MA 106: Spring 2014: Tutorial Sheet 2

 
1 3 −2
 
1. Find the inverse of the matrix 
 2 5  using elementary row operations.
−7 
0 1 −4
Solution:
   
1 3 -2 1 0 0 1 3 -2 1 0 0
   
[A|I] = 
 2 5 -7  R2 − 2R1 −→  0 -1 -3 -2 1 0 
0 0 1  
0 1 -4 0 0 1 0 1 -4 0 0 1
 
1 3 -2 1 0 0
 
R3 + R2 −→ 
 0 -1 -3 -2  − 1 × R2 & − 1/7 × R3 −→
1 0 
0 0 -7 -2 1 1
   
1 3 -2 1 0 0 ) 1 3 0 11/7 -2/7 -2/7
  R2 − 3R3  
 0 1
 3 2 -1 0 
 R + 2R −→  0 1 0 8/7 -4/7 3/7 

1 3
0 0 1 2/7 -1/7 -1/7 0 0 1 2/7 -1/7 -1/7
   
1 0 0 -13/7 10/7 -11/7 -13/7 10/7 -11/7
  −1 −1
 
R1 −3R2 −→  0 1 0
 8/7 -4/7 3/7  = [I|A ]. Thus A =  8/7 -4/7
 
.
3/7 
0 0 1 2/7 -1/7 -1/7 2/7 -1/7 -1/7
 
1 1 1 0
 
 -3 -17 1 2 
2. Let A =  4 -24
 . Assuming that A is invertible, find the last row and the last
 8 -5 

0 -7 2 2
−1
column of A .
Solution: The last column of A−1 is the unique solution of the system Ax = [0, 0, 0, 1]t , which is
[11/4, −1/2, −9/14, 1]t .
The last row of A−1 is the last column of (A−1 )t = (At )−1 , and it is the unique solution of the
system At y = [0, 0, 0, 1]t , which is [−3/2, −1/2, 0, 1]t .

3. Let A be an m × n matrix, Eij be an n × n matrix unit, and E be an n × n elementary matrix.

Prove the following. (i) Post-multiplying A by Eij replaces the jth column of A by the ith column
by A, and all other columns of A by 0. (ii) If E = I + αEij is of type I, then post-multiplying A by
E adds α times the ith column of A to its jth column. (iii) If E is of type II, then post-multiplying
A by E interchanges the ith and the jth columns of A. (iv) If E = I + (α − 1)Ejj is of type III,
then post-multiplying A by E multiplies the jth column of A by α.
Solution: Observe that (Eij )t = Eji and that the transpose of an elementary matrix is an ele-
mentary matrix of the same type. The result follows from the row case dealt with in the class and
the identity (AE)t = E t At .

4. Prove that
 
a11 a12 a13
 
det 
 a21 a22 =
a23 
a31 a32 a33

a11 a22 a33 − a11 a32 a23 − a21 a12 a33 + a21 a32 a13 + a31 a12 a23 − a31 a22 a13 .

1
 Solution: Denote the matrix by A. From the class work and the lecture notes, we know that
X
det(A) = ah(1)1 ah(2)2 ah(3)3 det(eh(1 , eh(2) , eh(3) ),
h

where the sum is over all bijections h : {1, 2, 3} −→ {1, 2, 3}. There are 3! = 6 such functions
h. The corresponding determinant det(eh(1) , eh(2) , eh(3) ) is 1 or −1 (depending on whether the
permutation (h(1), h(2), h(3)) of (1, 2, 3) is even or odd). Check that this gives the usual formula
for the determinant of a 3 × 3 matrix.

5. Prove the
 following : 
0 0 0 ... 0 0 1
 
 0 0 0 ... 0 1 0 
 
.
 
 
 = (−1)n(n−1)/2 ,
 
(a) det 
 . 
 
 . 
 
 0 1 0 ... 0 0 0 
 
1 0 0 ... 0 0 0
 
1 2 3 ... n − 1 n
 

 2 2 3 ... n − 1 n  
3 3 3 ... n − 1 n 
 
n+1

(b) det 
 .. .. ..  = (−1)
.. ..  n,
 . . . . . 
 
 n−1 n−1 n−1 ... n − 1 n 
 
n n n ... n n
 
2 −1
 
 −1 2 −1 
 
−1 2 −1
 
 
 
(c) det 
 −1 ·  = n + 1.

·
 
 
 

 2 −1 

−1 2

Solution: (a) Call the matrix An . Prove by induction on n, the cases n = 1, 2 being clear.
Observe that A = [en , en−1 , . . . , e1 ]. By interchanging e1 successively with e2 , e3 , . . . , en , we obtain
the matrix [e1 , en , en−1 , . . . , e2 ] by n − 1 interchanges of columns. Now, expanding by the first
column, and using induction, we obtain det(An ) = (−1)n−1 (−1)(n−1)(n−2)/2 = (−1)n(n−1)/2 .
(b) Call the given matrix An . Clearly, det A1 = 1. Let n ≥ 2. Perform row operations on An in
the following order: Rn → Rn + (−1)Rn−1 , Rn−1 → Rn−1 + (−1)Rn−2 , . . ., R2 → R2 + (−1)R1
to obtain the row equivalent matrix
 
1 2 3 ... n−2 n−1 n
 
 1 0 0 ... 0 0 0 
 
 1 1 0 ... 0 0 0 
 
Bn =  .. .. .. .. .. .
.. 
 . . . ... . . . 
 
 1 1 1 ... 1 0 0 
 
1 1 1 ... 1 1 0

2
 Expand Bn using the last column to obtain
 
1 0 0 ... 0
 
 1 1 0 ... 0 
det An = det Bn = (−1)n+1 n det  . . .  = (−1)n+1 n,
 
 .. .. .. ..
 ... . 

1 1 1 ... 1
as the determinant of an upper triangular matrix is the product of its diagonal elements.
 
2 −1
 
 −1 2 −1 
 
−1 2 −1
 
 
 
(c) Let An = 
 −1 · .

·
 
 
 

 2 −1  
−1 2
Clearly, det A1 = 2. Let n ≥ 2 and assume that det Ak = k + 1 for k = n − 2 and n − 1. Expand
An by the first row to obtain
   
2 −1 −1 −1
 −1 2 −1 2 −1
   
  
   

 −1 2 −1 


 −1 2 −1 

   
 −1 ·  
 − (−1)  −1 2 −1 
det An = 2 det  .

 · 


 −1 · 

   

 2 −1 


 · 

−1 2  2 −1 
   
 
1 2 −1 2
The first (n − 1) × (n − 1) determinant on the right is An−1 . Expand the second (n − 1) × (n − 1)
determinant on the right by the first column. Using det An−1 = n and det An−2 = n − 1, we obtain
det An = 2n + (−1)(n − 1) = 2n − n + 1 = n + 1.
 
1 1 1
 
6. (a) Prove that det  a b c  = (b − a)(c − a)(c − b).
 
a2 b2 c2
(b) Prove an analogous formula for n × n matrices. (Vandermonde determinant)
   
1 1 1 1 1 1
   
Solution: (a) A =  a b c  R3 → R3 + (−a)R2 →  a b c 
 

2 2 2
a b c 0 b(b − a) c(c − a)
 
1 1 1
 
R2 → R2 + (−a)R1 →  0 b−a c−a  
0 b(b − a) c(c − a)
Expanding by the first column, we obtain
" #
b−a c−a
det A = det = (b − a)(c − a)(c − b).
b(b − a) c(c − a)

3
 (b) Let x1 , x2 , . . . , xn ∈ R and
 
1 1 ... 1
x1 x2 ... xn 
 

 
2
A(x1 , x2 , . . . , xn ) = 
 x1 x22 ... x2n  .
 .. .. .. 
. . ... . 
 

xn−1
1 xn−1
2 ... xn−1
n
Y
We prove that det A(x1 , x2 , . . . , xn ) = (xj − xi ).
1≤i<j≤n

We may assume that x1 , . . . , xn are distinct as otherwise both sides are zero.
 
1 1 ... 1 1
x x2 . . . xn−1 xn 
 

 
2
Let p(x) = det 
 x x22 . . . x2n−1 x2n  .
 .. .. .. .. 
. . ... . . 
 

xn−1 xn−1
2 ... xn−1
n−1 xn−1
n

Then p(x) is a polynomial in x of degree ≤ n − 1 and x2 , . . . , xn are its roots, because when
x = xi , 2 ≤ i ≤ n, the determinant will have two identical columns, and so p(xi ) = 0. Thus
p(x) = α(x − x1 )(x − x2 ) . . . (x − xn ) for some α ∈ R. Note that α is the coefficient of xn−1 on the
R.H.S, whereas by expanding the determinant p(x) by the first column, we see that the coefficient
of xn−1 on the LHS is (−1)n+1 det A(x2 , x3 , . . . , xn ). Hence

p(x) = (−1)n+1 det(A(x2 , . . . , xn ))(x − x2 ) . . . (x − xn ).

By induction, we see that

 Y 
p(x) = (−1)n+1 (xj − xi ) (x − x2 ) . . . (x − xn ).
2≤i<j≤n

Since det(A(x1 , . . . , xn )) = p(x1 ), the proof is complete.

7. Examine whether the following sets of vectors constitute a vector space.

(a) The set of all real functions of the form a cos x + b sin x + c, where a, b, c ∈ R.
(b) Homogeneous polynomials of degree 3 in two variables together with the zero polynomial.
(c) Let n ≥ 2. The set of all n×n real matrices [aij ] which are: (i) diagonal, (ii) upper triangular,
(iii) trace-zero, (iv) symmetric, (v) skew-symmetric (vi) invertible, (vii) noninvertible.
(d) The set of all real polynomials of degree 5 together with the zero polynomial.
(e) The set of all real polynomials of degree ≤ 5 with p(0) = p(1).
(f) The real functions of the form (ax + b)ex , a, b ∈ R.

Solution: (a) Yes. This is the linear span of {cos x, sin x, 1}.
P3
(b) Yes. A typical term is k=0 ak xk y 3−k with a0 , a1 , a2 , a3 ∈ R.
(c) (i) Yes. (ii) Yes. (iii) Yes. (iv) Yes. (v) Yes. (vi) No. In and −In are in the set but their sum
is not. (vii) No. A = [e1 , · · · , en−1 , 0] and B = [0, . . . , 0, en ] are in the set, but A + B is not.

4
 (d) No. x5 + x, −x5 + x are in the set but not their sum.
(e) Yes. The condition is equivalent to saying that sum all the coefficients of the polynomial except
the constant coefficient is equal to zero.
(f) Yes. The subspace is spanned by {ex , xex }.

8. Let Pn denote the vector space consisting of all real-valued polynomial functions of degree ≤ n,
where n is fixed. Let S be a subset of all polynomial functions p in Pn satisfying the following
conditions. Check whether S is a subspace, and if so, find a spanning set.
00
(i) p(0) = 0, (ii) p is an odd function, (iii) p0 (0) = p (0) = 0.
Solution: (i) Yes. This subspace is spanned by {x, x2 , x3 , . . . xn }.
(ii) Yes. Recall that a function p is odd means p(−x) = −p(x) for all x ∈ R. By comparing
coefficients on either side we see that all even degree terms vanish. This subspace is then spanned
by {x, x3 , x5 , . . . , xn } if n is odd, and by {x, x3 , x5 , . . . , xn−1 } if n is even.
(iii) Yes. The given condition is equivalent to saying that the terms of degrees 1 and 2 are missing.
This subspace is spanned by {1, x3 , x4 , . . . , xn }.

9. Examine whether the following sets are linearly independent.

2×1 1×3
(i) {[x1 , x2 ]t , [y1 , y2 ]t } ⊂ R , where x1 y2 − x2 y1 6= 0, (ii) {[1, −1, 0], [3, −5, 2], [1, −2, 1]} ⊂ R ,
4×1
(iii) {[1, 0, 0, 0]t , [1, 1, 0, 0]t , [1, 1, 1, 0]t , [1, 1, 1, 1]t } ⊂ R .
Solution: (i) Yes. Let α, β ∈ R. αx1 + βy1 = 0 and αx2 + βy2 = 0 imply α = 0 = β, since
x1 y2 − x2 y1 6= 0. (ii) No. [1, −1, 0] − [3, −5, 2] + 2[1, −2, 1] = [0, 0, 0]. (iii) Yes. α1 [1, 0, 0, 0]t +
α2 [1, 1, 0, 0]t + α3 [1, 1, 1, 0]t + α4 [1, 1, 1, 1]t = [0, 0, 0, 0]t implies α1 = α2 = α3 = α4 = 0.