Вы находитесь на странице: 1из 5

MATH1115, Tutorial Worksheet for Week 7 1

Linear Algebra
Question 1. Determinant
Let
a0
1 0 0 0
1 a1 1 0 0

0 1 a2 1 0
n = . .. .

.. .. ..
.. . . . .

0
0 1 an1 1
0 0 0 1 an
Prove that
1 n
an + = .
1 n1
an1 +
1
an2 +
1
+
1
a1 +
a0
Hint: Use induction and the cofactor expansion.
Answer
When n = 0, n = a0 .
Now consider the case of n = 1. Then

a 1
1 = 0 = a0 a1 + 1.
1 a1

So

0 (a0 a1 + 1) = a0 1
0 (a1 + 1/a0 ) = 1
1 1
a1 + =
a0 0

If n = 2,

a0 1 0
a 1
1 = a2 1 1 0

2 = 1 a1 = a2 1 + 1 a0 = a2 1 + 0 .
0 1
0 1 a2

So

a2 1 + 0 = 2
0 2
a2 + =
1 1
1 2
a2 + =
1 1
0
1 2
a2 + =
1 1
a1 +
a0
MATH1115, Tutorial Worksheet for Week 7 2

Assume the equality hold true for n = k, prove true for n = k + 1.


a0
1 0 0 0

1 a1 1 0 0

0 1 a2 1 0
k+1 = .

.. .. .. ..
..

. . . .

0
0 1 ak 1
0 0 0 1 ak+1

a0
1 0 0
1 a1 1 0

0 1 a2 1
= (1)k+1+k+1 ak+1 k + (1)k+1+k .

.. .. ..
..

. . .

0
1 ak1 1
0 0 0 1
= ak+1 k + k1 .

So

ak+1 k + k1 = k+1
1 k+1
ak+1 + = .
k k
k1
Which completes the proof.

Calculus
Question 2. Convergent Sequences

1
(a) Use the formal definition of the limit to prove that lim = 0 .
n n
That is, for all > 0 show that there exists a natural number N () such that

1
n N () = < .

n

2 + n2
(b) By applying limit laws, find the value L = lim .
n 1 + 2n2

[Hint: First divide the top and bottom of the fraction by n2 .]


Then prove your answer using only the definition of the limit. That is, for all > 0 find a
natural number N () such that
2 + n2


n N () = L < .
1 + 2n2

(c) Define a sequence (an ) inductively by a1 = 1 and an+1 = 21 an (4 an ).


(i) Show that 0 < an+1 < 2 if 0 < an < 2, and deduce that an (0, 2) for all n N.
(ii) Show that the sequence (an ) is increasing and bounded above.
MATH1115, Tutorial Worksheet for Week 7 3

(iii) Prove that the sequence converges, and find its limit.

(d) In parts (i) to (v) below, decide whether the sequences (an ) defined by the given formulae
are convergent, and determine their limits if possible using limit laws:
2n
(i) an = .
n2 + 1
1 + 2n 7n2
(ii) an =
3 + n + n2
(1)n
(iii) an =2 .
n
n
2
(iv) an = n.
n
en en
(v) an = n .
e + en
When using limit laws in the above you may take as given the facts that a a for a constant
a R, n1p 0 for a constant p > 0, and rn 0, for a constant 0 < r < 1. Additionally,
remember you have at your disposal the squeeze theorem, the monotone convergence theorem,
and the standard theorem on limit laws:
Theorem 1. Suppose
lim an = a, lim bn = b,
and c is a real number. Then the following limits exist and have the given values.

lim(an bn ) = a b,
lim can = ca,
lim an bn = ab,
an a
lim = , assuming b 6= 0.
bn b
Answer
(a) I want them to write it out properly, but choosing N () > 12 is what is necessary.
2 + n2

1 3
(b) Note that for n N we have 2
<
2
<
1 + 2n 2 2(1 + 2n )
s  
2
3 2 3 1 3 1
(1 + 2n ) >
n > n > max 0, .
2 4 2 4 2
s  
3 1
Therefore, given > 0, if we choose N () N such that N () > max 0, ,
4 2
s
2 + n2
 
3 1 1
we have that n N () = n > max 0, = < .
4 2 1 + 2n2 2
Hence,
2 + n2
 
1
( > 0) (N N) n N =
< .
1 + 2n2 2
2 + n2 1
That is, by definition, lim = .
n 1 + 2n2 2
MATH1115, Tutorial Worksheet for Week 7 4

(c) (i) Assume 0 < an < 2 for some n N.


Then 0 < 4 an , and so
1 1 2
0 < 2 an (4 an ) = 2 (4an an )
= 2 + 12 (4 + 4an a2n )
= 2 21 (4 4an + a2n )
= 2 21 (2 an )2 < 2 .

Hence 0 < an < 2 implies that 0 < an+1 < 2.


Since we also have that 0 < a1 = 1 < 2, by induction an (0, 2) for all
n N.
(ii) Part (i) shows that (an ) is bounded.
Also, part (i) gives that an+1 = 12 an (4 an ) > 21 an (4 2) = 12 2an = an .
So an is strictly increasing.
(iii) By part (ii), (an ) is monotone and bounded.
Therefore (an ) converges by the Monotone Convergence Theorem.
Let a = lim an . Then an+1 a also. Thus
n
  
a = lim an+1 = lim ( 21 an (4an )) = 21 lim an 4 lim an = 12 a(4a) .
n n n n

Therefore 2a = 4a a2 , hence a2 2a = 0, which implies a(a 2) = 0.


So a = 0 or a = 2.
It is not possible that an 0, because an is increasing and a1 > 0.
Hence an 2.
1
(d) (i) If an = n22n 2 2n
+1 , divide numerator and denominator by n to give an = 1+n2 .
1 2
Since n 0 and n 0, we have an 0 by the theorem on limit laws.
(ii) Since n 1 we can divide numerator and denominator by n2 to obtain an =
7+2/n+1/n2
1+1/n+3/n2 . We know n lim n1 = lim n2 = 0, so by the theorem on limit
n
laws we have limn an = 7.
n
(iii) If an = 2 (1) 1
n , use the squeeze theorem: We have 2 n an 2 + n ,
1
1
and the sequences on each side converge to 2 since n 0. Therefore by the
squeeze theorem an 2.
n
(iv) If an = n2 n , we can again use the squeeze theorem: For n 4 we have nn 4n
n
and therefore 0 < an 42n = 2n . Since 2n 0 we have an 0 by the
squeeze theorem.
1e2n
(v) Divide top and bottom by en to get an = 1+e2n . We have en 0, hence
an 0 by the theorem on limit laws.

Extra Question
Question 3. Matrix form of Differential Equations
Suppose is a scalar and that A() is an m n matrix with entries aij (). If aij () is a
d
differentiable function of for all i and j, then we denote by A() or d A() the matrix whose
ij-th entries are given by the following (note the many different ways of expressing the same thing):
 
  d d
A() = A() = aij () = a0ij () = aij ().
ij d ij d
MATH1115, Tutorial Worksheet for Week 7 5

(a) Suppose A() is an m r matrix and B() is a r n matrix whose entries are differentiable
functions of the scalar . Show
d d d
[A()B()] = [A()] B() + A() [B()] .
d d d

(b) Suppose A() is an n n matrix whose entries are differentiable functions of the scalar .
Assuming A() is always non-singular, show

d  d
A()1 = A()1 [A()] A()1 .

d d

Answer
(a)
 
d d  
[A()B()] = [A()B()]ij
d ij d
r
!
d X
= a()ik b()kj
d
k=1
r
X d
= (a()ik b()kj )
d
k=1
r  
X d d
= [a()ik ] b()kj + a()ik [b()kj ]
d d
k=1
r   r  
X d X d
= [a()ik ] b()kj + a()ik [b()kj ]
d d
k=1 k=1
   
d d
= [A()] B() + A() [B()]
d ij d ij

As this is true for all 1 i m, 1 j n


d d d
[A()B()] = [A()] B() + A() [B()] .
d d d
d
(b) Firstly note that if In is the n n identity matrix then d In = 0 as each entry
in In is a constant. So
d d  1 d  1  d
A () A() + A1 ()

0= [In ] = A ()A() = [A()] .
d d d d
Thus
d  1  d
A () A() = A1 () [A()]
d d
which then gives
d  1  d
A () = A1 () [A()] A1 ().
d d

Вам также может понравиться