UPSC Civil Services Main 2005 - Mathematics

Linear Algebra
Sunder Lal
Retired Professor of Mathematics
Panjab University
Chandigarh

September 13, 2007

Question 1(a) Find the values of k for which the vectors (1, 1, 1, 1), (1, 3, 2, k), (2, 2k
2, k 2, 3k 1) and (3, k 2, 3, 2k + 1) are linearly independent in R4 .

Solution. The given vectors are linearly independent if the matrix

1 1 1 1
1
3 2 k
2 2k 2 k 2 3k 1
3 k+2 3 2k + 1

is non-singular.
Now

1 1 1 1 1 0 0 0
2 3 k 1
1 3 2 k 1 2 3 k 1
2 2k 2 k 2 3k 1 = 2 2k 4 k 4 3k 3 = 2k 4 k 4 3k 3

k1 6 2k 2
3 k + 2 3 2k + 1 3 k 1 6 2k 2

2
3 k 1
= 2k 4 k 4 3k 3 = (k 5)[9k + 9 + (k 1)(k + 4)] 6= 0
k5 0 0

Clearly (k 5)[9k + 9 + (k 1)(k + 4)] = 0 k = 1, 5. Thus the vectors are linearly

independent when k 6= 1, 5.

1
Question 1(b) Let V be the vector space of polynomials in x of degree n over R. Prove
that the set {1, x, x2 , . . . , xn } is a basis for V. Extend this so that it becomes a basis for the
set of all polynomials in x.

Solution. {1, x, x2 , . . . , xn } are linearly independent over R If a0 + a1 x + . . . + an xn = 0

where ai R, 0 i n, then we must have ai = 0 for every i because the non-zero
polynomial a0 +a1 x+. . .+an xn can have at most n roots in R whereas a0 +a1 x+. . .+an xn = 0
for every x R.
Every polynomial in x of degree n is clearly a linear combination of 1, x, x2 , . . . , xn

S
with coefficients from R. Thus {1, x, x2 , . . . , xn } is a basis for V.
We shall show that = {1, x, x2 , . . . , xn , xn+1 , . . .} is a basis for the space of all polyno-
mials.
S
(i) Linear Independence: Let {xi1 , . . . , xir } be a finite subset of . Let n = max{i1 , . . . , ir },

S
then {xi1 , . . . , xir } being a subset of the linearly independent set {1, x, x2 , . . . , xn } is linearly
independent, which shows the linear independence of .

S S
(ii) Let f be any polynomial. If degree of f is m, then f is a linear combination of
{1, x, x2 , . . . , xm }, which is a subset of . Thus is a basis of W, the space of all
polynomials over R.
3
Question 2(a) Let T bea linear transformation on R whose matrix relative to the standard
3
2 1 1
basis of R is 1 2 2 . Find the matrix of T relative to the basis

3 3 4
B
= {(1, 1, 1), (1, 1, 0), (0, 1, 1)}.

Solution. Let the vectors

of the given
basis be v1 , v2 , v3 . (T(v1 ), T(v2 ), T(v3 )) =
2 1 1 1 1 0 2 3 0
1 2 2 1 1 1 = 5 3 4.
3 3 4 1 0 1 10 6 7
If (a, b, c) = v1 + v2 + v3 , then + = a, + + = b, + = c therefore
= a b + c, = b c, = b a. Consequently
T(v1 ) = 7v1 5v2 + 3v3
T(v1 ) = 6v1 3v2 + 0v3
T(v1 ) = 3v1 3v2 + 4v3

B

7 6 3
This shows that the matrix of T with respect to given basis is 5 3 3
3 0 4

Question 2(b) If S is a skew-Hermitian matrix, then show that A = (I + S)(I S)1 is a

unitary matrix. Show that a unitary matrix A can be expressed in the above form provided
1 is not an eigenvalue of A.

Solution. See related results of question 2(a) year 1989.

2
Question 2(c) Reduce the quadratic form

6x21 + 3x22 + 3x23 4x1 x2 4x2 x3 + 4x1 x3

to a sum of squares. Also find the corresponding linear transformation, index and signature.

Solution.

Q(x1 , x2 , x3 ) = 6x21 + 3x22 + 3x23 4x1 x2 4x2 x3 + 4x1 x3

2 2 1 1 2
= 6[x21 x1 x2 + x1 x3 + x22 + x23 x2 x3 ]
3 3 9 9 9
7 2 7 2 8
+ x2 + x3 x2 x3
3 3 3
1 1 2 7 4 33
= 6[x1 x2 + x3 ] + [x2 x3 ]2 + x23
3 3 3 7 21
Put X1 = x1 13 x2 + 31 x3 , X2 = x2 47 x3 , X3 = x3 , so that
1
1 3 17

x1 X1
x2 = 0 1 4 X2 (1)
7
x3 0 0 1 X3

and Q(x1 , x2 , x3 ) is transformed to 6X12 + 37 X22 + 33 2

X .
21 3
Let

1 Z1


X1 q 6
3
X2 = 7 Z2
q
X3 7
Z
11 3

then Q(x1 , x2 , x3 ) is transformed to Z12 + Z22 + Z32 , which is its canonical form. Thus
Q(x1 , x2 , x3 ) is positive definite. The Index of Q(x1 , x2 , x3 ) = Number of positive squares in
its canonical form = 3. The signature of Q(x1 , x2 , x3 ) = Number of positive squares - the
number of negative squares in its canonical form = 3.
The required linear transformation which transforms Q(x1 , x2 , x3 ) to sums of squares is
given by (1), and the linear transformation which transforms it to its canonical form is given
by
1

1 1 0 0
x1 1 3 7 6 q Z1
3
x2 = 0 1 4 0 7
0 Z2
7 q
x3 0 0 1 0 0 7 Z3
11