How many positive integer solutions satisfy this equation?

  • Thread starter Thread starter songoku
  • Start date Start date
  • Tags Tags
    Integer Positive
Click For Summary
The discussion revolves around finding positive integer solutions for the equation $$\frac{1}{x}+\frac{1}{y}+\frac{1}{xy}=\frac{1}{2021}$$, which simplifies to $$xy = 2021(y + x + 1)$$. Participants explore the implications of the prime factorization of 2021 (43 x 47) and how it relates to the variables x and y. Various approaches are suggested, including rewriting the equation and considering the divisibility conditions imposed by the factors of 2021. Ultimately, the conversation leads to the conclusion that there are 32 positive integer solutions to the equation.
songoku
Messages
2,503
Reaction score
402
Homework Statement
How many positive integer solutions (x, y) satisfy ##\frac{1}{x}+\frac{1}{y}+\frac{1}{xy}=\frac{1}{2021}##
Relevant Equations
Not sure
$$\frac{1}{x}+\frac{1}{y}+\frac{1}{xy}=\frac{1}{2021}$$
$$\frac{y+x+1}{xy}=\frac{1}{2021}$$
$$xy = 2021y + 2021 x + 2021$$

Then I am stuck. How to continue?

Thanks
 
Physics news on Phys.org
Did the question of whether 2021 is prime occur to you?
 
  • Like
Likes songoku
PeroK said:
Did the question of whether 2021 is prime occur to you?
No

2021 is not prime (2021 = 43 x 47) but sorry I don't know how to use the hint to continue. To be honest, idea of using prime number to solve this question never crosses my mind.

Thanks
 
songoku said:
To be honest, idea of using prime number to solve this question never crosses my mind.
That's a fundamental problem. It's like a golfer who cannot figure out how to get out of a bunker and it never crosses their mind to use a sand wedge! Or a chef who can't light a stove and never thinks of using a match!

We cannot help you to think for yourself. Primes are so fundamental to number theory that it's impossible for me to understand why you didn't think of the prime factorisation of 2021.
 
  • Like
Likes epenguin and songoku
PeroK said:
That's a fundamental problem. It's like a golfer who cannot figure out how to get out of a bunker and it never crosses their mind to use a sand wedge! Or a chef who can't light a stove and never thinks of using a match!

We cannot help you to think for yourself. Primes are so fundamental to number theory that it's impossible for me to understand why you didn't think of the prime factorisation of 2021.
I never learn about number theory. This question is a challenge question from the teacher, just for practice and optional for the students to do it or not.

What do I need to learn and understand to be able to do this question?

Thanks
 
songoku said:
What do I need to learn and understand to be able to do this question?
I don't know. I don't know what you've studied and what you haven't and why you are doing these problems.

The last problem was the same. You had a circle and a point. My first thought was to use geometry and symmetry before diving into the algebra. You dived straight into the algebra. Why do I think of geometry in such a problem and you don't? I can't explain that.
 
  • Like
Likes hutchphd and songoku
Thank you very much PeroK
 
  • Like
Likes Hall
Hmm I can't solve this, though I probably have 25-30 years more experience than @songoku

@PeroK if I understand your hint, x and y must relate somehow to 43 and 47 but i can't think exactly how, you got to give us a secondary hint
 
  • Like
Likes songoku and Hall
Delta2 said:
Hmm I can't solve this, though I probably have 25-30 years more experience than @songoku

@PeroK if I understand your hint, x and y must relate somehow to 43 and 47 but i can't think exactly how, you got to give us a secondary hint
I haven't looked at it other than to check whether 2021 was prime. It's clear, therefore, that ##x## and/or ##y## must have factors of ##43## and/or ##47##.

It might still be difficult, but that must be the first thing to look at.
 
  • Like
Likes songoku and Delta2
  • #10
Here's one solution: ##x = 16512, \ y = 2303##.
 
  • Like
  • Skeptical
Likes songoku and Delta2
  • #11
Hmmm, did you use a python program to get those?
 
  • Like
Likes songoku
  • #12
Delta2 said:
Hmmm, did you use a python program to get those?
No. Pen and some Pret-a-Manger napkins!
 
  • Like
  • Haha
Likes malawi_glenn, songoku and Delta2
  • #13
I think I've found them all. Not easy, but ultimately all about prime factorisation!
 
  • Like
Likes songoku
  • #14
(\frac{1}{x}+1)(\frac{1}{y}+1)=\frac{1}{2021}+1
\frac{(x+1)(y+1)}{xy}=\frac{2022}{2021}=\frac{2*3*337}{43*47}
Is this helpful ?
 
  • Like
Likes songoku, Delta2, hutchphd and 1 other person
  • #15
Does it help to "discover" that $${1\over 2z} + {1\over 2z+1} + {1\over 2z(2z+1)} = {1\over z}\qquad ?$$(for which I needed to replace the 2021 by 4 and do some trial-and-error, to find 8,9 does the trick. But then so does 6,14 and I still have to find out how and why...)
 
  • Like
Likes songoku
  • #16
anuttarasammyak said:
Is this helpful ?
Very!
 
  • Like
Likes songoku and BvU
  • #17
To give some hints for what I did (which is not necessarily the quickest way);

There were two cases: a) ##x = 2021k## for some integer ##k##; and, b) ##x = 43k, y = 47m## for some ##k, m##.

a) leads quickly to a condition on ##k - 1##.

For b) I showed first that ##k > 47## and ##m > 43##.
 
  • Like
Likes songoku and Delta2
  • #18
As @songoku has mentioned that he has not been confronted (if that’s the right word here) with these types of problems before, I think we should at least guide him in a direction which would let him handle these types of problems instead of scathing him for why he didn’t do what he ought to.
 
  • Like
Likes songoku and Delta2
  • #19
Hello pal @songoku, we can rewrite your second equation as
$$
x+y+1= \frac{xy}{2021}$$

Now, we want ##x## and ##y## to be integer, that implies ##xy## must be a multiple of 2021, right? So, we can write it as ##xy =2021 k##.

Now, we can write out your third equation as:
$$
2021x + 2021y +2021= 2021k$$
Now, we have got a system:
$$
x+y =k-1$$

And $$xy =2021k$$
Taking ##y=k## and ##x=2021## would not satisfy, the first equation of our system, hence ##x## and ##y## must be the factors (or their multiples) of 2021. [Edited: or at most one of them can be a factor of ##k##, but not k itself].
 
Last edited:
  • Like
Likes anuttarasammyak, songoku, Delta2 and 1 other person
  • #20
Hall said:
Hello pal @songoku, we can rewrite your second equation as
$$
x+y+1= \frac{xy}{2021}$$

Now, we want ##x## and ##y## to be integer, that implies ##xy## must be a multiple of 2021, right? So, we can write it as ##xy =2021 k##.

Now, we can write out your third equation as:
$$
2021x + 2021y +2021= 2021k$$
Now, we have got a system:
$$
x+y =k-1$$

And $$xy =2021k$$
Taking ##y=k## and ##x=2021## would not satisfy, the first equation of our system, hence ##x## and ##y## must be the factors (or their multiples) of 2021.
Well, ##x = 4042, y = 4043## is a solution. So, your conclusion that ##y## must be a multiple of the factors of ##2021## is false.
 
  • Like
Likes Hall
  • #21
PeroK said:
Well, ##x = 4042, y = 4043## is a solution. So, your conclusion that ##y## must be a multiple of the factors of ##2021## is false.
A nice counter-one. I have edited my post.
 
  • #22
Hall said:
A nice counter-one. I have edited my post.
More formally, ##43## divides ##xy##, hence ##43## divides ##x## or ##43## divides ##y##. Similarly, ##47## divides ##x## or ##47## divides ##y##.

This gives us four possibilities, which reduces to two by the symmetry of ##x## and ##y##:

a) ##2021## divides ##x## (or, ##2021## divides ##y##)

b) ##43## divides ##x## and ##47## divides ##y## (or, vice versa).

Which is expanded on in post #17.
 
  • Like
Likes songoku and Delta2
  • #23
@PeroK It's so charming that there is no solution for first 2021 values of ##k##!
 
  • #24
Delta2 said:
Hmm I can't solve this, though I probably have 25-30 years more experience than @songoku

@PeroK if I understand your hint, x and y must relate somehow to 43 and 47 but i can't think exactly how, you got to give us a secondary hint
Edit:This is not a solution, because I can't think of one, but a suggestion for experiments:
Notice the function is symmetric in x,y, i.e., f(x,y)=f(y,x).
0)Test to see what happens when
0.1)Both x,y are small
0.2)Both are large
0.3)x is significantly larger
0.4) x=y ; x=y=2021, or x=y =k2021; k an integer.

1)Go to Wolfram and plot the function. Find it's max and min. Maybe there's a way of restricting to Integers.

2)Look for other interpretations. This is close to being the ratio of the perimeter of a rectangle to its area.

Hopefully this brainstorming will lead to an insight. Edit: Tl; dr: There is no royal road to Geometry, nor to any advanced Mathematics. Need to explore, struggle until something falls through.
 
Last edited:
  • Like
Likes songoku and Delta2
  • #25
From my post #14 I would rewrite the problem as

Find x,y satisfying
Pa=Rc-1\equiv x
Qb=Sd-1\equiv y
ab=cd
where a,b,c,d are positive integers,
\{P,Q\}=\{1,43*47\},\{43,47\},\{47,43\},\{43*47,1\} of four cases, and
\{R,S\}=\{1,2*3*337\},\{2,3*337\},\{3,2*337\},..., \{2*3*337,1\} of eight cases.

I am not sure such a rewriting would be useful.
 
Last edited:
  • Like
Likes WWGD, songoku and Delta2
  • #26
WWGD said:
Edit:This is not a solution, because I can't think of one, but a suggestion for experiments:
Notice the function is symmetric in x,y, i.e., f(x,y)=f(y,x).
0)Test to see what happens when
0.1)Both x,y are small
0.2)Both are large
0.3)x is significantly larger
0.4) x=y ; x=y=2021, or x=y =k2021; k an integer.

1)Go to Wolfram and plot the function. Find it's max and min. Maybe there's a way of restricting to Integers.

2)Look for other interpretations. This is close to being the ratio of the perimeter of a rectangle to its area.

Hopefully this brainstorming will lead to an insight. Edit: Tl; dr: There is no royal road to Geometry, nor to any advanced Mathematics. Need to explore, struggle until something falls through.
I have got the threshold ##k=8086## for the solution. (It can be worked out by focusing on the discriminant of ##x^2 -(k-1)x +2021k=0##) Now, we got to look for an upper limit, if it is possible.
 
  • Like
Likes songoku and WWGD
  • #27
PeroK said:
To give some hints for what I did (which is not necessarily the quickest way);

There were two cases: 1) ##x = 2021k## for some integer ##k##; and, 2) ##x = 43k, y = 47m## for some ##k, m##.

1) leads quickly to a condition on ##k - 1##.

For 2) I showed first that ##k > 47## and ##m > 43##.
The posts seem to be going in the wrong direction! A bit more help:

The starting equation is$$xy = 2021(x + y + 1) = (43)(47)(x + y + 1)$$1) In the first case we have ##x = 2021k## giving:$$ky = 2021k + y + 1$$Hence$$y = \frac{2021k + 1}{k - 1} = \frac{2021(k-1) + 2022}{k - 1} = 2021 + \frac{2022}{k-1}$$As ##y## is an integer, ##k-1## must be a divisor of ##2022##.

The second case is slightly trickier. See the hint.
 
  • Like
Likes songoku and BvU
  • #28
To the OP: You have started correctly. Either your teacher taught you well, or he/she told you how to start. To continue, think about what you have been learning recently in pre-calc. Did you learn how to graph a hyperbola? Did you learn how to solve a degree-2 equation?

word
 
  • Like
Likes songoku
  • #29
There's a quicker way. Consider ##r = x - 2021## and ##s = y - 2021## and the product ##rs##.
 
  • Like
Likes songoku
  • #30
A solution of post #25
Let {P,Q}={1,2021}, {R,S}={1,2022}
we get easily a=d=2022, b=c=2023
x=2022, y=2021*2023=4088483
I think there is no other solution for these particular P,Q,R,S or as for
2021b=2022d-1
d=2022 is the only possible number so that RHS is multiple of 2021. But I am not still certain of it.
 
Last edited:
  • Like
Likes songoku

Similar threads

Replies
5
Views
2K
  • · Replies 8 ·
Replies
8
Views
2K
  • · Replies 10 ·
Replies
10
Views
2K
  • · Replies 18 ·
Replies
18
Views
2K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 29 ·
Replies
29
Views
6K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 17 ·
Replies
17
Views
3K
  • · Replies 6 ·
Replies
6
Views
1K