
where X is a vector in Rn. Show that X=0 (zero vector).
Rn=span{F1,F2,…,Fn } and XRn  X= k1F1+k2F2+…+knFn for some
kiRn.
We have X2=XX= X(k1F1+k2F2+…+knFn)= k1XF1+k2XF2 +
…+ knXFn=0 X=0  X=0
 Theorem

• If X, Y and Z are vectors in Rn, we have:

 d(X,Y)0
d(X,Y) = 0  X=Y
 d(X,Y)=d(Y,X)
 triangle inequality: d(X,Y)+d(Y,Z)d(X,Z)
• Corolary. X+YX+ Y
• Cauchy Inequality: XYX.Y
 Definitions
 Orthogonal set
A set {x1,x2,…,xm} is called orthogonal set if xi is not zero vector
and xi•xj=0 for all i≠j.
For example, {(1,-1);(1,1)} is an orthogonal set in R2
{(1,1,1);(-1,0,1);(0,1,0)} is not a orthogonal set but {(-1,0,1);
(0,1,0)} is a orthogonal set.
 Orthonormal set
A orthogonal set {xi} is called orthonormal set (hệ trực chuẩn)
is xi is unit vector for all i. For example, (1,0,0);(0,1,0)} is
orthonormal.
{(-3,0,4);(4,5,3)} is a orthogonal set, not a orthonormal set.
1 1 
However, the set  5  3, 0, 4  ;
5 2
 4,5,3  is orthonormal

 Examples
 The standard basis of Rn {E1,E2,…,En} is orthonormal
 If {F1,F2,…,Fk} is orthogonal then {a1F1,a2F2,…,akFk} is
also orthogonal for any nonzero scalar ai
 Every orthogonal set is a linearly independent set
 If u,v are unit orthogonal vectors then
(3u-5v)(4u+2v)=12uu+ 6uv-20vu-10vv =
12u2-10v2= 12-10=2
 Pythagoras’s Theorem
 If {F1,F2,…,Fk} is orthogonal then
2 2 2 2
F1  F2  ...  Fk  F1  F2  ...  Fk
 Expansion Theorem
 Let {F1,F2,…,Fk} be a orthogonal basis of a
subspace U and X is in U. Then
X  F1 X  F2 X  Fn
X 2
F1  2
F2  ...  2
Fk
F1 F2 Fn
5.4. Rank of a matrix
 Rank of a matrix
 If A is carried to row-echelon form then
rankA=number of leading 1’s
 If a is an mxn matrix then rankA≤min{n,m}
 rankA=rank(AT)
rowA and colA subspaces
 rowA=span{rows of matrix A}
 colA=span{columns of A}
 dim(rowA)=dim(colA)=rankA
 For example, find bases of colA and rowA if
 1 1 1 2 
3 2 0 5
 
A   2 3 3 4 
 
 1 1 1 3 
 0 1 1 2 
  1 1 1 2  1 1 1 2 
 3 2 0 5  0 1 3 1
   
A   2 3 3 4   0 1 1 0 
   
 1 1 1 3   0 0 0 1 
 0 1 1 2  0 1 1 2 
1 1 1 2  1 1 1 2 
0 1 3 1 0 1 3 1
   
 0 0 2 1   0 0 2 1 
   
 0 0 0 1   0 0 0 1 
0 0 3 3  0 0 0 0 

A basis of rowA is {r1,r2,r3,r4} and dim(rowA)=4

A basis of colA is {c1,c2,c3,c4} and dim(colA)=4
 Theorem
 An nxn matrix A is invertible if and only if rankA=n
 If an mxn matrix B has rank n then the n columns of B
is linearly independent
 If A is mxn matirx and m>n then the set of m rows of A
is not independent
For example, If A is an 3x5 matrix with rank 3 then 5 columns
are dependent and 3 rows are independent.
 Theorem
If an mxn matrix A has rank r then:
 The equation AX=0 has n-r basic solutions X1,X2,
…,Xn-r
 {X1,X2,…,Xn-r} is a basis of nullA={X:AX=0}
 dimnullA=n-r
 imA=colA and
 dimimA=dimcolA=rankA=r
Thanks