A triple summation with a restriction!

Algebra Level 5

i j , j k , k i 1 2 i 2 j 2 k = a b , \large \displaystyle \sum_{i\ne j, j\ne k, k\ne i } \frac{1}{2^i 2^j 2^k} = \frac{a}{b},
where i , j , k i,j,k are distinct whole numbers. If a , b a,b are coprime positive integers, determine the value of a + b a+b .

Clarifications on the summation
The summation runs over all triples ( i , j , k ) (i,j,k) of pairwise non-negative integers. As an example, ( 1 , 3 , 2 ) (1,3,2) and ( 1 , 0 , 8 ) (1,0,8) are allowed but ( 3 , 3 , 2 ) (3,3,2) , ( 0 , 0 , 0 ) (0,0,0) , ( 0 , 0 , 3 ) (0,0,3) and ( 7 , 7 , 7 ) (7,7,7) aren't.


The answer is 23.

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

Shaurya Gupta
Nov 16, 2015

We'll solve the problem using inclusion-exclusion principle.
We first calculate the value without the condition that i , j , k i,j,k are distinct. 1 2 i 2 j 2 k = 2 × 1 2 i 2 j = 2 × 2 × 1 2 i = 8 \sum\frac{1}{2^i2^j2^k} = \sum 2\times\frac{1}{2^i 2^j} = \sum 2\times 2\times\frac{1}{2^i} = 8
Now we need to remove the extra value in our sum by removing all those cases which do not satisfy our condition. We consider the case that two of our variables are equal without any restriction on the third variable.
1 2 i 2 j 2 k = 1 2 2 i 2 k = 2 × 1 2 2 i = 2. 4 3 = 8 3 \sum \frac{1}{2^i 2^j 2^k} = \sum \frac{1}{2^{2i} 2^k}= \sum 2\times \frac{1}{2^{2i}} = 2.\frac{4}{3} =\frac{8}{3}
There will be 3 3 such cases, namely i = j , j = k and k = i i=j, j=k \text{ and } k=i . So we need to subtract three times this value from our previously calculated value. But doing so, we've also removed the case when all three variables are equal thrice. Since we want to remove this case only once, we add twice its value. 1 2 i 2 j 2 k = 1 2 3 i = 8 7 8 3 × 8 3 + 2 × 8 7 = 16 7 \sum \frac{1}{2^i 2^j 2^k} = \sum \frac{1}{2^3i} = \frac{8}{7} \\ 8 - 3\times\frac{8}{3}+ 2\times\frac{8}{7} = \frac{16}{7}
Since 16 , 7 16,7 are co-prime, the answer will be 23 \boxed{23} .


Wow! I totally forgot about inclusion exclusion! Great

I wonder how to solve for n n for a given A = 2 ( a 1 + a 2 + + a n ) \large A = \sum 2^{-(a_1 + a_2+\cdots +a_n) } for pairwise distinct a i a_i 's.

Pi Han Goh - 5 years, 6 months ago
Mark Hennings
Feb 2, 2016

This is something of a special case of Euler's work on the generating function of the partition function P P .

Let us let S ( n ) S(n) be the sum S ( n ) = a 1 , a 2 , , a n 0 a 1 , a 2 , , a n distinct 2 ( a 1 + a 2 + + a n ) = n ! 0 a 1 < a 2 < < a n 2 ( a 1 + a 2 + + a n ) , S(n) \; =\; \sum_{{a_1,a_2,\ldots,a_n \ge0} \atop {a_1,a_2,\ldots,a_n \text{ distinct}}} 2^{-(a_1+a_2+\cdots+a_n)} \;= \; n! \sum_{0 \le a_1 < a_2 < \cdots < a_n} 2^{-(a_1+a_2+\cdots+a_n)} \;, Then S ( n ) = n ! N 1 2 n ( n 1 ) A ( N , n ) 2 N , S(n) \; = \; n! \sum_{N \ge \frac12n(n-1)} \big|A(N,n)\big| 2^{-N} \;, where A ( N , n ) A(N,n) is the set A ( N , n ) = { a N { 0 } n a 1 < a 2 < < a n , j = 1 n a j = N } , N 1 2 n ( n 1 ) . A(N,n) \; = \; \left\{ \mathbf{a} \in \mathbb{N}\cup\{0\}^n \; \Big| \; a_1 < a_2 < \cdots < a_n \;, \; \sum_{j=1}^n a_j \,=\, N \right\} \;, \; N \ge \tfrac12n(n-1) \;. If we also consider the set B ( L , n ) = { b N { 0 } n j = 1 n ( n + 1 j ) b j = L } , L 0 , B(L,n) \; = \; \left\{ \mathbf{b} \in \mathbb{N}\cup\{0\}^n \; \Big| \; \sum_{j=1}^n (n+1-j)b_j \,=\, L\right\}\;, \qquad L \ge 0 \;, then the mapping X : B ( N 1 2 n ( n 1 ) , n ) A ( N , n ) X \,:\, B\big(N-\tfrac12n(n-1),n\big) \,\to\, A(N,n) given by the formula ( X b ) j = j 1 + r = 1 j b r 1 j n , b B ( N 1 2 n ( n 1 ) , n ) (X\mathbf{b})_j \; = \; j-1 + \sum_{r=1}^j b_r \qquad 1 \le j \le n \;, \qquad \qquad \mathbf{b} \in B\big(N-\tfrac12n(n-1),n\big) is a bijection for any N 1 2 n ( n 1 ) N \ge \tfrac12n(n-1) , and hence A ( N , n ) = B ( ( N 1 2 n ( n 1 ) , n ) \big|A(N,n)\big| \,=\, \big|B(\big(N-\tfrac12n(n-1),n\big)\big| .

But B ( L , n ) \big|B(L,n)\big| is the coefficient of t L t^L in the series expansion of the product r = 1 n ( 1 t r ) 1 , \prod_{r=1}^n \big(1 - t^r\big)^{-1} \;, and hence A ( N , n ) \big|A(N,n)\big| is the coefficient of t N t^N in the series expansion of the product Q n ( t ) = t 1 2 n ( n 1 ) r = 1 n ( 1 t r ) 1 = r = 1 n ( t r 1 1 t r ) . Q_n(t) \; = \; t^{\frac12n(n-1)}\prod_{r=1}^n \big(1 - t^r\big)^{-1} \; = \; \prod_{r=1}^n \left(\frac{t^{r-1}}{1-t^r}\right) \;. Thus it follows that S ( n ) = n ! Q n ( 1 2 ) = n ! r = 1 n 2 2 r 1 = n ! 2 n ( r = 1 n ( 2 r 1 ) ) 1 . S(n) \; = \; n! Q_n(\tfrac12) \; =\; n! \prod_{r=1}^n \frac{2}{2^r-1} \; = \; n!2^n \left(\prod_{r=1}^n (2^r-1)\right)^{-1} \;. In this case, we have S ( 3 ) = 3 ! × 2 3 1 × 3 × 7 = 16 7 , S(3) \; =\; \frac{3! \times 2^3}{1\times3\times7} \; =\; \frac{16}{7} \;, making the answer 16 + 7 = 23 16+7 = \boxed{23} .

Pi Han Goh - 5 years, 4 months ago

Log in to reply

  1. Every term 2 a 1 a 2 a n 2^{-a_1-a_2-\cdots-a_n} with a 1 + a 2 + + a n = N a_1+a_2+\cdots +a_n = N contributes 2 N 2^{-N} to the sum. The number A ( N , n ) |A(N,n)| just counts how many different times 2 N 2^{-N} occurs in the sum. Basically, I am collecting terms.

  • Yes. N { 0 } n \mathbb{N}\cup\{0\}^n is the set of n n -tuples of nonnegative integers.

  • A mapping is another word for a function. Since I am saying that X X is a bijection, you can read this whole bit as saying that X X sets up a 1-1 correspondence between the two sets, if that is a more familiar expression to you.

  • Remember, X X is a function between sets of n n -tuples. If we start with the n n -tuple b \mathbb{b} in B ( N 1 2 n ( n 1 ) , n ) B(N-\tfrac12n(n-1),n) , then X b X\mathbf{b} is the n n -tuple (call it a \mathbf{a} ) whose coefficients are a j = j 1 + r = 1 j b r , 1 j n . a_j \; = \; j - 1 + \sum_{r=1}^j b_r \;, \qquad 1 \le j \le n \;. I am using the standard notational abbreviation a \mathbf{a} to represent the n n -tuple ( a 1 , a 2 , , a n ) (a_1,a_2,\ldots,a_n) , and similarly for b \mathbf{b} . You need to check that X b X\mathbf{b} actually belongs to A ( N , n ) A(N,n) (that is the point of X X ) and that X X has an inverse, so that every element of A ( N , n ) A(N,n) is equal to X b X\mathbf{b} for a unique element b \mathbf{b} in B ( N 1 2 n ( n 1 ) , n ) B(N-\tfrac12n(n-1),n) . You retrieve b \mathbf{b} from a = X b \mathbf{a} = X\mathbf{b} by the formula b j = { a 1 j = 1 a j a j 1 1 2 j n . b_j \; = \; \left\{ \begin{array}{lll} a_1 & \qquad \qquad j = 1 \\ a_j - a_{j-1} - 1 & \qquad \qquad 2 \le j \le n. \end{array} \right.

  • Mark Hennings - 5 years, 4 months ago

    Log in to reply

    Thank you for your reply. Even if I accept that the function A ( N , n ) A(N,n) is carefully constructed, I still couldn't for the life of me figure out how you managed to pull the function of B ( L , n ) B(L,n) out of nowhere. That's some outstanding detective work!

    Will read your solution again once my head is not about to explode.

    Pi Han Goh - 5 years, 4 months ago

    Log in to reply

    @Pi Han Goh There are standard combinatoric tricks for converting sets of strictly increasing n n -tuples into sets of just increasing n n -tuples, or even (as I did in this case) into sets of n n -tuples of nonnegative numbers. Generally speaking, it is easier to identify what a set of n n -tuples is doing if there are as few restrictions on the shape of its elements as possible; going from a 1 < a 2 < < a n a_1 < a_2 < \cdots < a_n to b 1 , b 2 , b n 0 b_1,b_2, \ldots b_n \ge 0 was a big simplification, and it seemed to me worth the while (to be more accurate, I have read about how to derive the generating function for the partition function P P , and this calculation is similar), to investigate what would happen if we made this conversion. As you can see, it worked!

    Where did I get B ( L , n ) B(L,n) from? It was forced on me. The formula b j = { a 1 j = 1 a j a j 1 1 2 j n . b_j \; = \; \left\{ \begin{array}{lll} a_1 & \qquad \qquad & j = 1 \\ a_j - a_{j-1} - 1 && 2 \le j \le n \;. \end{array} \right. is one of those standard tricks for converting an n n -tuple a \mathbf{a} with a 1 < a 2 < < a n a_1 < a_2 < \cdots < a_n into an n n -tuple b \mathbf{b} with b 1 , b 2 , , b n 0 b_1,b_2,\ldots,b_n \ge 0 . Check out what the requirement that the components of a \mathbf{a} add to N N has on the components of b \mathbf{b} , and out pops the sum requirement of B ( L , n ) B(L,n) , for L = N 1 2 n ( n 1 ) L = N - \tfrac12n(n-1) .

    Mark Hennings - 5 years, 4 months ago

    I have so many questions and I don't know where to begin, so let me just ask whatever that appears the most confusing:


    Question 01 : How did you convert this first S ( n ) S(n) to the second S ( n ) S(n) ?

    S ( n ) = a 1 , a 2 , , a n 0 a 1 , a 2 , , a n distinct 2 ( a 1 + a 2 + + a n ) = n ! 0 a 1 < a 2 < < a n 2 ( a 1 + a 2 + + a n ) , S(n) \; =\; \sum_{{a_1,a_2,\ldots,a_n \ge0} \atop {a_1,a_2,\ldots,a_n \text{ distinct}}} 2^{-(a_1+a_2+\cdots+a_n)} \;= \; n! \sum_{0 \le a_1 < a_2 < \cdots < a_n} 2^{-(a_1+a_2+\cdots+a_n)} \;,

    to

    S ( n ) = n ! N 1 2 n ( n 1 ) A ( N , n ) 2 N , S(n) \; = \; n! \sum_{N \ge \frac12n(n-1)} \big|A(N,n)\big| 2^{-N} \;,

    (Yes, I've read your definition of A ( N , n ) = A(N,n) = \ldots already and I'm still clueless)


    Question 02 : what does a N { 0 } n \mathbf{a} \in \mathbb{N}\cup\{0\}^n means? Does it mean a vector called a a , with all non-negative integer elements right?


    Question 03 : Can you explain to me what this means? Mapping? Do you mean a one-to-one relation?

    then the mapping X : B ( N 1 2 n ( n 1 ) , n ) A ( N , n ) X \,:\, B\big(N-\tfrac12n(n-1),n\big) \,\to\, A(N,n) given by the formula


    Question 04 : What does ( X b ) j = j 1 + r = 1 j b r \displaystyle(Xb)_j \; = \; j-1 + \sum_{r=1}^j b_r means? What ( X b ) j (Xb)_j means exactly? I didn't see this notation before in your solution.

    Pi Han Goh - 5 years, 4 months ago

    i dont know why i am getting 16/3

    Akash singh - 5 years, 4 months ago

    Log in to reply

    Unless you explain how you are getting 16 3 \tfrac{16}{3} , I cannot explain what you have done wrong!

    Mark Hennings - 5 years, 4 months ago

    Log in to reply

    oops! i did a calculation error

    Akash singh - 5 years, 4 months ago

    Hey, can you please give an easy explanation for maybe a high school grade math student (me)?

    I didn't get the step from where you changed the sum S(n) to something in |A(N, n)|. The rest is pretty overwhelming!

    Tapas Mazumdar - 3 years, 5 months ago

    Log in to reply

    Have a look at the reply I gave to Pi Han almost two years ago. That should explain things.

    Mark Hennings - 3 years, 5 months ago

    0 pending reports

    ×

    Problem Loading...

    Note Loading...

    Set Loading...