Can Natural Numbers m and n Satisfy x + 1/m < y - 1/n?

  • Thread starter Thread starter mathstudent1
  • Start date Start date
  • Tags Tags
    Inequality Numbers
Click For Summary
The discussion revolves around proving the existence of natural numbers m and n such that x + 1/m < y - 1/n, given that x < y. Participants explore the implications of y - x and the Archimedean property to find suitable m and n. They highlight the need to ensure that 1/m and 1/n are positive and less than y - x, particularly addressing cases where y - x is greater than or less than 1. The conversation emphasizes structuring arguments properly and generalizing findings from specific examples to broader cases. Ultimately, the proof hinges on finding natural numbers that satisfy the derived inequalities.
mathstudent1
Messages
18
Reaction score
0
No Effort - Member reminded that some effort must be shown in homework problems.
Homework Statement
prove x+1/m < y-1/n
Relevant Equations
.
Let x, y ∈ R be such that x < y. Prove that there exist natural numbers m and n such that x +1/m < y −1/n?
 
Physics news on Phys.org
Because this is a homework-type of problem, you must show your work and then we can give hints and guidance. I will cheat and give you an initial hint. Consider y-x and see what that tells you about n and m.
 
Can you make an attempt at this?
 
FactChecker said:
Because this is a homework-type of problem, you must show your work and then we can give hints and guidance. I will cheat and give you an initial hint. Consider y-x and see what that tells you about n and m.
 
Last edited by a moderator:
i try but i do not know how to complete and i think my assumption is false
ass real.jpg
 
PeroK said:
Can you make an attempt at this?
ass real.jpg

i think my assumption is false i don ot know how i complete
 
Please type yoru attempt into a post, with some explanation of what you are doing.
 
For your proof to work, you would need to begin with ##\frac{1}{m}+\frac{1}{n}<0##, which is false for any ##m,n>0##.

If you get stuck, try proving it differently. Can you choose ##m## and ##n## such that ##\frac{1}{m}+\frac{1}{n}## is less than ##y-x##?
 
  • Like
Likes FactChecker
by archimedean property right ? there is n,m which are integers s.t n>1/y and 1/m>-1/x then 1/n<y and 1/m<-x , add them we will get 1/m+1/n <y-x.
 
  • #10
mathstudent1 said:
by archimedean property right ? there is n,m which are integers s.t n>1/y and 1/m>-1/x then 1/n<y and 1/m<-x , add them we will get 1/m+1/n <y-x.
At some point, you'll need to learn to structure your arguments properly. That argument doesn't work as ##m, n## must be natural numbers. In particular, ##\frac 1 m## and ##\frac 1 n## cannot be negative.
 
  • #11
PeroK said:
At some point, you'll need to learn to structure your arguments properly. That argument doesn't work as ##m, n## must be natural numbers. In particular, ##\frac 1 m## and ##\frac 1 n## cannot be negative.
you are right , then how i prove it ?
 
  • #12
mathstudent1 said:
you are right , then how i prove it ?
My general advice if you are not sure how to prove something is to justify that it is true. In this case, you could start with, say, ##x = 0## and ##y = 1## and find suitable ##m, n##. Can you do that to start with?

Note that if you were more experienced, you might see a way to generalise that simple case to cover the case for any ##x## and ##y##.
 
  • #13
yes i can take m=3 and n=2
 
  • #14
mathstudent1 said:
yes i can take m=3 and n=2
If ##y - x \ge 1##, then those would do in general.

What about the case where ##y - x < 1##?
 
  • #15
PeroK said:
If ##y - x \ge 1##, then those would do in general.

What about the case where ##y - x < 1##?
they said in question x<y ,
PeroK said:
If ##y - x \ge 1##, then those would do in general.

What about the case where ##y - x < 1##?
also i can found m and n but i do not know who i sketch the proof
 
  • #16
mathstudent1 said:
they said in question x<y ,
Yes, but if ##x < y##, then ##y - x > 0##. That splits into two cases:

Case 1: ##y - x \ge 1##, where you have the solution ##m = \frac 1 3, n = \frac 1 2##.

Case 2: ##0 < y-x < 1##. Which looks harder, because ##y-x## is "small".
 
  • #17
Rewrite x+1/m < y-1/n as 1/m+1/n < y-x by subtracting x from both sides and adding 1/n to both sides.
So the problem can be restated as proving that you can find Natural numbers n,m such that 1/m+1/n < y-x.
Now, how can you make 1/m+1/n small enough? (Remember that x < y, so 0 < y-x.)
 
  • #18
mathstudent1 said:
also i can found m and n but i do not know who i sketch the proof
Let me give you a bit more help. When ##x =0## and ##y = 1##, here's what I think you probably did:

Note that ##\frac 1 2## is half way between ##0## and ##1##. So, let's take ##m, n \ge 2##. Although, it might have been better to take ##m = n = 3##.

For general ##x, y## you could generalise this:

Note that ##\frac {y- x}{2}## is half the distance between ##x## and ##y##. This is where you can apply the Archimedean property and take ##\frac 1 m = \frac 1 n < \frac{y-x}{2}##.
 
Last edited:
  • Like
Likes docnet and MatinSAR
  • #19
PeroK said:
Let me give you a bit more help. When ##x =0## and ##y = 1##, here's what I think you probably did:

Note that ##\frac 1 2## is half way between ##0## and ##1##. So, let's take ##m, n \le \frac 1 2##. Although, it might have been better to take ##m = n = \frac 1 3##.
. . .
@PeroK ,
Don't you mean ##\displaystyle \frac 1 m ## and ##\displaystyle \frac 1 n ## rather than ##m## and ##n## in this post ?
 
  • #20
FactChecker said:
Rewrite x+1/m < y-1/n as 1/m+1/n < y-x by subtracting x from both sides and adding 1/n to both sides.
So the problem can be restated as proving that you can find Natural numbers n,m such that 1/m+1/n < y-x.
Now, how can you make 1/m+1/n small enough? (Remember that x < y, so 0 < y-x.)

You can take n = m, so that you need \frac1{n} &lt; \frac{y - x}{2}.
 
  • Like
Likes FactChecker
  • #21
SammyS said:
@PeroK ,
Don't you mean ##\displaystyle \frac 1 m ## and ##\displaystyle \frac 1 n ## rather than ##m## and ##n## in this post ?
Yes. Fixed, thanks.
 
  • #22
pasmith said:
You can take n = m, so that you need \frac1{n} &lt; \frac{y - x}{2}.
That is a nice simplification that allows one to get a formula for n (=m).
##n > \frac {2}{y-x}##, and ##m=n##.
 
  • #23
It is sufficient to require <br /> \frac1n + \frac1m \leq \frac 2{\min(n,m)} &lt; y - x where the first inequality is true for all positive n and m. This again leads to the problem of proving the existence of a positive N such that N &gt; \frac{2}{y - x}.
 
  • Like
Likes FactChecker