Perfect AP

There is an arithmetic progression whose first term is 36, and the sum of the first n n terms of the AP is a perfect square for all positive integers n n .

What is the 100th term of this arithmetic progression?

Image credit: OEIS A000217 .


The answer is 7164.

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.

6 solutions

Eddie The Head
Apr 15, 2014

Solution. \textbf{Solution.} We will prove it in a more general case,let us take the 1 1 st term to be k 2 k^{2}

When n = 1 n=1 ,the sum of the first term is a perfect square = k 2 k^{2} .

The sum upto the n n -th term is S n = n 2 [ 2 a + ( n 1 ) d ] S_n = \frac{n}{2}[2a+(n-1)d] ,since these are all perfect squares we have [ 2 a + ( n 1 ) d ] > 0 [2a+(n-1)d] > 0 and hence d > 0 d>0

Now let us choose any odd prime p p . S p = p 2 [ 2 a + ( p 1 ) d ] S_p = \frac{p}{2}[2a+(p-1)d]

Since p S p p|S_p and S p S_p is a perfect square, p 2 S p p^{2} | S_p , p ( 2 a d ) + p d p|(2a-d)+pd p ( 2 a d ) p|(2a - d) This relation is true for all primes p p ,but since ( 2 a d ) (2a-d) is finite,we must have 2 a d = 0 2a - d = 0 and hence d = 2 a = 2 k 2 d = 2a = 2k^{2}

So the series is k 2 , 3 k 2 , 5 k 2 , . . . . . . . {k^{2},3k^{2},5k^{2},.......} .

In other words,the n n -th term is t n = ( 2 n 1 ) k 2 t_n = (2n-1)*k^{2} .

In this problem clearly k = 6 k = 6 ,So the 100 100 -th term of the series will be = ( 2 100 1 ) 6 2 = 7164 (2*100-1)*6^{2} = \boxed{7164}

What if after 36, we added 13, 15, 17, 19..., which would yield sums of 49, 64, 81, 100...? Wouldn't this series also met all the terms and conditions. Pun intended.

Nanayaranaraknas Vahdam - 7 years, 1 month ago

Log in to reply

Is it an ap?

Eddie The Head - 7 years, 1 month ago

Log in to reply

Sorry, my mistake. I didn't realise it.

Nanayaranaraknas Vahdam - 7 years, 1 month ago

Log in to reply

@Nanayaranaraknas Vahdam Sometimes it happens...

Eddie The Head - 7 years, 1 month ago

Ugg, dude, that's what I did. :O

Finn Hulse - 7 years, 1 month ago

U cannot add diff nos in an A.P..... In A.P, the difference b/w any 2 consecutive nos is always constant...... But in ur series, the difference is an A.P is not constant.........

Akshay Mujumdar - 7 years, 1 month ago

Please explain what is p|S

Pinak Wadikar - 7 years, 1 month ago

Log in to reply

It means p divides S

Eddie The Head - 7 years, 1 month ago

Cool problem!

Shreyansh Vats - 7 years, 1 month ago

Why should we have (2a-d)=0....

Sriram Vudayagiri - 7 years, 1 month ago

Log in to reply

We have proved that every even prime divided 2a-d which is only possible if it is infinite but actually it is a finite quantity so it must be 0.

Eddie The Head - 7 years, 1 month ago

But 7164 36 99 = 72 \frac{7164 - 36}{99} = 72 , which would be the number we add on to the sum S n S_n to get S n + 1 S_{n+1} , and 36 + 72 = 108 36 + 72 = 108 , which is not a perfect square.

Also, why can't 36 36 be an answer to this? If we start with S 1 = 36 S_1 = 36 (as in the problem), and add 0 0 to every term after, then the conditions stated in the problem would be satisfied, as the sum would always be 36 36 , which is a perfect square.

Milly Choochoo - 7 years, 1 month ago

Log in to reply

I recommend reading the question carefully once again....I said the sum of the 1 s t 1st n n terms should be a perfect square not all the terms individually

1. 1. Following my method we get the series 36 , 108 , 180 , . . . . {36,108,180,....}

Hence S 1 = 36 = 6 2 S_1 = 36 = 6^{2} S 2 = 36 + 108 = 144 = 1 2 2 S_2 = 36+108 = 144 = 12^{2} S 3 = 36 + 108 + 180 = 1 8 2 S_3 = 36+108+180 = 18^{2}

2. 2. Following your method we get the series 36 , 36 , 36 , . . . . . {36,36,36,.....}

Hence S 1 = 36 S_1 = 36 S 2 = 36 + 36 = 72 = ( 6 2 ) 2 S_2 = 36 + 36 = 72 = (6\sqrt{2})^{2} S 3 = 108 = ( 6 3 ) 2 S_3 = 108 = (6\sqrt{3})^{2} . None of which are perfect squares...

Firstly maybe you misunderstood the question and secondly your counter example is not true............

Eddie The Head - 7 years, 1 month ago

Log in to reply

Oh yes, I didn't read the question carefully. Woops

Milly Choochoo - 7 years, 1 month ago

Your solution is very nice but I have doubt in one step, How did you conjecture that p divides Sp i.e p s p p | sp \\

gaurav pathak - 7 years ago

That's the best solution I've seen so far.I too did it that way.Good problem with a nice solution.Upvoted&liked!!

rajdeep brahma - 4 years, 1 month ago
Alfred Lam
Apr 21, 2014

A relatively well-known fact about the AP of a sequence of odd numbers is that the sum is always a perfect sqaure: 1, 1+3 = 4, 1+3+5=9 etc then we will just have to match the first term of this to the one of the question by multiplying each term by 36

Since 36 is a perfect square this property is preserved: for a n = 2 n 1 where n Z : k = 1 n a k = n 2 \text{for } a_{n}=2n-1 \quad \text{where } n\in\mathbb{Z}\colon \quad \sum_{k=1}^{n}a_{k}=n^2 then 36 k = 1 n a k = 36 n 2 = 6 2 n 2 = ( 6 n ) 2 36\sum_{k=1}^{n}a_{k}=36n^2=6^{2}n^{2}=(6n)^{2} which is still a perfect square

So now it reduces to finding the 100th term of this new sequence: a 100 = 36 a 100 = 36 ( 2 × 100 1 ) = 7164 a'_{100}=36a_{100}=36(2\times 100 - 1) = \boxed{7164} The proof of sum of odd numbers equal to a perfect square can be trivially proven by induction or graphically and hence is omitted here.

I did it this way:

S(n) = n 2 \frac{n}{2} [2a+(n-1)d] = m 2 m^{2} ,where m is some positive integer.

Rearranging terms in L.H.S,

( d 2 \frac{d}{2} ) n 2 n^{2} + (a- d 2 \frac{d}{2} ) n = m 2 m^{2}

The only way to complete the square on the L.H.S is by making d=2a=72,

and so the sum of first n terms will be of the form S(n)= m 2 m^{2} = ( 6 n ) 2 (6n)^{2} .

Hence the 100th term will be a+(100-1)d= 7164 \boxed{7164} .

But you should have proven that d=72.Otherwise u can not claim that it's the only possible value for d.

rajdeep brahma - 4 years, 1 month ago

This is my solution (Sorry, I don't know how to use LaTeX here): The terms of this AP are 36, 36+a, 36+2·a, ... So the sum of the first n terms is: S n = 36 + (36+a) + (36+2·a) + ... + (36 + (n-1)·a) = (36 + ... + 36) + (0·a + 1·a + 2·a + ... + (n-1)·a) Therefore, S n = 36·n + a·n·(n-1)/2 If we expand that expression, we have S n = 36·n - (a/2)·n + (a/2)·n^2 = (36 - a/2)·n + (a/2)·n^2 So one realizes that it would be enough if a/2 were 36, since S n would be a perfect square: S_n = 36·n^2. Hence, we are looking for a = 72. With that, the 100th term of the AP is 36 + 99·72 = 7164

Marco Massa
Jun 20, 2014

Let d be the common difference of the AP. We want that i = 1 k a k = s q u a r e , \sum_{i=1}^{k} a_k = square , for any k { 1 , 2 , . . . , 100 } k \in \lbrace 1,2,...,100 \rbrace .Is simple to see that the sum is 36 k + d k ( k + 1 ) 2 36k + d \frac{k(k+1)}{2} , and so we want

36 k + d k ( k + 1 ) 2 = s q u a r e 36k + d \frac{k(k+1)}{2} = square for any k k .

Rearranged the sum like 36 ( k + d k ( k + 1 ) 72 ) 36( k + d \frac{k(k+1)}{72})

and see that if d = 72 d= 72 we have

i = 1 k a k = 36 k 2 \sum_{i=1}^{k} a_k = 36k^2 for any k.

So d = 72 d=72 and a 100 = 36 + 99 × 72 = 1764 a_{100}= 36+99 \times 72 = \boxed{1764}

Fredrik Meyer
Apr 22, 2014

First of all, nothing in the problem description says the constant sequence a n = 36 a_n = 36 $ cannot be a solution. Secondly, the solution is best done by drawing. Draw a 1x1 square. To get a 2x2 square, you add 3. To get a 3x3 square you add 5. Conclusion: the sum of the first n n odd numbers is a perfect square. So let a n = 2 n 1 a_n=2n-1 . However, we want our sequence to start with 36, with is a perfect square, so we multiply a n a_n with 36 to get b n = 72 n 36 b_ n = 72n-36 . This is our sequence. It works, because the product of a square and a square is a square.

Can you please clarify how the constant sequence can be a solution???I said that the sum of the n terms should be a perfect square not each term individually.....To learn why the constant sequence cannot be an answer read my reply to Milly ChooChoo...........

Eddie The Head - 7 years, 1 month ago

Log in to reply

Oh, my mistake! Silly thinking!

Fredrik Meyer - 7 years, 1 month ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...