11

th
International Mathematical Competition for University Students
Skopje, 2526 July 2004
Solutions for problems on Day 2
1. Let A be a real 4 2 matrix and B be a real 2 4 matrix such that
AB =
_
_
_
_
1 0 1 0
0 1 0 1
1 0 1 0
0 1 0 1
_
_
_
_
.
Find BA. [20 points]
Solution. Let A =
_
A
1
A
2
_
and B =
_
B
1
B
2
_
where A
1
, A
2
, B
1
, B
2
are 2 2 matrices. Then
_
_
_
_
1 0 1 0
0 1 0 1
1 0 1 0
0 1 0 1
_
_
_
_
=
_
A
1
A
2
_
_
B
1
B
2
_
=
_
A
1
B
1
A
1
B
2
A
2
B
1
A
2
B
2
_
therefore, A
1
B
1
= A
2
B
2
= I
2
and A
1
B
2
= A
2
B
1
= I
2
. Then B
1
= A
1
1
, B
2
= A
1
1
and A
2
= B
1
2
=
A
1
. Finally,
BA =
_
B
1
B
2
_
_
A
1
A
2
_
= B
1
A
1
+ B
2
A
2
= 2I
2
=
_
2 0
0 2
_
2. Let f, g : [a, b] [0, ) be continuous and non-decreasing functions such that for each x [a, b] we
have
_
x
a
_
f(t) dt
_
x
a
_
g(t) dt
and
_
b
a
_
f(t) dt =
_
b
a
_
g(t) dt.
Prove that
_
b
a
_
1 + f(t) dt
_
b
a
_
1 + g(t) dt. [20 points]
Solution. Let F(x) =
_
x
a
_
f(t) dt and G(x) =
_
x
a
_
g(t) dt. The functions F, G are convex, F(a) = 0 =
G(a) and F(b) = G(b) by the hypothesis. We are supposed to show that
_
b
a
_
1 +
_
F

(t)
_
2
dt
_
b
a
_
1 +
_
G

(t)
_
2
dt
i.e. The length ot the graph of F is the length of the graph of G. This is clear since both functions are
convex, their graphs have common ends and the graph of F is below the graph of G the length of the
graph of F is the least upper bound of the lengths of the graphs of piecewise linear functions whose values
at the points of non-dierentiability coincide with the values of F, if a convex polygon P
1
is contained in
a polygon P
2
then the perimeter of P
1
is the perimeter of P
2
.
3. Let D be the closed unit disk in the plane, and let p
1
, p
2
, . . . , p
n
be xed points in D. Show that there
exists a point p in D such that the sum of the distances of p to each of p
1
, p
2
, . . . , p
n
is greater than or
equal to 1. [20 points]
Solution. considering as vectors, thoose p to be the unit vector which points into the opposite direction as
n

i=1
p
i
. Then, by the triangle inequality,
n

i=1
|p p
i
|

np
n

i=1
p
i

= n +

i=1
p
i

n..
4. For n 1 let M be an nn complex matrix with distinct eigenvalues
1
,
2
, . . . ,
k
, with multiplicities
m
1
, m
2
, . . . , m
k
, respectively. Consider the linear operator L
M
dened by L
M
(X) = MX +XM
T
, for any
complex n n matrix X. Find its eigenvalues and their multiplicities. (M
T
denotes the transpose of M;
that is, if M = (m
k,l
), then M
T
= (m
l,k
).) [20 points]
Solution. We rst solve the problem for the special case when the eigenvalues of M are distinct and all sums

r
+
s
are dierent. Let
r
and
s
be two eigenvalues of M and v
r
, v
s
eigenvectors associated to them, i.e.
Mv
j
= v
j
for j = r, s. We have Mv
r
(v
s
)
T
+v
r
(v
s
)
T
M
T
= (Mv
r
)(v
s
)
T
+v
r
_
Mv
s
_
T
=
r
v
r
(v
s
)
T
+
s
v
r
(v
s
)
T
,
so v
r
(v
s
) is an eigenmatrix of L
M
with the eigenvalue
r
+
s
.
Notice that if
r
=
s
then vectors u, w are linearly independent and matrices u( w)
T
and w(u)
T
are
linearly independent, too. This implies that the eigenvalue
r
+
s
is double if r = s.
The map L
M
maps n
2
dimensional linear space into itself, so it has at most n
2
eigenvalues. We already
found n
2
eigenvalues, so there exists no more and the problem is solved for the special case.
In the general case, matrix M is a limit of matrices M
1
, M
2
, . . . such that each of them belongs to the
special case above. By the continuity of the eigenvalues we obtain that the eigenvalues of L
M
are
2
r
with multiplicity m
2
r
(r = 1, . . . , k);

r
+
s
with multiplicity 2m
r
m
s
(1 r < s k).
(It can happen that the sums
r
+
s
are not pairwise dierent; for those multiple values the multiplicities
should be summed up.)
5. Prove that
_
1
0
_
1
0
dx dy
x
1
+| ln y| 1
1. [20 points]
Solution 1. First we use the inequality
x
1
1 | ln x|, x (0, 1],
which follows from
(x
1
1)

x=1
= | ln x||
x=1
= 0,
(x
1
1)

=
1
x
2

1
x
= | ln x|

, x (0, 1].
Therefore
_
1
0
_
1
0
dx dy
x
1
+| ln y| 1

_
1
0
_
1
0
dx dy
| ln x| +| ln y|
=
_
1
0
_
1
0
dx dy
| ln(x y)|
.
Substituting y = u/x, we obtain
_
1
0
_
1
0
dx dy
| ln(x y)|
=
_
1
0
__
1
u
dx
x
_
du
| ln u|
=
_
1
0
| ln u|
du
| ln u|
= 1.
Solution 2. Substituting s = x
1
1 and u = s ln y,
_
1
0
_
1
0
dx dy
x
1
+| ln y| 1
=
_

0
_

s
e
su
(s + 1)
2
u
duds =
_

0
__
u
0
e
s
(s + 1)
2
ds
_
e
u
u
dsdu.
Since the function
e
s
(s+1)
2
is convex,
_
u
0
e
s
(s + 1)
2
ds
u
2
_
e
u
(u + 1)
2
+ 1
_
so
_
1
0
_
1
0
dx dy
x
1
+| ln y| 1

_

0
u
2
_
e
u
(u + 1)
2
+ 1
_
e
u
u
du =
1
2
__

0
du
(u + 1)
2
+
_

0
e
u
du
_
= 1.
6. For n 0 dene matrices A
n
and B
n
as follows: A
0
= B
0
= (1) and for every n > 0
A
n
=
_
A
n1
A
n1
A
n1
B
n1
_
and B
n
=
_
A
n1
A
n1
A
n1
0
_
.
Denote the sum of all elements of a matrix M by S(M). Prove that S(A
k1
n
) = S(A
n1
k
) for every n, k 1.
[20 points]
Solution. The quantity S(A
k1
n
) has a special combinatorical meaning. Consider an n k table lled with
0s and 1s such that no 2 2 contains only 1s. Denote the number of such llings by F
nk
. The lling of
each row of the table corresponds to some integer ranging from 0 to 2
n
1 written in base 2. F
nk
equals
to the number of k-tuples of integers such that every two consecutive integers correspond to the lling of
n 2 table without 2 2 squares lled with 1s.
Consider binary expansions of integers i and j i
n
i
n1
. . . i
1
and j
n
j
n1
. . . j
1
. There are two cases:
1. If i
n
j
n
= 0 then i and j can be consecutive i i
n1
. . . i
1
and j
n1
. . . j
1
can be consequtive.
2. If i
n
= j
n
= 1 then i and j can be consecutive i i
n1
j
n1
= 0 and i
n2
. . . i
1
and j
n2
. . . j
1
can be
consecutive.
Hence i and j can be consecutive i (i +1, j +1)-th entry of A
n
is 1. Denoting this entry by a
i,j
, the sum
S(A
k1
n
) =

2
n
1
i
1
=0

2
n
1
i
k
=0
a
i
1
i
2
a
i
2
i
3
a
i
k1
i
k
counts the possible llings. Therefore F
nk
= S(A
k1
n
).
The the obvious statement F
nk
= F
kn
completes the proof.