How to prove the existence of this progression? Struggling with a BDMO problem.

86 Views Asked by At

This problem is from Bangladesh Mathematical Olympiad $2023$, The problem statement is as follows-

Prove that there is sequence of $2023$ distinct positive integers such that the sum of the squares of two consecutive integer is a perfect square itself

I can come up with such a sequence but I can't prove it. Such as-

$(2^k-1)2^0 , (2^{k-1}-1)2^k , (2^{k-2}-1)2^{k+(k+1)} , (2^{k-3}-1)2^{k+(k+1)+(k+2)}\cdots\cdots\cdots$

Where $k$ represents the number of elements in the sequence I want to create. Say, I want such a sequence with 5 numbers. Plugging in $k=5$,

$(2^5-1)2^0=31$
$(2^{5-1}-1)2^5=480$
$(2^{5-2}-1)2^9=3584$
$(2^{5-3}-1)2^{12}=12288$
$(2^{5-4}-1)2^{14}=16384$

We get the sequence $31,480,3584,12288,16384$ which obviously meet up the conditions.

I also tried to find out the $n^{th}$ term of the sequence like this:

$n^{th}$ term = ${2^{k-(n-1)}-1}[2^{(n-1)(k+1-\frac{n}{2})}]$ (based on the observed property of the sequence , some simplifications will lead to this)

Then what is left to prove is :

$\sqrt{({2^{k-(n-1)}-1}[2^{(n-1)(k+1-\frac{n}{2})}])^2+({2^{k-(n)}-1}[2^{(n)(k+1-\frac{n+1}{2})}])^2}$ AND
$\sqrt{({2^{k-(n-1)}-1}[2^{(n-1)(k+1-\frac{n}{2})}])^2+({2^{k-(n-2)}-1}[2^{(n-2)(k+1-\frac{n-2}{2})}])^2}$

Are integers, that's where I am stuck, I can't conclude because huge calculations arise with seemingly no positive deduction.

Thanks, at least for reading this:)

2

There are 2 best solutions below

0
On

A different approach

All we need is the Egyptian triangle. Or the most common Pythagorean triple. Let us substitute $2023$ with $5$ for the sake of simplicity. Then the following sequence works:

$$3^5\cdot 4,\;3^4\cdot 4^2,\; 3^3\cdot 4^3,\; 3^2\cdot 4^4,\; 3\cdot 4^5$$

Now sum of squares of two consecutive numbers (say, second and third one) is $$(3^3\cdot4^2)^2\cdot(3^2+4^2)= (3^3\cdot4^2)^2\cdot5^2= (3^3\cdot4^2\cdot5)^2.$$

0
On

HINT: Consider the geometric sequence $a,ar,ar^2,ar^3,\dots,ar^{2022}$. Choose any Pythagorean triplet $(x,y,z)$. Set $r=x/y$ and $a=y^{2022}$. Do you see what the individual terms become? Does it satisfy your required property? It becomes clear that there are infinitely many sequences that satisfy your given property and they all may be of differing lengths.