Consider the set K of all positive integers k such that 6 k is a triangular number . Then if k ∈ K ∑ k 1 = C + 2 π ( tan A B π − tan A π − tan B π − ( cot A B π − cot A π − cot B π ) ) where A , B , C are integers, find ( A − B ) 2 + C 2 .
Notation:
cot
(
⋅
)
denotes the
cotangent 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.
If 6 k is a triangular number, it may be written as 6 k = 2 1 n ( n ± 1 ) . Double this to find the equation 1 2 k = n ( n ± 1 ) . Since both sides of this equation are multiples of 12, we have three possibilities
n is a multiple of 12, say n = 1 2 c . Then k = 1 2 c 2 ± c for c = 1 , 2 , … are all distinct solutions. Thus the sum contains the terms c = 1 ∑ ∞ 1 2 c 2 ± c 1 = c = 1 ∑ ∞ ( c ± 1 2 1 1 − c 1 ) .
n is a multiple of 6 but not of 12, say n = 6 c for odd c . Then k = 3 c 2 + 2 1 c . But since c is odd, this cannot be an integer.
n is a multiple of 4 but not of 12, say n = 4 a . Then n ± 1 = 4 a ± 1 must be a multiple of 3. But this implies that a = 3 c ∓ 1 . It is now easy to show that k = 1 2 c 2 ± 7 c + 1 . For c = 1 , 2 , … this gives distinct solutions; for c = 0 we find the additional solution k = 1 . Thus the sum contains the terms 1 + c = 1 ∑ ∞ 1 2 c 2 ± 7 c + 1 1 = 1 + c = 1 ∑ ∞ ( c ± 3 1 1 − c ± 4 1 1 ) .
Therefore the required sum is k ∈ K ∑ k 1 = 1 + c = 1 ∑ ∞ ( c + 1 2 1 1 + c − 1 2 1 1 − c 2 + c + 3 1 1 + c − 3 1 1 − c + 4 1 1 − c − 4 1 1 ) . Add some magic to find that this sum is actually equal to k ∈ K ∑ k 1 = 1 2 − π ( 3 4 + 1 ) , but for unknown reason we must express the term in brackets using a cotangent. Now cot π / 3 = 1 / 3 , so that k ∈ K ∑ k 1 = 1 2 − π ( 4 cot 3 π + 1 ) . We see that A = 1 2 , B = 4 , C = 3 , and A 2 + B 2 + C 2 = 3 2 + 4 2 + 1 2 2 = 1 6 9 .
Note: Since this solution was posted, the problem has been edited. The answer is still 169.
Hi, thanks for posting this nice solution. The reason for the cotengent being used derives from the way I dealt with the problem. Indeed, after some modular considerations, I came across a series of a kind that got me stuck but which involved a cotangent formula (for instance, in the first case of values of k , I retrieved the series k = 1 ∑ ∞ k 2 − z 2 1 = 2 z 1 ( z 1 − π cot ( π z ) ) , but in the following 3 cases the series to be solved was more challenging). However as a result, the original solution had a sum of cotangents as its expression, so maybe it might be better to edit the problem in this direction.
Log in to reply
Summing these series is not easy, and my solution was so messy that I didn't even bother posting the details. I like the cotangent series you found.
Log in to reply
In the end, I have posted my partial solution, highliting the series of which I lack expert knowledge
Problem Loading...
Note Loading...
Set Loading...
[This is not a complete solution. Indeed, in the following it is used a series identity that I found (the one coloured in blue ) of which I don't know a proof nor I have a link to one. All hints are welcome.]
A positive integer T is a triangular number if and only if 8 T + 1 is a perfect square. So, in the problem we have 4 8 k + 1 = x 2 , for x ∈ Z . Hence, x 2 ≡ 1 mod 4 8 and then k = 4 8 x 2 − 1 .
Now, each 0 ≤ b ≤ 4 7 such that b 2 ≡ 4 8 1 generates a set K b = { 4 8 ( b + 4 8 h ) 2 − 1 , h ∈ Z } of solutions for k . Since, moreover, every solution is contained in a set of this form, we have that K is the union of all K b . Let also S b denote the sum of all the reciprocals of elements of K b .
For two distinct b 1 , b 2 of this kind, K b 1 and K b 2 are not disjoint if and only if b 2 = 4 8 − b 1 . Indeed, if there is a common element, then exist h 1 , h 2 for which ( b 1 + 4 8 h 1 ) 2 = ( b 2 + 4 8 h 2 ) 2 , and so ± ( b 1 + 4 8 h 1 ) = b 2 + 4 8 h 2 . Thus b 1 ≡ 4 8 b 2 ∨ − b 1 ≡ 4 8 b 2 . But since b 1 = b 2 and they lie in the interval above, b 2 = 4 8 − b 1 . Conversely, given b 2 = 4 8 − b 1 , if an integer h determines by the respective formula an element of K b 1 or K b 2 , then we can plug − ( h + 1 ) in the other formula to see that the element also belongs to the other set. Moreover, notice that b 1 and b 2 actually generate the same set of solutions, therefore if K b 1 and K b 2 are not disjoint, they are equal.
Hence, we conclude that all integers b between 0 and 2 4 such that b 2 ≡ 4 8 1 provide pair-wise disjoint sets K b and that their union equals K . The possible values for b are 1 , 7 , 1 7 , 2 3 . Thus
S = k ∈ K ∑ k 1 = S 1 + S 7 + S 1 7 + S 2 3
Knowing that k = 1 ∑ ∞ k 2 − z 2 1 = 2 z 1 ( z 1 − π cot ( π z ) ) , we can evaluate
S 1 = 4 8 h = 1 ∑ ∞ ( 1 + 4 8 h ) 2 − 1 1 + ( 1 − 4 8 h ) 2 − 1 1 = 2 4 1 h = 1 ∑ ∞ h 2 − 2 4 2 1 1 = 1 2 − 2 1 π cot 2 4 π
While for b = 1 , S b = 4 8 h = − ∞ ∑ ∞ ( b + 4 8 h ) 2 − 1 1 = 2 π ( cot 4 8 π b − π − cot 4 8 π b + π )
Hence
S = 1 2 + 2 π ( − cot 2 4 π + cot 8 π − cot 6 π + cot 3 π − cot 8 3 π + cot 2 4 1 1 π − cot 2 π )
However, if α + β = 2 π , with α , β = k π , then cot α cot β = 1 and so cot α = tan β . For instance, cot 2 4 1 1 π = tan 2 4 π . So
S = 1 2 + 2 π ( tan 2 4 π − tan 8 π − tan 3 π − ( cot 2 4 π − cot 8 π − cot 3 π ) )
And ( A − B ) 2 + C 2 = ( 8 − 3 ) 2 + 1 2 2 = 1 6 9