Linear Algebra
The Date
1. Find all 4 7 real, row-reduced echelon matrices A = (ai;j ) of rank 4 with a1;3 = a3;5 = 2 so
that the linear system of equations with the augmented matrix A has the solutions [1; 0; 1; 3; 1; 0] and
[1; 1; 1; 1; 0; 0]. (Explain your work)
2 3
: : 2 : : : :
6: : : : : : :7
ANSWER We have A = 6 4: : :
7 : Since rank(A) = 4 the last row of A is a nonzero
: 2 : :5
: : : : : : :
row. If the …rst nonzero entry of the last row appears in column k then k 4 and 3 6= k 6= 5:
Thus we get k = 6 or k = 7: If k = 7 then the last row is (0; 0; 0; 0; 0; 0; 1) which implies that the
system 2 with the augmented matrix 3 A has no solution, which is not the case.m Thus k = 6: So we get
: : 2 : : 0 :
6: : : : : 0 :7
A=6 4: :
7 : Similarly if the …rst nonzero entry of the third row appears in the
: : 2 0 :5
0 0 0 0 0 1 :
column l then 3 l 4 and l 6= 3 That is l = 4: And we determine that the f¬rst nonzero entry of the
second row appears 2 in the second column and 3 the …rst nonzero entry of the f¬rst row appears in the
1 0 2 0 : 0 :
60 1 : 0 : 0 :7
…rst column: A = 6 40 0 0 1
7:
2 0 :5
0 0 0 0 0 1 :
That the vectors [1; 0; 1; 3; 1; 0] and [1; 1; 1; 1; 0; 0] are solutions of the linear system the augmented
matrix A give that
1 2 + a1 ;5 = a1;7
a2;3 + a_ 2;5 = a2;7
3 2 = a3;7
0 = a4;7
and
1 = a1;7
1 + a2;3 = a2;7
1 = a3;7
2 3
1 0 2 0 0 0 1
60 1 t 0 1 0 1 + t7
We get A = 6
40 0
7 ; t 2 R:
0 1 2 0 1 5
0 0 0 0 0 1 0
2. Given a linear system with the following augmented matrix over R depending on two parameters t and
u: 2 3
t t t+1 3
4 t t 2 t+1 u 5:
t t 5t 1 u 1
For which values of the parameters does this system have
(a) a unique solution,
(b) more than one solution,
(c) no solutions.
ANSWER
2 3 2 3
t t t+1 3 t t t+1 3
We reduce the given matrix to echelon form 4 t t 2 t+1 u 5 ! 40 2t 2 2t + 2 u + 35
t t 5t 1 u 1 0 0 4t u+2
and see that if 0 6= t 6= 1 the rank of the augmented matrix is equal to the rank of the coe¢ -
cient matrix which is equal to 3 and the number of variables, So2that in this case the 3 system has a
1 1 2 3
unique solution . If t = 1 then the matrix is row equivalent to 4 0 0 2 u + 35 and hence to
0 0 4 u+2
2 3
1 1 2 3
4 0 0 2 u + 3 5 So we get in this case that the system has no solution if 3u + 8 6= 0 and in…nitely
0 0 0 3u + 8
many2 solutions if 3u +38 = 0. And …nally we have t = 0: In this case the given matrix is row equivalent
0 0 1 3
to 40 2 2 u + 35 and the system has no solutions if u 6= 2 and in…nitely many solutions if
0 0 0 u+2
u = 2: So the system
(a) a unique solution for arbitrary u if 2 6= t 6= 0;
8
(b) more than one solution, if (t = 0 and u = 2) or (t = 1 and u = ):
3
8
(c) no solutions if (t = 0 and u 6= 2) or (t = 1 and u 6= ):
3
3. Let V = R3 and W be the subspace of V generated by = (1; 3; 5); = (2; 5; 9); = ( 1; 2; 0):
(a) Which of the vectors (a; b; c) 2 V belongs to W ?
(b) If (0; 1; 1) is in W then write it as a linear combination of the vectors ; ; :
ANSWER
a) (a; b; c) 2 V belongs to W , 9 x; y; z 2 R such that (a; b; c) = x + y + z = (x + 2y z; 3x +
5y + 2z; 5x 9y) , 9 x; y; z 2 R such that
x + 2y z = a
3x + 5y + 2z = b
5x 9y = c
,the system with the augmented matrix
2 3
1 2 1 a
43 5 2 b5
5 9 0 c
is consistent. And to see this we row reduce this last matrix to its row reduced echelon form
2 3
1 2 1 a
43 5 2 b5 :
5 9 0 c
2 3 2 3 2 3
1 2 1 a 1 2 1 a 1 2 1 a
43 5 2 b 5 ! 40 1 5 b 3a5 ! 40 1 5 b 3a 5
5 9 0 c 0 1 5 c + 5a 0 0 0 c + b + 2a
We deduce that
(a; b; c) 2 V belongs to W , 2a + b + c = 0:
b) Clearly (0; 1; 1) 2 W: As
2 3 2 3 2 3
1 2 1 0 1 2 1 0 1 0 9 2
43 5 2 1 5 ! 40 1 5 15 ! 40 1 5 15
5 9 0 1 0 0 0 0 0 0 0 0
we get that (0; 1; 1) = 2 +0 :
4. Let U and W be the subspaces of V = R3 given by U = h(1; 0; 1); (0; 1; 1)i and W = f(x; y; z) :
2x 3y + 4z = 0g: Find U \ W and U + W: Is the sum U + W a direct sum?
ANSWER
f(1; 0; 1); (0; 1; 1)g is given as a generating set for U: Since we can write W = f( 3y 2 4z ; y; z) : y; zg we
see that f(3; 2; 0); (2; 0; 1)g is a generating set for W . Therefore we get that f(1; 0; 1); (0; 1; 1)g [
f(3; 2; 0); (2; 0; 1)g is a generating set for U +W = hU [ W i and hence U +W = h(1; 0; 1); (0; 1; 1); (3; 2; 0); (2; 0; 1)i
Now we can describe U +W as the set of all (a; b; c) 2 V which can be written as a linear combination of
the vectors in the generating set f(1; 0; 1); (0; 1; 1); (3; 2; 0); (2; 0; 1)g So (a; b; c) 2 U +W if and only
if there exist t; x; y; z 2 R such that (a;2b; c) = t(1; 0; 1) + 3
x(0; 1; 1) + y(3; 2; 0)2
+ z(2; 0; 1); if and 3
only
1 0 3 2 a 1 0 3 2 a
if the system withe augmented matrix 4 0 1 2 0 b 5 is consistent . As 4 0 1 2 0 b 5 !
2 3 1 1 0 1 c 1 1 0 1 c
1 0 3 2 a
40 1 2 0 b 5 we see that the system has a unique solution for any a,b,c. This means that
0 0 1 1 c+a b
U +W =V.
On the other hand (a; b; c) 2 U \ W if and only if 2a 3b + 4c = 0 and there exist x; y 2 R such that
(a; b; c) = x(1; 0; 1)+y(0; 1; 1): Thus we get x = a; y = b; x+y = c and hence 2x 3y +4( x+y) = 0
Thus 2x + y = 0 and
U \ W = fx(1; 2; ; 1) : x 2 Rg
5. (Bonus) Suppose that A = (ak;l ) 2 F k m
and B 2 F m n
:
(a) Show that the row space of AB is a subspace of the row space of B and deduce that rank(AB)
rank(B):
(b) Suppose that m < n = k: Show that it is not possible to have AB = In and …nd an example such
that BA = Im .:
ANSWER
a) Row space of B is the subspace of F 1 n consisting of all linear combinations of rows of B: The
k-th row of the matrix AB is the k-th row of A multiplied by B and hence
(AB)(k) = A(k) B = ak;1 B(1) + ak ;2 B(2) + + ak;m B(m)
is clearly a vector in the row space of B: So the row space of AB is a subspace of the row space
The set of nonzero vectors of the row reduced echelon form of a matrix X is a basis for the row
space of X that is the dimension of the row space of X is equal to the row rank of X: Therefore
rank(AB) = dim(Row space of AB) dim(Row space of B) = rank(B) that is
rank(AB) rank(B):
b) If AB = In we have n = rank(In ) = rank(AB) rank(B) m =number of rows: This is not
possible as m < n: So AB 6= In :
2
But BA = In is possible.The smallest example m = 1 < 2 = n :A = ;B = 1 1 then
1
2 2
BA = [1] = I1 and AB = 6= I2 :
1 1
