The problem asks to prove that if $0 < x < y$ then $x^{n} < y^{n}$, where $n$ is a positive integer, so I started by assuming that $0 < x < y$. I then wrote this chain of inequalities: $x^{n} < x^{n-1}y^{1} < x^{n-2}y^{2} < x^{n-3}y^{3} < ... < x^{2}y^{n-2} < x^{1}y^{n-1} < y^{n}$. It is true that $x^{n} < x^{n-1}y^{1}$, since dividing both sides by $x^{n-1}y^{0}$ yields $x < y$. Similarly, dividing both sides of $x^{n-1}y^{1} < x^{n-2}y^{2}$ by $x^{n-2}y^{1}$ yields $x < y$. It seems to me that this pattern continues for all the inequalities, but I haven't been able to prove this. Does that make the proof invalid? Also, how can I prove that this pattern always holds true?
-
9Use induction on $n$. – Wuestenfux Aug 14 '19 at 09:58
-
I think it is ok. If you want to formalise it you have to use induction – Lucio Tanzini Aug 14 '19 at 10:10
-
I appreciate you for finding your own proof. – Divya Prakash Sinha Aug 14 '19 at 11:15
5 Answers
Your proof is absolutely correct. But to make the proof clearer and standard I would recommend you to write these first.
$\because$ $0<x<y$.
$\implies \frac{y}{x}>1$ .....$(i)$
Multiplying both sides by $x^n$ we get,
$x^{n-1}y^1>x^n$
Again Multiplying both sides of $(i)$ by $x^{n-1}y$ we get,
$x^{n-2}y^2>x^{n-1}y$
Continuing this process of multiplying LHS of the obtained inequality to $(i)$, we finally obtain,
$y^n>x^1y^{n-1}$
Now you can write
$x^{n} < x^{n-1}y^{1} < x^{n-2}y^{2} < x^{n-3}y^{3} < ... < x^{2}y^{n-2} < x^{1}y^{n-1} < y^{n}$.$\blacksquare$
Hint: Use that $$a^n-b^n=(a-b)(a^{n-1}+a^{n-2}b^1+\cdots ab^{n-1}+b^n)$$
- 95,283
The comments are trending pretty strongly that you should be using induction. And, in one sense, they're not wrong. Whenever you use the symbol "..." in the reasoning of a proof, it's usually a pretty blatant sign that you're trying to avoid induction.
But, you know, if this wasn't for a formal proof class, I'd give you high marks for this. I understand your chain of reasoning well, and sometimes an intuitive argument can get lost in a sea of formality. And I've never thought of doing it this way myself, so I'm glad you were able to convey your original thoughts effectively. Gold star!
This is equivalent to
$$x>0,\\1<\frac yx\implies 1<\left(\frac yx\right)^n$$
or
$$1<t\implies 1<t^n.$$
By induction,
$$1<t\implies 1<t^1$$
and
$$1<t,1<t^n\implies 1<t^{n+1}.$$
Induction is a workable approach here as mentioned by Matthew. Also you can use algebra as Graubner writes.
I would probably do it with calculus. Investigate the properties of the function $f(x) = x^n$ or maybe $f(x) = x^n - x^{n-k}$. Can you conclude where it is growing or where it could have zeroes?
- 25,824