Let a , b , c be distinct positive integers such that a + b + c ≤ 3 0 0 0 0 0 0 . Find the maximum value of gcd ( a b + 1 , b c + 1 , c a + 1 ) .
Notation:
g
cd
(
⋅
)
denotes the
greatest common divisor
function.
This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try
refreshing the page, (b) enabling javascript if it is disabled on your browser and,
finally, (c)
loading the
non-javascript version of this page
. We're sorry about the hassle.
I did exactly what you did up until the "at this point" part.
Writing n = k a 2 + 1 , we get a = n k − 1 . So n + n k − 1 ≤ 1 0 0 0 0 0 0 .
For k = 1 , it's easy to check that n = 9 9 9 2 + 1 = 9 9 8 0 0 2 is the max.
For k = 2 , it's easy to check that n = 2 1 4 1 3 2 + 1 = 9 9 8 2 8 5 is the max.
For k ≥ 3 , we get n + 3 n − 1 ≤ n + n k − 1 ≤ 1 0 0 0 0 0 0 . The left side is an increasing function of n , and if we evaluate at n = 9 9 8 2 8 5 we get 1 0 0 0 0 1 5 . 5 6 … , so any value of n that corresponds to k ≥ 3 is smaller than 9 9 8 2 8 5 .
So we're done.
We can assume that a > b > c , let b − c = x , a − b = y . We have 3 c + 2 x + y ≤ 3 0 0 0 0 0 0 .
Let d = gcd ( a b + 1 , b c + 1 , c a + 1 ) , we get gcd ( d , a ) = gcd ( d , b ) = gcd ( d , c ) = 1 .
So d ∣ a − c , a − b ⇒ d ∣ gcd ( x , y ) .
Also, d ∣ c 2 + 1 . We get d ≤ c 2 + 1 , x , y ⇒ 3 0 0 0 0 0 0 ≥ 3 c + 3 d ≥ 3 d − 1 + 3 d .
This gives us d ≤ 9 9 9 0 0 1 .
If d = c 2 + 1 , we have d ≤ 9 9 9 2 + 1 = 9 9 8 0 0 2 .
If d = 2 c 2 + 1 , we get 3 0 0 0 0 0 0 ≥ 3 2 d − 1 + 3 d . This gives us d ≤ 9 9 8 5 8 7 , so d ≤ 2 1 4 1 3 2 + 1 = 9 9 8 2 8 5 .
If d ≤ 5 c 2 + 1 , we get 3 0 0 0 0 0 0 ≥ 3 5 d − 1 + 3 d . This gives us d ≤ 9 9 7 7 6 6 .
So, the maximum value of d is 9 9 8 2 8 5 . Equality holds at c = 1 4 1 3 , x = y = 9 9 8 2 8 5 .
Problem Loading...
Note Loading...
Set Loading...
Without loss of generality, we can assume that a < b < c . If n is the greatest common divisor of a b + 1 , a c + 1 , b c + 1 , then since n divides a b + 1 it follows that n and a are coprime. Similarly n and b are coprime and n and c are coprime. Also note that n divides a ( b − c ) = ( a b + 1 ) − ( a c + 1 ) , and so n divides b − c . Similarly we can show that n divides a − b . Thus we must have b = a + u n , c = a + v n where u , v are positive integers with u < v . Since n divides a b + 1 we deduce that n divides a 2 + 1 .
Since n divides a 2 + 1 we deduce that n − 1 ≤ a . Since b ≥ a + n and c ≥ a + 2 n we deduce that 3 0 0 0 0 0 0 ≥ a + b + c ≥ 3 ( a + n ) , so that a ≤ 1 0 0 0 0 0 0 − n . Thus n must be such that n − 1 ≤ 1 0 0 0 0 0 0 − n , and hence n ≤ 2 1 ( 2 0 0 0 0 0 1 − 3 9 9 9 9 9 7 ) = 9 9 9 0 0 0 . 5 0 0 4 , so that n ≤ 9 9 9 0 0 0 . We note that n = 9 9 9 2 + 1 = 9 9 8 0 0 2 a = 9 9 9 b = a + n = 9 9 9 0 0 1 c = a + 2 n = 1 9 9 7 0 0 3 is a particular solution, and hence 9 9 8 0 0 2 ≤ n ≤ 9 9 9 0 0 0 .
At this point we only need conduct a computer search to find the largest value of n giving a solution to the problem (there are fewer than 1 0 0 0 cases to check): n = 9 9 8 2 8 5 a = 1 4 1 3 b = a + n = 9 9 9 6 9 8 c = a + 2 n = 1 9 9 7 9 8 3