Don't trip

Let S S be the set of all distinct Pythagorean triples ( a , b , c ) (a,b,c) , where a , b , c a,b,c are positive integers with a < b < c a \lt b \lt c , such that a b = 6 ( a + b + c ) . ab = 6(a + b + c).

Find the sum of all distinct values for c c of the elements of S . S.


The answer is 267.

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.

5 solutions

The side lengths will be of the form k ( m 2 n 2 ) , 2 k m n k(m^{2} - n^{2}), 2kmn and k ( m 2 + n 2 ) k(m^{2} + n^{2}) for positive integers k , m , n k,m,n with m > n . m \gt n. Now we don't know at this point which of the first two of these terms will be the greater, so we can't say which is a a and which is b b , but since the condition we must adhere to, namely a b = 6 ( a + b + c ) ab = 6(a + b + c) , is symmetric for a , b a,b we don't need to know which term is a a and which is b b until the end of the solution. We do know, however, that the third term will be c . c.

Now regardless of which term is a a and which is b b , the given condition becomes

2 k 2 m n ( m 2 n 2 ) = 6 k ( 2 m 2 + 2 m n ) 2k^{2}mn(m^{2} - n^{2}) = 6k(2m^{2} + 2mn)

k n ( m n ) ( m + n ) = 6 ( m + n ) k n ( m n ) = 6. \Longrightarrow kn(m - n)(m + n) = 6(m + n) \Longrightarrow kn(m - n) = 6.

Now we need to look in turn at the cases of k = 1 , 2 , 3 , 6 k = 1,2,3,6 and then determine the possible values for m , n . m,n. After these values are determined, we can then determine the associated triples, gather those that are distinct, and then add the c c values to obtain the desired solution. We have the following cases:

  • k = 1 ( m , n ) = ( 7 , 1 ) , ( 5 , 2 ) , ( 5 , 3 ) , ( 7 , 6 ) k = 1 \Longrightarrow (m,n) = (7,1), (5,2), (5,3), (7,6) ;

  • k = 2 ( m , n ) = ( 4 , 1 ) , ( 4 , 3 ) k = 2 \Longrightarrow (m,n) = (4,1), (4,3) ;

  • k = 3 ( m , n ) = ( 3 , 1 ) , ( 3 , 2 ) k = 3 \Longrightarrow (m,n) = (3,1), (3,2) ;

  • k = 6 ( m , n ) = ( 2 , 1 ) . k = 6 \Longrightarrow (m,n) = (2,1).

Plugging all these values in to our original side length expressions and eliminating the duplicate triples, we find that

S = { ( 14 , 48 , 50 ) , ( 20 , 21 , 29 ) , ( 16 , 30 , 34 ) , ( 13 , 84 , 85 ) , ( 18 , 24 , 30 ) , ( 15 , 36 , 39 ) } . S = \{(14,48,50), (20,21,29), (16,30,34), (13,84,85), (18,24,30), (15,36,39)\}.

The sum of all hypotenuse values is then 50 + 29 + 34 + 85 + 30 + 39 = 267 . 50 + 29 + 34 + 85 + 30 + 39 = \boxed{267}.

I couldn't pursue some approaches which I suspect can also help solve this problem. It's also good to know why a solution doesn't work.

Since a b = 6 ( a + b + c ) ab=6\left( a+b+c \right) is symmetric in a a and b b . We can get another equation as a 2 + b 2 = c 2 { a }^{ 2 }+{ b }^{ 2 }={ c }^{ 2 } which is also symmetric in a a and b b . Using these two equations can we find a polynomial equation in c c as c n + S 1 c n 1 + S 2 c n 2 + S 3 c n 3 . . . + S n c 0 = 0 { c }^{ n }+{ S }_{ 1 }{ c }^{ n-1 }+{ { S }_{ 2 }c }^{ n-2 }+{ S }_{ 3 }{ c }^{ n-3 }...+{ S }_{ n }{ c }^{ 0 }=0 ? Then the sum of all the c s c's will be S 1 -{ S }_{ 1 } i.e., sum of all the roots taken one at a time.

We can also try a c = sin θ , b c = cos θ \displaystyle \cfrac { a }{ c } =\sin { \theta } ,\cfrac { b }{ c } =\cos { \theta } . With help of trigonometry and co-ordinate geometry can we arrive at a solution?

Both the approaches may not yield a solution. I tried both but couldn't go far enough. Plz look into it in your free time. It may be waste of time.

Soumo Mukherjee - 6 years, 2 months ago

Log in to reply

For the first approach, while the equations are indeed symmetric in a a and b b , the two must in the end have different values, so I can't, for the moment, see how we will be able to get a workable polynomial in c c . I like the concept, but the execution of that concept may be more trouble than it's worth.

The second concept is interesting too. We would end up with the equation

sin ( 2 θ ) = 12 ( 1 + sin ( θ ) + cos ( θ ) ) , \sin(2\theta) = 12*(1 + \sin(\theta) + \cos(\theta)),

which I believe would only have solutions when both sides equal 0. 0. This would end up requiring one of sin ( θ ) \sin(\theta) or cos ( θ ) \cos(\theta) being 1 -1 and the other 0 0 , which would leave one of a a or b b to be 0 0 , which is not promising.

So while these both are good ideas, they don't appear to give us the desired results, (at least not easily). So for know the algebraic approach will have to do. :)

Brian Charlesworth - 6 years, 2 months ago

Log in to reply

Thanks...

I also revisited the approaches. In the first approach, we need to eliminate a a and b b to get an equation in c c . Which may not be possible because we need three condition to eliminate two variables. So the first approach must be put to hold, till other alternatives are found. I didn't have much faith in the 2nd approach, I just noticed the form a 2 + b 2 = c 2 { a }^{ 2 }+{ b }^{ 2 }={ c }^{ 2 } and was trying to reduce the variables somehow.

Did you see this one . I think there is some implicit weakness in the problem which can help to tackle it in some different way. For e.g. take the two red-squares to the opposite end of the diagonals. Then remove the squares. We then get a board with no diagonally opposite corners. Also notice the position of the squares, both are interchangeable. There could be some more factors to observe. After all "A problem's weak points are in the eye of the beholder." And I am a blind today ;( .... :D

Soumo Mukherjee - 6 years, 2 months ago

Log in to reply

@Soumo Mukherjee Yes, I'm all to familiar with that question. I calculated the value that is now considered the correct and was told it was incorrect, as Kevin had originally posted 249 as the answer. I then proceeded trying six other ways of solving the problem, all of which resulted in my original answer. I figured I must be wrong because over 50 people had already answered "correctly", so I never reported the problem. I should have, because my answer was correct all along, (and I got credit for it once the posted answer was changed). The silver lining, I suppose, is that I learned that just because dozens of people come to a similar answer doesn't necessarily mean that the answer they agree on is correct. :P

Brian Charlesworth - 6 years, 2 months ago

Log in to reply

@Brian Charlesworth 6 ways... wow

But, you solved it. Combinatorics seems very hard to apply. I always make the wrong application. I will keep in mind moral of the story ;) Til then ... waiting for awesome problems with funny names from Brian Charlesworth :D

Soumo Mukherjee - 6 years, 2 months ago
Eilon Lavi
Mar 24, 2015

36 ( a + b ) 2 = ( 6 c ) 2 = ( a b 6 ( a + b ) ) 2 36 ( a + b ) 2 = a 2 b 2 12 a b ( a + b ) + 36 ( a 2 + 2 a b + b 2 ) 0 = a 2 b 2 12 a b ( a + b ) + 72 a b 0 = a b 12 ( a + b ) + 72 36(a+b)^2=(6c)^2=(ab-6(a+b))^2 \\36(a+b)^2=a^2b^2-12ab(a+b)+36(a^2+2ab+b^2) \\0=a^2b^2-12ab(a+b)+72ab \\0=ab-12(a+b)+72

Now:

( a b ) 2 = ( a + b ) 2 4 a b = ( a + b ) 2 48 ( a + b ) + 288 = ( a + b 24 ) 2 288 ( a + b + 24 ) 2 ( a b ) 2 = 288 ( 2 a 24 ) ( 2 b 24 ) = 288 ( a 12 ) ( b 12 ) = 72 (a-b)^2=(a+b)^2-4ab=(a+b)^2-48(a+b)+288=(a+b-24)^2-288 \\(a+b+24)^2-(a-b)^2=288 \\(2a-24)(2b-24)=288 \\(a-12)(b-12)=72

And from here, we can check all factors of 72 = 2 3 3 2 72=2^3 \cdot 3^2 . Here is a shortcut, though arguably unwarranted:

12 ( a + b ) 72 = a b = 6 ( a + b + c ) a + b 12 = c 12(a+b)-72=ab=6(a+b+c) \\a+b-12=c

0 < j < k , j k = 72 ( j + 12 ) + ( k + 12 ) 12 = 0 < j < k , j k = 72 j + k + 12 = ( 0 < j < k , j k = 72 j + k ) + ( 0 < j < k , j k = 72 12 ) = 0 < k 72 k + 12 ( N f a c t o r s o f 72 / 2 ) = ( 1 + 2 + 4 + 8 ) ( 1 + 3 + 9 ) + 12 ( ( 3 + 1 ) ( 2 + 1 ) / 2 ) = 15 13 + 12 ( 6 ) = 267 \sum\limits_{0<j<k,jk=72} (j+12)+(k+12)-12 \\=\sum\limits_{0<j<k,jk=72} j+k+12 \\=(\sum\limits_{0<j<k,jk=72} j+k) +( \sum\limits_{0<j<k,jk=72} 12) \\=\sum\limits_{0<k|72} k +12 (N_{factors-of-72}/2) \\=(1+2+4+8)(1+3+9)+12((3+1)(2+1)/2) \\=15\cdot 13+12(6) \\=267

I did the same way.

Ronak Agarwal - 6 years, 1 month ago
Andrea Palma
Mar 29, 2015

Since c = a 2 + b 2 c = \sqrt{a^2 + b^2} the equation is

a b = 6 ( a + b + a 2 + b 2 ) ab = 6\left(a + b + \sqrt{a^2 + b^2} \right)

a b 6 a 6 b = 6 a 2 + b 2 ab -6a - 6b = 6\sqrt{a^2 + b^2}

Setting a b > 6 ( a + b ) ab > 6(a+b) we can square the sides and get

a b ( a b 12 a 12 b + 72 ) = 0 ab(ab - 12a - 12b + 72) = 0

This means a b 12 a 12 b + 72 = a ( b 12 ) ( 12 b 72 ) = 0 ab -12a - 12 b + 72 = a(b - 12) -(12b - 72) = 0

So ( b 12 ) ( 12 b 72 ) = ( 12 b 144 ) + 72 ( b 12 ) 72 = 2 3 3 2 (b -12) \vert (12b - 72) = (12b - 144 )+ 72 \ \Rightarrow \ (b - 12) \vert 72 = 2^33^2 .

Keeping in mind that b > a = 12 + 72 b 12 > 0 b > a = 12 + \dfrac{72}{b-12} > 0 we get the following acceptable values for b b , and consequently values for a a , and hence for c c

( a , b , c ) = { ( 18 , 24 , 30 ) ( 15 , 36 , 39 ) ( 20 , 21 , 29 ) ( 16 , 30 , 34 ) ( 14 , 48 , 50 ) ( 13 , 84 , 85 ) \displaystyle{(a,b,c) = \cases{ (18,24,30) \\ (15,36,39) \\ (20,21,29) \\ (16,30,34) \\ (14,48,50) \\ (13,84,85) }}

Summing the c c s we get the right answer.

Rushikesh Joshi
Mar 27, 2015

My soln is not as fancy as yours... Observe A(ABC)÷s=6(where s is semiperimeter). But A(ABC)÷s=r (where r is inradius) We know that for right angled triangle 2r=a+b+c. So a+b+c=12. So b >a> 12. Also we can get an equation in (b-a) as

(b-a)^2 + (12 - c)(b-a) + 72 = 0

put condition for (b-a) is real. we get c^2 - 24c - 144 > 0. that is c>28.97 Check value of a from 13 ,14, 15, ... So the values of c are 85 50 39 34 30 29.

D G
Mar 24, 2015

b = 6 a 6 c 6 a b = \frac{-6a - 6c}{6 - a}

a 2 c 2 + b 2 = a 2 c 2 + ( 6 a 6 c 6 a ) 2 = 0 a^2 - c^2 + b^2 = a^2 - c^2 + {(\frac{-6a - 6c}{6 - a}})^2 = 0

a 6 , ( a 2 c 2 ) ( 6 a ) 2 + ( 6 a 6 c ) 2 = 0 a \neq 6, (a^2 - c^2)(6-a)^2 + (-6a - 6c)^2 = 0

a ( a + c ) ( 72 12 a + 12 c a c + a 2 ) = 0 a(a + c)(72 - 12a + 12c - ac + a^2) = 0

a = 0 a = 0 : breaks constraints

a = c a = -c : breaks constraints

72 12 a + 12 c a c + a 2 = 0 72 - 12a + 12c - ac + a^2 = 0 : Diophantine equation which can be solved using standard methods, and then filtered to fit the constraints.

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...