Proving LCM Inequality for Positive Integers

Click For Summary
SUMMARY

The discussion proves the inequality for positive integers \(m > n\): \(\operatorname{lcm}(m,n)+\operatorname{lcm}(m+1,n+1)>\frac{2mn}{\sqrt{m-n}}\). The proof utilizes the definition of least common multiple (LCM) and greatest common divisor (GCD), specifically \(\text{lcm}(x,y)=\frac{x\cdot y}{\gcd(x,y)}\). By applying the AM-GM inequality, it establishes that if \(\gcd(m,n)\cdot \gcd(m+1,n+1) \leq m-n\), the inequality holds true, concluding with the fact that \(\gcd(d_1,d_2)=1\) ensures \(d_1\cdot d_2\) divides \(m-n\).

PREREQUISITES
  • Understanding of least common multiple (LCM) and greatest common divisor (GCD)
  • Familiarity with the AM-GM inequality
  • Basic knowledge of number theory concepts
  • Ability to manipulate mathematical inequalities
NEXT STEPS
  • Study the properties of LCM and GCD in number theory
  • Explore advanced applications of the AM-GM inequality
  • Investigate other inequalities involving LCM and GCD
  • Learn about the implications of GCD in modular arithmetic
USEFUL FOR

Mathematicians, number theorists, and students studying inequalities and number theory concepts will benefit from this discussion.

pedja
Messages
14
Reaction score
0
For all positive integers $$m > n$$, prove that :

$$\operatorname{lcm}(m,n)+\operatorname{lcm}(m+1,n+1)>\frac{2mn}{\sqrt{m-n}}$$
 
Mathematics news on Phys.org
Remember that $\text{lcm}(x,y)=\displaystyle\frac{x\cdot y}{\gcd(x,y)}$ so we have

\[A(m,n):=\text{lcm}(m,n)+\text{lcm}(m+1,n+1) = \frac{m\cdot n}{\gcd(m,n)}+\frac{(m+1)\cdot (n+1)}{\gcd(m+1,n+1)}\]

and then clearly

\[A(m,n) > \frac{m\cdot n}{\gcd(m,n)}+\frac{m\cdot n}{\gcd(m+1,n+1)} = m\cdot n \cdot \left( \tfrac{1}{\gcd(m,n)}+\tfrac{1}{\gcd(m+1,n+1)} \right)\]

Now let us use the AM-GM inequality ( $x+y \geq 2\sqrt{x\cdot y}$ for $x,y\geq 0$) to get

\[\frac{1}{\gcd(m,n)}+\frac{1}{\gcd(m+1,n+1)} \geq \frac{2}{\sqrt{\gcd(m,n)\cdot \gcd(m+1,n+1)}}\]

Next note that if we get $\gcd(m,n)\cdot \gcd(m+1,n+1) \leq m-n$, we are done.

To prove it, note that $d_1 = \gcd(m,n) = \gcd(n,m-n)$ which divides $m-n$, and $d_2=\gcd(m+1,n+1)=\gcd(n+1,m-n)$ which also divides $m-n$. But $d_1$ divides $n$ and $d_2$ divides $n+1$ ... and $\gcd(n,n+1)=1$ :p so in fact $\gcd(d_1,d_2)=1$ !.

Hence $d_1\cdot d_2$ must divide $m-n$, and so $d_1\cdot d_2 \leq m - n$ completing the proof $\square$.
 

Similar threads

  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 1 ·
Replies
1
Views
1K
Replies
2
Views
1K
Replies
1
Views
1K
  • · Replies 4 ·
Replies
4
Views
3K