Number theoryDifficulty 8.0National olympiad, round 2Prove itIMO
Find all triples (a,b,c) of positive integers such that a3+b3+c3=(abc)2.
Solution
Solution 1. We will start by proving that c=1. Note that 3a3⩾a3+b3+c3>a3 So 3a3⩾(abc)2>a3 and hence 3a⩾b2c2>a. Now b3+c3=a2(b2c2−a)⩾a2, and so 18b3⩾9(b3+c3)⩾9a2⩾b4c4⩾b3c5 so 18⩾c5 which yields c=1.
Now, note that we must have a>b, as otherwise we would have 2b3+1=b4 which has no positive integer solutions. So a3−b3⩾(b+1)3−b3>1 and 2a3>1+a3+b3>a3, which implies 2a3>a2b2>a3 and so 2a>b2>a. Therefore 4(1+b3)=4a2(b2−a)⩾4a2>b4 so 4>b3(b−4); that is, b⩽4.
Now, for each possible value of b with 2⩽b⩽4 we obtain a cubic equation for a with constant coefficients. These are as follows: b=2:b=3:b=4:a3−4a2+9=0a3−9a2+28=0a3−16a2+65=0. The only case with an integer solution for a with b⩽a is b=2, leading to (a,b,c)=(3,2,1).
Solution 2. Again, we will start by proving that c=1. Suppose otherwise that c⩾2. We have a3+b3+c3⩽3a3, so b2c2⩽3a. Since c⩾2, this tells us that b⩽3a/4. As the right-hand side of the original equation is a multiple of a2, we have a2⩽2b3⩽2(3a/4)3/2. In other words, a⩽1627<2, which contradicts the assertion that a⩾c⩾2. So there are no solutions in this case, and so we must have c=1.
Now, the original equation becomes a3+b3+1=a2b2. Observe that a⩾2, since otherwise a=b=1 as a⩾b. The right-hand side is a multiple of a2, so the left-hand side must be as well. Thus, b3+1⩾a2. Since a⩾b, we also have b2=a+a2b3+1⩽2a+a21 and so b2⩽2a since b2 is an integer. Thus (2a)3/2+1⩾b3+1⩾a2, from which we deduce a⩽8.
Now, for each possible value of a with 2⩽a⩽8 we obtain a cubic equation for b with constant coefficients. These are as follows: a=2:a=3:a=4:a=5:a=6:a=7:a=8:b3−4b2+9=0b3−9b2+28=0b3−16b2+65=0b3−25b2+126=0b3−36b2+217=0b3−49b2+344=0b3−64b2+513=0. The only case with an integer solution for b with a⩾b is a=3, leading to (a,b,c)=(3,2,1).
Solution 3. Set k=(b3+c3)/a2⩽2a, and rewrite the original equation as a+k=(bc)2. Since b3 and c3 are positive integers, we have (bc)3⩾b3+c3−1=ka2−1, so a+k⩾(ka2−1)2/3 As in Comment 1.2, k is a positive integer; for each value of k⩾1, this gives us a polynomial inequality satisfied by a: k2a4−a3−5ka2−3k2a−(k3−1)⩽0. We now prove that a⩽3. Indeed, 0⩾k2k2a4−a3−5ka2−3k2a−(k3−1)⩾a4−a3−5a2−3a−k⩾a4−a3−5a2−5a, which fails when a⩾4. This leaves ten triples with 3⩾a⩾b⩾c⩾1, which may be checked manually to give (a,b,c)=(3,2,1).
Solution 4. Again, observe that b3+c3=a2(b2c2−a), so b⩽a⩽b2c2−1. We consider the function f(x)=x2(b2c2−x). It can be seen that that on the interval [0,b2c2−1] the function f is increasing if x<32b2c2 and decreasing if x>32b2c2. Consequently, it must be the case that b3+c3=f(a)⩾min(f(b),f(b2c2−1)) First, suppose that b3+c3⩾f(b2c2−1). This may be written b3+c3⩾(b2c2−1)2, and so 2b3⩾b3+c3⩾(b2c2−1)2>b4c4−2b2c2⩾b4c4−2b3c4 Thus, (b−2)c4<2, and the only solutions to this inequality have (b,c)=(2,2) or b⩽3 and c=1. It is easy to verify that the only case giving a solution for a⩾b is (a,b,c)=(3,2,1).
Otherwise, suppose that b3+c3=f(a)⩾f(b). Then, we have 2b3⩾b3+c3=a2(b2c2−a)⩾b2(b2c2−b) Consequently bc2⩽3, with strict inequality in the case that b=c. Hence c=1 and b⩽2. Both of these cases have been considered already, so we are done.
Want a route through all this instead of an archive? The track
puts 2,000 problems in a working order, from AMC 10 level to the IMO shortlist.
Source: MathNet,
licensed CC-BY-4.0.
Statement and solution reproduced as published; topic and difficulty added by this site.