XIX Asian Pacic Mathematics Olympiad

Problem 1. Let S be a set of 9 distinct integers all of whose prime factors are at most 3. Prove that S contains 3 distinct integers such that their product is a perfect cube.

Solution. Without loss of generality, we may assume that S contains only positive integers. Let S = {2ai 3bi | ai , bi Z, ai , bi 0, 1 i 9}. It suces to show that there are 1 i1 , i2 , i3 9 such that ai1 + ai2 + ai3 bi1 + bi2 + bi3 0 (mod 3). ()

For n = 2a 3b S , lets call (a (mod 3), b (mod 3)) the type of n. Then there are 9 possible types : (0, 0), (0, 1), (0, 2), (1, 0), (1, 1), (1, 2), (2, 0), (2, 1), (2, 2). Let N (i, j ) be the number of integers in S of type (i, j ). We obtain 3 distinct integers whose product is a perfect cube when (1) N (i, j ) 3 for some i, j , or (2) N (i, 0)N (i, 1)N (i, 2) = 0 for some i = 0, 1, 2, or (3) N (0, j )N (1, j )N (2, j ) = 0 for some j = 0, 1, 2, or (4) N (i1 , j1 )N (i2 , j2 )N (i3 , j3 ) = 0, where {i1 , i2 , i3 } = {j1 , j2 , j3 } = {0, 1, 2}. Assume that none of the conditions (1)(3) holds. Since N (i, j ) 2 for all (i, j ), there are at least ve N (i, j )s that are nonzero. Furthermore, among those nonzero N (i, j )s, no three have the same i nor the same j . Using these facts, one may easily conclude that the condition (4) should hold. (For example, if one places each nonzero N (i, j ) in the (i, j )-th box of a regular 3 3 array of boxes whose rows and columns are indexed by 0,1 and 2, then one can always nd three boxes, occupied by at least one nonzero N (i, j ), whose rows and columns are all distinct. This implies (4).)

Second solution. Up to (), we do the same as above and get 9 possible types : (a (mod 3), b (mod 3)) = (0, 0), (0, 1), (0, 2), (1, 0), (1, 1), (1, 2), (2, 0), (2, 1), (2, 2) for n = 2a 3b S . Note that (i) among any 5 integers, there exist 3 whose sum is 0 (mod 3), and that (ii) if i, j, k {0, 1, 2}, then i + j + k 0 (mod 3) if and only if i = j = k or {i, j, k } = {0, 1, 2}. Lets dene T : the set of types of the integers in S ; N (i) : the number of integers in S of the type (i, ) ; M (i) : the number of integers j {0, 1, 2} such that (i, j ) T . If N (i) 5 for some i, the result follows from (i). Otherwise, for some permutation (i, j, k ) of (0, 1, 2), N (i) 3, N (j ) 3, N (k ) 1. If M (i) or M (j ) is 1 or 3, the result follows from (ii). Otherwise M (i) = M (j ) = 2. Then either (i, x), (i, y ), (j, x), (j, y ) T or (i, x), (i, y ), (j, x), (j, z ) T for some permutation (x, y, z ) of (0, 1, 2). Since N (k ) 1, at least one of (k, x), (k, y ) and (k, z ) contained in T . Therefore, in any case, the result follows from (ii). (For example, if (k, y ) T , then take (i, y ), (j, y ), (k, y ) or (i, x), (j, z ), (k, y ) from T .)

Problem 2. Let ABC be an acute angled triangle with BAC = 60 and AB > AC . Let I be the incenter, and H the orthocenter of the triangle ABC . Prove that 2AHI = 3ABC.

Solution. Let D be the intersection point of the lines AH and BC . Let K be the intersection point of the circumcircle O of the triangle ABC and the line AH . Let the line through I perpendicular to BC meet BC and the minor arc BC of the circumcircle O at E and N , respectively. We have 1 1 BIC = 180 (IBC + ICB ) = 180 (ABC + ACB ) = 90 + BAC = 120 2 2 and also BN C = 180 BAC = 120 = BIC . Since IN BC , the quadrilateral BICN is a kite and thus IE = EN . Now, since H is the orthocenter of the triangle ABC , HD = DK . Also because ED IN and ED HK , we conclude that IHKN is an isosceles trapezoid with IH = N K . Hence AHI = 180 IHK = 180 AKN = ABN. Since IE = EN and BE IN , the triangles IBE and N BE are congruent. Therefore 1 N BE = IBE = IBC = IBA = ABC 2 and thus 3 AHI = ABN = ABC. 2

Second solution. Let P, Q and R be the intersection points of BH, CH and AH with AC, AB and BC , respectively. Then we have IBH = ICH . Indeed, 1 IBH = ABP ABI = 30 ABC 2 1 1 ICH = ACI ACH = ACB 30 = 30 ABC, 2 2 because ABH = ACH = 30 and ACB + ABC = 120 . (Note that ABP > ABI and ACI > ACH because AB is the longest side of the triangle ABC under the given conditions.) Therefore BIHC is a cyclic quadrilateral and thus 1 BHI = BCI = ACB. 2 and

On the other hand, BHR = 90 HBR = 90 (ABC ABH ) = 120 ABC. Therefore, 1 AHI = 180 BHI BHR = 60 ACB + ABC 2 3 1 = 60 (120 ABC ) + ABC = ABC. 2 2

Problem 3. Consider n disks C1 , C2 , . . . , Cn in a plane such that for each 1 i < n, the center of Ci is on the circumference of Ci+1 , and the center of Cn is on the circumference of C1 . Dene the score of such an arrangement of n disks to be the number of pairs (i, j ) for which Ci properly contains Cj . Determine the maximum possible score.

Solution. The answer is (n 1)(n 2)/2. Lets call a set of n disks satisfying the given conditions an n-conguration. For an nconguration C = {C1 , . . . , Cn }, let SC = {(i, j ) | Ci properly contains Cj }. So, the score of an n-conguration C is |SC |. Well show that (i) there is an n-conguration C for which |SC | = (n 1)(n 2)/2, and that (ii) |SC | (n 1)(n 2)/2 for any n-conguration C . Let C1 be any disk. Then for i = 2, . . . , n 1, take Ci inside Ci1 so that the circumference of Ci contains the center of Ci1 . Finally, let Cn be a disk whose center is on the circumference of C1 and whose circumference contains the center of Cn1 . This gives SC = {(i, j ) | 1 i < j n 1} of size (n 1)(n 2)/2, which proves (i). For any n-conguration C , SC must satisfy the following properties: (1) (i, i) SC , (2) (i + 1, i) SC , (1, n) SC , (3) if (i, j ), (j, k ) SC , then (i, k ) SC , (4) if (i, j ) SC , then (j, i) SC . Now we show that a set G of ordered pairs of integers between 1 and n, satisfying the conditions (1)(4), can have no more than (n 1)(n 2)/2 elements. Suppose that there exists a set G that satises the conditions (1)(4), and has more than (n 1)(n 2)/2 elements. Let n be the least positive integer with which there exists such a set G. Note that G must have (i, i + 1) for some 1 i n or (n, 1), since otherwise G can have at most n n(n 3) (n 1)(n 2) n= < 2 2 2 elements. Without loss of generality we may assume that (n, 1) G. Then (1, n 1) G, since otherwise the condition (3) yields (n, n 1) G contradicting the condition (2). Now let G = {(i, j ) G | 1 i, j n 1}, then G satises the conditions (1)(4), with n 1. We now claim that |G G | n 2 : Suppose that |G G | > n 2, then |G G | = n 1 and hence for each 1 i n 1, either (i, n) or (n, i) must be in G. We already know that (n, 1) G and (n 1, n) G (because (n, n 1) G) and this implies that (n, n 2) G and (n 2, n) G. If we keep doing this process, we obtain (1, n) G, which is a contradiction.

Since |G G | n 2, we obtain |G | (n 1)(n 2) (n 2)(n 3) (n 2) = . 2 2

This, however, contradicts the minimality of n, and hence proves (ii).

 Problem 4. Let x, y and z be positive real numbers such that x + y + z = 1. Prove that x2 + yz y 2 + zx z 2 + xy + + 1. 2x2 (y + z ) 2y 2 (z + x) 2z 2 (x + y )

Solution. We rst note that x2 + yz 2x2 (y + z ) = = x2 x(y + z ) + yz 2x2 (y + z ) (x y )(x z ) 2x2 (y + + x(y + z ) 2x2 (y + z )

y+z 2 + z) y+ z (x y )(x z ) + . 2 2x2 (y + z ) Similarly, we have y 2 + zx 2y 2 (z + x) z 2 + xy 2z 2 (x + y ) We now add (1)(3) to get x2 + yz + y 2 + zx + z 2 + xy (y z )(y x) 2y 2 (z + x) (z x)(z y ) 2z 2 (x + y ) + + z+ 2 x y , .

(1)

(2) (3)

x+ 2

2x2 (y + z ) 2y 2 (z + x) 2z 2 (x + y ) (x y )(x z ) (y z )(y x) (z x)(z y ) + + + x+ y+ z 2x2 (y + z ) 2y 2 (z + x) 2z 2 (x + y ) (x y )(x z ) (y z )(y x) (z x)(z y ) = + + + 1. 2x2 (y + z ) 2y 2 (z + x) 2z 2 (x + y ) Thus, it suces to show that (x y )(x z ) 2x2 (y + z ) + (y z )(y x) 2y 2 (z + x) + (z x)(z y ) 2z 2 (x + y ) 0. (4)

Now, assume without loss of generality, that x y z . Then we have (x y )(x z ) 2x2 (y + z ) 0

and (z x)(z y ) 2z 2 (x + y ) (y z )(x y ) 2z 2 (x + y) + (y z )(y x) 2y 2 (z + x) (y z )(x y ) 2y 2 (z + x) = (y z )(x z ) 2z 2 (x + y ) (y z )(x y ) 2y 2 (z + x) 1 2z 2 (x + y) 1 2y 2 (z + x)

= (y z )(x y )

The last quantity is non-negative due to the fact that y 2 (z + x) = y 2 z + y 2 x yz 2 + z 2 x = z 2 (x + y ). This completes the proof. Second solution. By Cauchy-Schwarz inequality, x2 2x2 (y + z ) + y2 2y 2 (z + x) 2(z + x) + + z2 2(x + y )) ( x + y + z )2 = 1, 2z 2 (x + y ) (5)

( 2(y + z ) + and yz 2x2 (y + z) +

zx 2y 2 (z + x)

xy 2z 2 (x + y) yz + x zx + y xy z
2

(6) .

( 2(y + z ) +

2(z + x) +

2(x + y ))

We now combine (5) and (6) to nd x2 + yz + y 2 + zx + z 2 + xy 2z 2 (x + y ) 2(z + x)) 2 yz 2 + x

2x2 (y + z ) 2y 2 (z + x) ( 2(x + y ) + 2(y + z ) + yz zx xy + + 1+ x y z Thus, it suces to show that 2 yz + x zx + y xy z

zx + y

xy z

2(y + z ) +

2(z + x) +

2(x + y ) .

(7)

Consider the following inequality using AM-GM inequality yz + x 1 2 zx 1 + y 2 xy z

yz x

1 2

zx 1 + y 2

xy z

= 2(y + z ),

or equivalently yz + x Similarly, we have zx + y xy + z 1 2 1 2 xy 1 + z 2 yz 1 + x 2 yz x zx y 2(z + x) , 2(x + y ) . 1 2 zx 1 + y 2 xy z 2(y + z ) .

Adding the last three inequalities, we get 2 yz + x zx + y xy z 2(y + z ) + 2(z + x) + 2(x + y ) .

This completes the proof.

Problem 5. A regular (5 5)-array of lights is defective, so that toggling the switch for one light causes each adjacent light in the same row and in the same column as well as the light itself to change state, from on to o, or from o to on. Initially all the lights are switched o. After a certain number of toggles, exactly one light is switched on. Find all the possible positions of this light.

Solution. We assign the following rst labels to the 25 positions of the lights: 1 0 1 0 1 1 0 1 0 1 0 0 0 0 0 1 0 1 0 1 1 0 1 0 1

For each on-o combination of lights in the array, dene its rst value to be the sum of the rst labels of those positions at which the lights are switched on. It is easy to check that toggling any switch always leads to an on-o combination of lights whose rst value has the same parity(the remainder when divided by 2) as that of the previous on-o combination. The 90 rotation of the rst labels gives us another labels (let us call it the second labels) which also makes the parity of the second value(the sum of the second labels of those positions at which the lights are switched on) invariant under toggling. 1 1 0 1 1 0 0 0 0 0 1 1 0 1 1 0 0 0 0 0 1 1 0 1 1

Since the parity of the rst and the second values of the initial status is 0, after certain number of toggles the parity must remain unchanged with respect to the rst labels and the second labels as well. Therefore, if exactly one light is on after some number of toggles, the label of that position must be 0 with respect to both labels. Hence according to the above pictures, the possible positions are the ones marked with i s in the following picture:

2 0 3

1 4

Now we demonstrate that all ve positions are possible : Toggling the positions checked by t (the order of toggling is irrelevant) in the rst picture makes the center(0 ) the only position with light on and the second picture makes the position 1 the only position with light on. The other i s can be obtained by rotating the second picture appropriately. t t t t t t t t t t t t t t t t t t t t t t