UIUC Mock Putnam Exam 1/2005 Solutions

UIUC Mock Putnam Exam 1/2005
Solutions
Problem 1. Without any numerical calculations, determine which of

the two numbers e and e is larger.
Solution. Let a = e and b = e . We will show that a > b. Since
ln a = ln e = and ln b = e ln , and since taking logarithms preserves
inequalities, we see that a > b holds if and only if () (ln e)/e > (ln )/.
Now consider the function f (x) = (ln x)/x. We have f 0 (x) = (1 ln x)/x2 ,
so f is decreasing for x > e, and since > e, it follows that f () < f (e)
which is equivalent to ().
Problem 2. Prove that, given any set of 10 distinct positive integers

below 100, there exist two disjoint non-empty subsets whose elements have
the same sum.
Solution. Let A be a set of 10 distinct positive integers 100. Note
that the sum over the elements of any subset of A must be between 1 and
10 100 = 1000, so there are at most 10 1000 possible values for such a
sum. On the other hand, there are 210 1 = 1023 nonempty subsets of A.
Hence, by the pigeonhole principle, there exist two distinct nonempty subsets
B, C A whose elements have the same sum. If B and C are disjoint, we are
done. Otherwise, let B 0 = B \ B C, C 0 = C \ B C. Then, by construction,
B 0 and C 0 are disjoint. Moreover, B 0 and C 0 still have the same element
sum since they were obtained from B and C by removing the same set of
elements. Finally, each of B 0 and C 0 is nonempty, since if they were both
empty, then we would have B = C, contradicting our assumption that B and
C are distinct, and if one of B 0 and C 0 is empty and the other non-empty,
then B 0 and C 0 could not have the same element sum.
Problem 3. How many positive integers are there which, in their decimal
representation, have strictly decreasing digits? Explain!
1
Solution. The key to this problem is the fact that there is a one-to-
one correspondence between nonnegative integers (i.e., positive or zero) with
decreasing digits and nonempty subsets of {0, 1, . . . , 9}. This is seen as fol-
lows: First, given such a subset {a1 , . . . , ak }, with the ai s written in de-
creasing order, let n be the integer obtained by concatenating the ai s, i.e.,
n = a1 a2 . . . ak . Then n is a nonnegative integer with decreasing digits. Con-
versely, given a nonnegative integer with decreasing digits, let a1 , a2 , . . . , ak
denote the sequence of its digits. Then the set {a1 , . . . , ak } is a nonempty
subset of {0, 1, . . . , 9}.
Thus, the number of nonnegative integers with decreasing digits is equal
to the number of non-empty subsets of the 10-element set {0, 1, . . . , 9}, i.e.,
210 1 = 1023. The integer 0 is included in this count, so the number of
positive integers with decreasing digits sought in the problem is one less, i.e.,
1022.
Problem 4. Let x = 0.0102040816326428 . . . be the real number in the

interval (0, 1) whose decimal expansion (after the decimal period) is obtained
by concatenating the last two digits of the sequence of powers of 2 (padded
by 0 in case there is only one digit). Is x is rational? Explain!
Solution. We claim that x is rational. To show this, it suffices to
show that its decimal expansion is ultimately periodic. Now, by definition,
x = 0.a0 a1 a2 . . . , where an denotes the last two digits of 2n , padded with
zeros when necessary. Thus, it suffices to show that the sequence {an } n=0 is
ultimately periodic.
To this end, first note that an is the remainder of 2n upon division by
100, i.e., an 2n mod 100. Since there are only finitely many possible values
for an , there exist m < n such that am = an . Now, if am = an , then
2n+i = 2i 2n 2i an = 2i am 2i 2m = 2m+i mod 100,
and hence an+i = am+i for all i 1. Thus, the sequence an is ultimately
periodic (with period nm). This proves our claim and completes the proof.
Problem 5. Find a function f (x) satisfying

1
f (x) + f =x
1x
for all x 6= 0, 1.
Solution. Let (1) denote the given equation for f (x). Substituting (1x)1
for x in (1), we get a second equation,

1 1 1
f +f 1 = , (2)
1x x 1x
2
valid for all x 6= 0, 1. We repeat the process one more time, substituting
(1 x)1 for x in (2). The result is

1 1
f 1 + f (x) = 1 . (3)
x x
We stop here, since the next substitution would only lead back to the original
equation (1). However, we now have enough equations to determine f (x):
Setting X = f (x), Y = f (1/(1x), and Z = f (11/x)), the three equations
(1), (2), (3), can be written as
X + Y = x,
1
Y +Z = ,
1x
1
Z +X =1 .
x
This is a linear system of three equations in three unknowns X, Y, Z, which
is easily solved. The result is

1 1 1
X = f (x) = x +1 .
2 1x x
One can check that this function is indeed a solution to (1).
Problem 6. Let A denote the set of positive integers whose decimal

expansion
P does not have a zero. Determine, with proof, whether the infinite
series nA 1/n converges.
Solution. We break the range of summation P into finite intervals Ik =
k1 k
10 n < 10 , k = 1, 2, . . ., and let Sk = nIk A 1/n denote the corre-
sponding partial sum. Now, note that an integer n Ik has at most k digits,
and if n is also in the set A, then there are only 9 possible values for each of
these digits. Hence the total number of elements in Ik A is at most 9k . On
the other hand, if n Ik then n 10k1 . Thus,
X 1 1 9k
Sk = k1
#{n : n Ik A} k1
,
nIk A
10 n 10
and therefore
X1 X k
X 9 9
= Sk 10 = = 90 < .
nA
n k=1 k=1
10 1 9/10
3
Hence the series converges.
Problem 7. Let P (t) be a non-constant polynomial with real coefficients.

Prove that there exist only finitely many positive real numbers x such that
Z x Z x
P (t) sin tdt = P (t) cos tdt = 0.
0 0
Rx it
Solution. RLet f (x) = 0 P (t)e dt. Then f (x) = 0 if and only if the
x R x
two integrals 0 P (t) sin tdt and 0 P (t) cos tdt are simultaneously R x zero. By
k it
induction, one can show that for each nonnegative integer k, 0 t e dt =
Qk (x)eix + ck , where Qk (x) is a polynomial of degree
Pn k with real coefficients
k
and ck a constant. It follows that, if P (t) = k=0 ak t is a polynomial
ix
of
Pndegree n with real coefficients, then f (x) = Q(x)e + c where Q(x) =
k=0 ak Q
Pkn(x) is a polynomial of the same degree as P (t) with real coefficients
and c = k=0 ak ck a constant. Setting x = 0, we see that c must be equal to
Q(0). Thus, f (x) = 0 is equivalent to () Q(x) = Q(0)eix , and we have
to show that () has only finitely many solutions.
By our assumption, the polynomial P (x) is non-constant. Since Q(x) has
the same degree as P (x), Q(x) is also non-constant. Hence |Q(x)| if
|x| . Therefore, all solutions to () must fall into an interval of the form
|x| C, where C is a constant.
If Q(0) = 0, then () reduces to the equation Q(x) = 0 which has at
most n solutions by the fundamental theorem of algebra, and we are done.
Suppose therefore that Q(0) 6= 0. Taking real parts, we obtain from ()
Q(x) = Q(0) cos x. If this equation had infinitely many solutions in the
interval [C, C], then by applying Rolles theorem to the function Q(x) +
Q(0) cos x and its successive derivatives, we obtain that, for each positive
integer k, the equation () Q(4k) (x) = Q(0) cos x, also had infinitely many
solutions x in the same interval. (Note that Rolles theorem does not hold
for complex-valued functions; thus it was necessary to take real parts in ()
before applying the theorem.) However, since Q(x) is a polynomial of degree
n, the left side of () is zero if k > n/4, whereas (since Q(0) 6= 0) the
right side can only be zero at the zeroes of cos x, of which there are only
finitely many in the interval [C, C]. Hence () can have only finitely many
solutions.
Remark. The last part of the solution can be considerably shortened
with a bit of complex analysis magic. Namely, let f (z) = Q(z) + Q(0)eiz .
Since Q(z) is a non-constant polynomial, this function is an analytic function
of z that is not constant. By a theorem from complex analysis, such a function
can have at most finitely many zeros in any disk |z| C. In particular, it
can have at most finitely many real zeros in this disk, and this is exactly
what we showed above using more elementary methods.

UIUC Mock Putnam Exam 1/2005 Solutions

Uploaded by

Copyright:

Available Formats

UIUC Mock Putnam Exam 1/2005 Solutions

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

UIUC Mock Putnam Exam 1/2005 Solutions

Uploaded by

Copyright:

Available Formats

UIUC Mock Putnam Exam 1/2005

Problem 1. Without any numerical calculations, determine which of

Problem 2. Prove that, given any set of 10 distinct positive integers

Problem 4. Let x = 0.0102040816326428 . . . be the real number in the

Problem 5. Find a function f (x) satisfying

One can check that this function is indeed a solution to (1).

Problem 6. Let A denote the set of positive integers whose decimal

Problem 7. Let P (t) be a non-constant polynomial with real coefficients.

You might also like