2 # 1
(a) x is orthogonal to W x W w W x w w W xw 0 . (b) W the orthogonal complement of W the set of all vectors in R n that are orthogonal to W. x Rn : x W x Rn : w W x w x Rn : w W x w 0 . Theorem (5.9) Let W be a subspace of R n . Then: (a) W is also a subspace of R n . (b) W W , 0 . (c) if W span w 1 , w 2 , , w k then W v R n : v w i 0 for i 1, 2, 1. Proof (a) . W , is not empty: w W and y w w W W 0w 0 0 W W y w 0 x y w x w y w 00 0 W. W W . W , is closed under addition: x W x w 0 and w
,k .
x y
$ 5.2 # 2 k R and x w w W
W xw 0 W kx w k x w k 0 0 W.
kx
(b) . 0
W W and
.W
0 W,
(c) W span w 1 , w 2 , x
Rn :
i v wi 0
, wk , wk ,k
R n : v w i 0 for i 1, 2,
,k
(a) RS A row A the row space of A. (b) CS A col A the column space of A. (c) NS A null A the null space of A. Theorem (5.10) Let A be an m n matrix. Then: (a) RS A , NS A . (b) CS A , NS A T . Proof (a) x RS A , x , RS A x , every row of A Ax 0 x (b) CS A
,
NS A .
,
RS A
NS A T , by part (a).
1 0
2 1
a. Determine the dependency equation(s) of the columns in A. Solution The linear dependence or independence in RREF A correspondingly determines the linear dependence or independence in A. All the columns in RREF A containing leading 1s are linearly independent, that is, c 1 , c 2 and c 4 in RREF A are linearly independent. The columns c 3 and c 5 are linearly dependent on c 1 , c 2 and c 4 as follows: c 3 1c 1 2c 2 0c 4 c 5 1 c 1 3 c 2 4 c 4. Therefore the columns dependency equations in A are: col 3 A col 1 A 2 col 2 A col 5 A Check: col 1 A 3 col 2 A 4 col 4 A .
col 1 A 3 col 2 A 4 col 4 A 1 2 3 4 6 1 1 3 b. Determine the dependency equation(s) of the rows in A. Solution c 4 c 1 6 c 2 3 c 3 in RREF A T col 4 A T col 1 A T 6 col 2 A T 3 col 3 A T row 4 A row 1 A 6 row 2 A 3 row 3 A Check: col 5 A . 3 1 1 2 1 4 1 1 2 1
row 4 A .
$ 5.2 # 7 c. Determine a basis in RREF A for RS A . Solution RS A RS RREF A for RS A in RREF A . d. Determine a basis in A for RS A . Solution col 1, col 2, col 3 are linearly independent in RREF A T col 1 A T , col 2 A T , col 3 A T is a basis for CS A T in A T row 1 A , row 2 A , row 3 A is a basis for RS A in A since RS A CS A T . e. Determine a basis in A for CS A . Solution col 1 A , col 2 A , col 4 A row 1, row 2, row 3 is a basis
f. Determine a basis in RREF A T for CS A . Solution row 1, row 2, row 3 is a basis for CS A in RREF A T .
1 span 6 3 1
NS A .
Solution It is enough to show that each basis vector of RS A is orthogonal to each basis vector of NS A : 1 1 0 1 0 0 1 2 0 0 0 0 1 1 3 4 2 1 0 0 or 1 1 2 3 1 2 3 1 1 1 2 6 1 1 2 1 0 0 j. CS A
,
1 3 0 4 1 0 0 0 0 0 0
1 3 0 4 1 0 0 0 0 0 0
1 0
NS A T .
Solution It is enough to show that each basis vector of CS A is orthogonal to each basis vector of NS A T :
$ 5.2 # 9 1 1 1 or 1 0 0 1 0 1 0 6 0 0 1 3 1 6 3 1 0 0 0 2 1 1 3 4 2 1 2 1 1 6 3 1 0 0 0
$ 5.2 # 10 Example 2. Let W be the subspace of R 5 spanned by 1 3 w1 5 0 5 Find a basis for W . Example Solution Let 1 3 A w1 w2 w3 5 0 5 Now, W CS A RS A T W CS A Therefore, 1 AT 1 0 so that 3 5 1 2 1 4 0 5 1 0 0 3 4 0 1 0 1 3 0 0 1 0 2 RREF A T 2 3 1 5
,
1 1 , w2 2 2 3 w3
0 1 4 1 5
1 1 2 2 3
0 1 4 1 5
RS A T
NS A T NS RREF A T
$ 5.2 # 11 3x 4 x4 W x4 x5 4x 5 3x 5 2x 5 span 3 1 0 1 0 , 4 3 2 0 1
$ 5.2 # 12 Orhtogonal projections Definition Let v and u 0 in n . Then: the component v parallel to u the projection of v onto u proj u v v u u uu and the component of v orthogonal to u perp u v Remark Since v proj u v perp u v it follows that perp u v v proj u v . W and If W span u , then w proj u v , w , proj u v perp u v W , . Therefore, there is a decompostion of v into the sum v w w , such that w W and w , W , . Definition Let S u 1 , u 2 , , u k be an orthogonal basis for the subspace W in n . For any v in n , the component of v in W proj W v v u1 u v u2 u 1 2 u1 u1 u2 u2 proj u 1 v proj u 2 v proj u k v and v uk uk uk
uk
$ 5.2 # 13 the component of v orthogonal to W perp W v proj W , v proj W v v proj W v . x Example (1.) P y z 3 v 1 . Find the orthogonal prjection of v onto P and the 2 component of v orthogonal to P. Solution (1.) 1 P, n 1 2 Therefore, :x y 2z 0 is a plane in 3 and let
,
$ 5.2 # 14 the component of v orthogonal to P proj P , v perp P v proj n v 3 1 2 1 1 2 and the component of v in P proj P v v 3 1 2 Solution (2.) y P y z 2z span u1 1 1 0 , u2 1 1 1 , perp P v 4 3 4 3 8 3 5 3 1 3 2 3 5 1 3 1 2 1 1 2 1 1 2 1 1 2 4 3 1 1 2 4 3 4 3 8 3 v n nn n
hence, P has an orthogonal basis Look at Example 5.3, page 367-368 . Therefore,
$ 5.2 # 15 the component of v in P proj P v proj u 1 v proj u 2 v v u1 u v u2 u 1 2 u1 u1 u2 u2 3 1 2 1 1 0 1 1 0 and the component of v orthogonal to P proj P , v perp P v v 3 1 2 Check: 5 3 1 3 2 3 proj P v 4 3 4 3 8 3 1 4 3 1 2 . 2 3 1 1 0 1 1 0 1 1 1 5 3 1 3 2 3 1 1 0 3 1 2 1 1 1 5 1 3 1 2 1 1 1 1 1 1 1 1 1
Theorem (5.11) The Orthogonal Decomposition Theorem Let W be a subspace of n and let v n . Then there are unique vectors w in W and w , in W , such that v w w ,. Proof (a) . Show that the decomposition exists, that is, w W w , W , v w w , : Let S u 1 , u 2 , , u k be an orthogonal basis for the subspace W in n and let v n . Let v u1 u v u2 u v uk w proj W v 1 2 u1 u1 u2 u2 uk uk
k
uk
i1
v ui ui ui
ui
and let w , perp W v v proj W v . v proj W v v. Then w w , proj W v perp W v proj W v . w W: w proj W v v u1 u1 u1 . w, W, : i,
u1
v u2 u2 u2
u2
v uk uk uk
uk
span S W.
$ 5.2 # 17 w, u i v v proj W v ui uk ui
v ui
v u1 u v uk 1 u1 u1 uk uk v ui ui ui , ui ui 0 ui
2
ui uj
if i uj
2
if i j
W,
Theorem 5. 9 (c). :
(b) Show that the uniqueness of the decomposition exists, that is, v x y x w and y w
W and b w w a b a b
w w, a b w , 0 , since W