AMS210 Homework 4

2.1:10,17,23,32; 2.2:3,6,18,35; 2.3:5,6,8,20 DUE: Friday, Oct. 6, Beginning of Class

Solutions to Problems:


2.1:10)   
	AB = [ 3  -4 ]  [ 7  4 ]  =  [  1    12-4k ]
	     [-5   1 ]  [ 5  k ]     [ -30   -20+k ]

	BA = [ 7   4 ]  [ 3 -4 ]  =  [  1     -24  ]
	     [ 5   k ]  [-5  1 ]     [15-5k  -20+k ]

Comparing entries, we get AB = BA only if 

	12-4k = -24  and  15-5k = -30,

which is satisfied for k = 9.
	

2.1:17)  First note that A is 2x2 and AB is 2x3, so B must be 2x3.

B = [ b_1  b_2  b_3 ].  We are asked to find the column vectors b_1 and b_2.
By definition of matrix multiplication, we have

	A b_1 = [-1 ]	so b_1 is the solution of the system whose 
		[ 6 ]    augmented matrix is below

   [ 1  -2  -1 ]       [ 1  -2  -1 ]       [ 1   0   7 ]   so b_1 = [ 7 ]
   [-2   5   6 ]   ~   [ 0   1   4 ]   ~   [ 0   1   4 ]	    [ 4 ]

Likewise, A b_2 = [ 2 ] gives the system
		  [-9 ]

   [ 1  -2   2 ]       [ 1  -2   2 ]       [ 1   0  -8 ]   so b_2 = [-8 ]
   [-2   5  -9 ]   ~   [ 0   1  -5 ]   ~   [ 0   1  -5 ]	    [-5 ]


2.1:23) We are given CA = I (I is nxn identity matrix).  Note if A is mxn, 
then C must be nxm, and below x must be a nx1 vector (that is, x is in R^n).
Show if  Ax  =  0  (the mx1 zero vector), then x = 0.

	    Ax  =  0   (Multiply both sides on left by C)
  -->     C(Ax) = C0 = 0  (here 0 is the nx1 zero vector).
	Associativity of matrix multiplication gives C(Ax) = (CA)x, so
	 (CA)x  =  0
  -->      I x  =  0 
  -->        x  =  0.


2.1:32)  (ABx)^T  =  x^T B^T A^T  (where ^T means transpose).
Apply Theorem 3(d) twice to see this.


2.2:3)  A = [ a   b ]  =  [ 3  -7 ]
	    [ c   d ]     [-6  13 ]
 
First note ad - bc = 3(13) - (-7)(-6) = -3 does not equal 0, so A is invertible

By Thm.4 in text A inverse = A^-1 = (-1/3) [ 13   7 ]
			                   [  6   3 ]


2.2:6)  The system is Ax = b = [-4 ] , where A is same matrix as in 2.2:3.
			       [ 1 ]

A is invertible by the previous problem so we have a unique solution

	x = (A^-1)b = (-1/3) [ 13   7 ] [-4 ] = (-1/3) [-45 ] = [ 15 ]
		             [  6   3 ] [ 1 ] 	       [-21 ]   [  7 ]


2.2:18) A = PB(P^-1) . Note B must be square (same size as P), so A is also
square and of same size.
Multiply both sides of equality on left by P^-1 (inverse) and on right by P:
	       A = PBP^-1
  -->   (P^-1)AP = (P^-1)(PBP^-1)P
		 = ((P^-1)P) B ((P^-1)P)
		 =  IBI 
		 =   B


2.2:35)  We could perform row operations on [ A  I ] to get [ I  (A^-1) ],
but then we would be finding all 3 columns of (A^-1).
   Since we only need the third column of A^-1, denote this column vector by
x = [ x_1 ]   and more efficiently solve the system   Ax = [ 0 ]  (the third
    [ x_2 ]						   [ 0 ]   column of
    [ x_3 ]						   [ 1 ]   matrix I)

	[ -1  -5  -7   0 ]             [ -1  -5  -7   0 ]
	[  2   5   6   0 ]      ~      [  0  -5  -8   0 ]    ~
	[  1   3   4   1 ]             [  0  -2  -3   1 ]

	[  1   5   7     0  ]             [  1   0  -1/2   5/2 ]
	[  0   1  3/2  -1/2 ]      ~      [  0   1   3/2  -1/2 ]    ~
	[  0  -5  -8     0  ]             [  0   0  -1/2  -5/2 ]

	[  1   0   0   5 ]   Thus the 3rd column of A^-1 is  [ 5 ]
	[  0   1   0  -8 ]				     [-8 ]
	[  0   0   1   5 ]				     [ 5 ]


2.3:5)
 [  5  -9   3 ]		[ 1   0   3 ]	     [ 1   0   3 ]       [ 1   0   3 ]
 [  0   3   4 ]    ~    [ 0   3   4 ]	~    [ 0   3   4 ]   ~   [ 0   3   4 ]
 [  1   0   3 ]		[ 5  -9   3 ]	     [ 0  -9  -12]	 [ 0   0   0 ]

Only 2 pivot positions so the matrix is not invertible.


2.3:6) Easiest to row reduce to an echelon form.

	[  1   4   2 ]         [  1   4   2 ]	    [ 1   4   2 ]
	[  2   7   3 ]    ~    [  0  -1  -1 ]   ~   [ 0   1   1 ]
	[  1   7   5 ]         [  0   3   3 ]	    [ 0   0   0 ]

Only 2 pivot positions so not invertible.


2.3:8)  Yes, invertible since 4 pivot positions for the 4x4 square matrix.


2.3:20) No, Q can not be invertible.  Apply the Invertible Matrix Thm. 
 Statement (h) false implies statement (a) is false.