0% found this document useful (0 votes)
11 views

2014 Junior Solution

The document presents solutions to various problems from the Second HKUST Undergraduate Math Competition at the junior level. It covers topics such as polynomial differences, eigenvalues, function behavior, series summation, calculus, and lighting configurations in a grid. Each solution is detailed with mathematical reasoning and proofs to support the conclusions drawn.

Uploaded by

nikhilkumawat303
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
11 views

2014 Junior Solution

The document presents solutions to various problems from the Second HKUST Undergraduate Math Competition at the junior level. It covers topics such as polynomial differences, eigenvalues, function behavior, series summation, calculus, and lighting configurations in a grid. Each solution is detailed with mathematical reasoning and proofs to support the conclusions drawn.

Uploaded by

nikhilkumawat303
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Solution of the Second HKUST Undergraduate Math Competition – Junior Level

1. Below we will write each term ak xk as a difference of two increasing polynomials.


For a0, let g0 (x) = x + a0 and h0(x) = x, then a0 = g0 (x) − h0 (x).
Let k ∈ {1, 2, . . ., n}. For ak xk with k odd, if ak ≥ 0, then let gk (x) = ak xk and hk (x) = 0 Otherwise
ak < 0, then let gk (x) = 0 and hk (x) = −ak xk . In both cases, we have ak xk = gk (x) − hk (x).
For ak xk with k even, observe that the polynomial pk (x) = xk+1 + xk + xk−1 is increasing since
0
pk (x) = (k + 1)xk + kxk−1 + (k − 1)xk−2 = xk−2 (k + 1)x2 + kx + (k − 1) ≥ 0 due to the discriminant
of the quadratic factor is k2 − 4(k2 − 1) = −3k2 + 4 < 0.
If ak ≥ 0, then let gk (x) = ak (xk+1 + xk + xk−1) and hk (x) = ak (xk+1 + xk−1). Otherwise ak < 0,
then let gk (x) = −ak (xk+1 + xk−1) and hk (x) = −ak (xk+1 + xk + xk−1). In both cases, we have
ak xk = gk (x) − hk (x).
Finally f = g − h, where g(x) = g0 (x) + g1(x) + g2(x) + · · · + gn (x) and h(x) = h0(x) + h1(x) +
h2 (x) + · · · + hn (x) are (strictly) increasing.

2. (1) By assumption, v 6= 0, Av = cv for some nonzero scalar c by invertibility of A. Now ABv = λBAv =
λcBv. So Bv is an eigenvector of A with eigenvalue λc.
(2) Solution 1 By induction B m v is an eigenvector of A with eigenvalue λm c for any positive integer m.
0
Since A has at most n eigenvalues, so λm c = λm c for some m0 > m. Then λ is an (m0 − m)-th root of
unity.
Solution 2 We have det(λBA) = det(AB). So λn det(B) det(A) = det(A) det(B), which is nonzero
due to invertibility of A and B. So λn = 1.

1 1
3. (Solution due to Kwok Ling Hon) For every x ∈ (−∞, +∞), |f 0(x)| = ≤ .
|3 + 2 sin f(x) + x2 | 1 + x2
Let c = f(0). Now
Z x nZ 0 Z ∞ o Z ∞
dt π
|f(x) − f(0)| = f 0 (t) dt ≤ max |f 0 (t)| dt, |f 0 (t)| dt ≤ 2
= .
0 −∞ 0 0 1+t 2
h π πi
So, f(x) is always in c − , c + .
2 2

4. By interchanging i and j in one of the sums, it is clear that the second sum is the negative of the first
sum. So it suffices to compute the first sum.

X

(−1)i+j (i − j) 1X
∞  1 1 
1 = (−1)i+j 1 + =
j=0
(i − j)2 − 4
2
j=0
i−j + 2 i − j − 12

(−1)i h 1 1   1 1   1 1  i
1
+ − + + + − · · · .
2 i+ 2 i − 12 i− 1
2 i − 32 i− 3
2 i − 52

This is clearly a telescoping series whose sum is (−1)i /(2i + 1). Therefore, the first sum is

X
∞ X

(−1)i+j (i − j) X∞
(−1)i 1 1 1
2 1
= = 1− + − +···
i=0 j=0
(i − j) − 4 i=0
2i + 1 3 5 7
Z 1 Z 1
1 π
= (1 − x2 + x4 − x6 + · · ·) dx = 2
dx = Arctan 1 = .
0 0 1 + x 4
Z a2
1 x−1 1 x−1
5. Define f(a) = ln dx and let F (x) be an antiderivative of g(x) = ln for x ∈ (0, +∞).
a x 32 x 32
2
By the fundamental theorem of calculus, f(a) = F (a ) − F (a). Hence,

2 a2 − 1 1 a − 1
f 0 (a) = 2aF 0(a2 ) − F 0 (a) =
ln − ln
a 32 a 32
ln{(a − 1)(a + 1)2} − ln 32
= .
a

It follows that f 0 (a) < 0 if a ∈ (1, 3) and f 0 (a) > 0 if a ∈ (3, +∞). Then f(a) achieves its minimum for
a = 3.

6. It is not possible. The total number of lighting configurations in the grid is 2169. We will show that the
number of configurations achievable using the switches is less than 2169.
We first observe that we never need to use a given switch more than once. If we flip it an even
number of times, it has no net effect; if we flip it an odd number of times, it has the same effect as
flipping it once. Secondly, we observe that the order in which we flip switches has no effect, since all
the matters for any given light is the number of times it is reversed. Together, these observations imply
that all achievable configurations can be obtained by flipping some subset of the switches one time each,
in any order.
The total number of switches is (13 − 9 + 1)2 + (13 − 2 + 1)2 = 52 + 122 = 169, which means there
are 2169 subsets of switches we may choose from. In order to show that the number of available lightings
is strictly less than 2169, we need to find two unequal sets of switches that result in the same lighting.
One example is as follows: we can get all lights except the center 5 × 5 square to be on in 2 ways. One
is to light the lower-left and upper-right 9 × 9 squares, then light the upper-left and lower-right 4 × 4
squares (using the 2 × 2 squares). This same lighting can be achieved by rotating the above example by
90◦ .
Thus, there are at most 2169 − 1 available patterns.

You might also like