Inspired by Problem 4, IMO 2006

p 3 a + 1 + p a + 1 = b p \large p^{3a+1}+p^a+1=b^p

Find all solutions to the Diophantine equation above, where a a and b b are positive integers and p p is an odd prime, and enter your answer as ( a + b + p ) \sum (a+b+p) .


The answer is 18.

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.

2 solutions

Kazem Sepehrinia
May 12, 2017

At first sight, there is no bound for variables. If you fix p p , for example, then how large a a might be, you can pick b b such that RHS balances LHS of the equation. Let us find a relation between b b and p p . This is reasonable because they are both bases. Then we might be able to bound one of variables.

Observe that gcd ( b , p ) = 1 \gcd(b, p)=1 . Look mod p p , we get b p 1 ( m o d p ) b^p \equiv 1 \pmod{p} and b p 1 b b 1 ( m o d p ) b^{p-1}b \equiv b \equiv 1 \pmod{p} , using FLT . Thus, p b 1 p|b-1 and we can apply LTE lemma in the following form.

Rearrange the equation to get p a ( p 2 a + 1 + 1 ) = b p 1. p^a(p^{2a+1}+1)=b^p-1.

Note that the highest power of p p dividing LHS is a a , hence a = v p ( b p 1 ) = v p ( b 1 ) + v p ( p ) = v p ( b 1 ) + 1 a=v_p(b^p-1)=v_p(b-1)+v_p(p)=v_p(b-1)+1 So v p ( b 1 ) = a 1 v_p(b-1)=a-1 and there is an even integer c c , relatively prime to p p , such that b = p a 1 c + 1 b=p^{a-1}c+1 . Now we have the following equation p 3 a + 1 + p a + 1 = ( p a 1 c + 1 ) p ( 1 ) p^{3a+1}+p^a+1=(p^{a-1}c+1)^p \qquad (1)

Now, from this equation, you see that the RHS would be greater than the LHS for almost any value of a a and p p and we have found our way to bound variables. We are going to narrow it down fairly quickly, i.e. prove that for most values RHS > LHS, and see what's left?

If a = 1 a=1 , then b 3 b \ge 3 and original equation becomes p 4 + p + 1 = b p ( 2 ) p^4+p+1=b^p \qquad(2) which results in p 4 + p + 1 = b p 3 p p^4+p+1=b^p \ge 3^p and p < 8 p<8 , but none of p = 3 , 5 , 7 p=3, 5, 7 give an integer for b b from equation ( 2 ) (2) .

Therefore, a > 1 a>1 . According to the binomial theorem, RHS of the equation ( 1 ) (1) consists of a lot of terms with powers of p p greater than or equal to a a . Therefore, there mustn't be a power of p p in the RHS greater than or equal to the highest power of p p in the RHS, which is 3 a + 1 3a+1 . But note that for p 5 p \ge 5 and a > 1 a>1 , since c 2 c\ge2 and c p > p 2 c^p>p^2 the first term of binomial expansion becomes p ( a 1 ) p c p > p ( a 1 ) p p 2 = p ( a 1 ) p + 2 p^{(a-1)p}c^p>p^{(a-1)p}p^2=p^{(a-1)p+2} and ( a 1 ) p + 2 5 a 3 3 a + 1 (a-1)p+2\ge 5a-3 \ge 3a+1 . This is a contradiction and leads to p < 5 p<5 and the only remaining value p = 3 p=3 . Thus, equation ( 1 ) (1) becomes 3 3 a + 1 + 3 a + 1 = ( 3 a 1 c + 1 ) 3 = 3 3 a 3 c 3 + 3 a c ( 3 a 1 c + 1 ) + 1 3^{3a+1}+3^a+1=(3^{a-1}c+1)^3=3^{3a-3}c^3+3^a c(3^{a-1}c+1)+1 or 3 2 a 3 ( 81 c 3 ) = c ( 3 a 1 c + 1 ) 1 3^{2a-3} (81-c^3)=c(3^{a-1}c+1)-1 LHS is positive. It follows that 81 c 3 > 0 81-c^3>0 and c 4 c\le 4 . Note that c c is even and, thus, c = 2 , 4 c=2, 4 . c = 2 c=2 gives 77 × 3 2 a 3 = 4 × 3 a 1 + 1 , 77\times 3^{2a-3}=4\times 3^{a-1}+1, which is impossible, because LHS is always greater than RHS for a > 1 a>1 . For c = 4 c=4 we get 17 × 3 2 a 3 = 16 × 3 a 1 + 3. 17\times 3^{2a-3}=16\times 3^{a-1}+3. For a > 2 a>2 LHS of this equation becomes greater than its RHS and a = 2 a=2 satisfies the equation. Therefore, the only solution to the given equation is ( a , b , p ) = ( 2 , 13 , 3 ) (a, b, p)=(2, 13, 3) .

L e t f ( p , a , b ) = p 3 a + 1 + p a + 1 b p = 0. T r y w i t h s m a l l e s t p = 3. T h e n f ( 3 , a , b ) = 3 3 a + 1 + 3 a + 1 = b 3 . F o r f ( 3 , 1 , b ) = 85 b u t t h i s i s n o t a c u b e o f a n y i n t e g e r . F o r f ( 3 , 2 , b ) = 2197 , 2197 3 = 13 , a n i n t e g e r . 3 + 2 + 13 = 18. Let ~~f(p,a,b)=p^{3a+1} +p^a+1-b^p=0.\\ Try~ with~ smallest~ p=3.\\ Then~f(3,a,b)=3^ {3a+1}+3^a+1=b^3.\\ For~f(3,1,b)=85~~but~ this~ is~ not ~a ~cube~ of~ any~integer.\\ For~f(3,2,b)=2197, \sqrt[3]{2197}=~13,~~an~ integer.\\ 3+2+13=\huge~~\color{#D61F06}{18}.

This is how I got the result. But it lacks the proof that this is the only solution.
Interesting observation.
( 5 25 + 5 8 + 1 ) 312 5 5 = 0 (5^{25}+5^8+1)-3125^5=0 if we use accuracy up to 10 digits. Obviously it is not true since the term in the (.... ) has unit digit as 6 while the negative term has 5.
But when accuracy is 14 , we get that ( 5 25 + 5 8 + 1 ) 312 5 5 = 391000. (5^{25}+5^8+1)-3125^5=391000. And that changes in the last four digits as we go in for more accuracy.


0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...