Proof verification of some properties of the exponent.

954 Views Asked by At

I'd like to know if the proof of the following laws of exponent is correct. First of all thanks in advance and I apologize for the extension of the following proof but it's one exercise. I hope there will not problem with this.

Definition: Let $x$ be a positive real number, and let $n\ge 1$ be a positive integer. We define $,x^{1/n}$ also known as the $n^{th}$ root of $x$ by the following formula:

$$\,x^{1/n}:=\text{ sup }(\{y\in \mathbb{R}: y\ge 0 \text{ and } y^n \le x \})$$

Lemma: Let $x,y>0$ be positive reals and let $n,m\ge 1$ be positive integers

(a) If $y = x^{1/n}$, then $y^n = x$.

(b) if $y^n = x$, then $y = x^{1/n}$

(c) $x^{1/n}$ is a positive real number

(d) We have $x<y \iff x^{1/n}<y^{1/n}$

(e) If $x>1$, then $x^{1/k}$ is an decreasing function of $k$. If $x<1$, then $x^{1/k}$ is an increasing function of $k$. If $x=1$, then $x^{1/k}=1$ for all $k$

(f) We have $(xy)^{1/n} = x^{1/n}y^{1/n} $

(g) We have $(x^{1/n})^{1/m}=x^{1/mn}$.

Proof:

(a)

Let $E=\{y\in \mathbb{R}: y\ge 0 \text{ and } y^n \le x \}$ and let $\,x^{1/n}=\text{ sup } (E)$. We have to show that either $y^n < x$ and $y^n > x$ leads to a contradiction.

Suppose $y^n < x$. If we can find an $\epsilon>0$ such that $(y+\epsilon)^n<x$ this means that $\,(y+\epsilon)\in E$ which contradicts the fact that $y$ is an upper bound for $E$.

Claim 1: If $\,0< h \le 1$ and $z>0$. Then $\,(z+h)^n \le z^n+h ((z+1)^n-z^n)$ for all $n\ge 0$.

Proof of claim 1: We use induction on $n$. Let $P(n)$ be the statement of the above claim. Clearly the base case holds. Now suppose inductively that $n$ is a natural number and $P(n)$ has already been proven. We now prove $P(n+1)$.

\begin{align} (z+h)^{n+1}=(z+h)(z+h)^{n} &\le(z+h)[z^n+h ((z+1)^n-z^n)]\\ &=(z^{n+1}+hz^n)+h (z+h)((z+1)^n-z^n)\\ &\le(z^{n+1}+hz^n)+h (z+1)((z+1)^n-z^n)\\ &=z^{n+1}+h ((z+1)^{n+1}-z^{n+1}) \end{align}

which close the induction and prove the claim. $\;\Box$

Since $y^n<x$, then $x-y^n$ is a positive real number. Let define $\epsilon< \text{min} \bigg( \frac{x-y^n}{(y+1)^n-y^n},1 \bigg)$. Thus, $0<\epsilon \le 1$ and by hypothesis $y>0$. We thus have $(y+\epsilon)^n\le x$ which is a contradiction.

Now suppose that $y^n>x$. We shall show that there is an $\epsilon>0$ such that $y-\epsilon$ is an upper bound of $E$ contradicting that $y$ is the least upper bound of $E$.

Claim 2: If $\,0< h \le 1$ and $z>h$. Then $\,(z-h)^n \ge z^n-h ((z+1)^n-z^n)$ for all $n\ge 0$.

Proof: We induct on n. The base case is when $n=0$ and clearly holds since both sides are equal. Now suppose inductively that the claim holds for some $n\ge 0$; we wish to show that also holds for $n+1$.

\begin{align} (z-h)^{n+1}=(z-h)(z-h)^{n} &\ge(z-h)[z^n-h ((z+1)^n-z^n)]\\ &=(z^{n+1}-hz^n)-h(z-h)((z+1)^n-z^n)\\ &\ge(z^{n+1}-hz^n)-h(z+1)((z+1)^n-z^n)\\ &=z^{n+1}-h((z+1)^{n+1}-z^{n+1})\\ \end{align}

which close the induction. $\; \Box$

Since $y^n>x$, then $y^n-x$ is a positive real number. Let $\epsilon < \text{min} \bigg( \frac{y^n-x}{(y+1)^n-y^n},1, y \bigg)$. Thus, $0\le \epsilon < 1$ and $y> \epsilon$. Then, we have $(y-\epsilon)^n \ge x$ contradicting that $y$ is the least upper bound since we find an upper bound which is less than $y$.

Hence, we must have $y^n=x$.

(b)

Suppose $y^n = x$, we need to show that $y = x^{1/n}$. We argue for contradiction, suppose $y \not= x^{1/n}$. Then by the trichotomy of the real numbers either $x^{1/n}<y$ or $x^{1/n}<y$.

If $\,x^{1/n}<y$ and since $x^{1/n}$ is the lub of $E$ so that it is non-negative (otherwise we have a contradiction, because $0 \in E$) and it is not zero, because by (a) if were zero that implies $x=0$ but contradicts our hypothesis, then $x^{1/n}$ is positive. This means $0<x^{1/n}<y$, which implies $0<x<y^n=x$ a contradiction

If $\,y<x^{1/n}$ and since $y>0$ using a similar argument as above we have $0<y^n=x<x$ which is a contradiction. Therefore $y = x^{1/n}$.

(c)

We can use the same argument as above when shows that $x^{1/n}$ is non-negative. But also cannot be zero, since $x>0$.

(d)

($\Rightarrow$) For the sake of contradiction suppose $x^{1/n}\ge y^{1/n}$ by (c) we already know that $0<y^{1/n}$. Moreover by (a) we have $y=(y^{1/n})^{n}$ and $x=(x^{1/n})^{n}$. Thus, $0<y^{1/n}\le x^{1/n}$ implies $0<y \le x$ which is a contradiction.

($\Leftarrow$) Suppose $x^{1/n}<y^{1/n}$ then $x<y$

(e) When $x>1$ we have to show that if $k,\ell \in \mathbb{N}-\{0\}$ and $\mathcal k<\ell$, then $x^{1/k}\ge x^{1/\ell}$. We argue this for contradiction.

Suppose $x^{1/k}< x^{1/\ell}$. Then using (d) and the properties of the exponentiation with natural numbers we have $x^{\ell}< x^{k}$. Let $m=\ell-k$. Thus, $x^{k+m}=x^k x^m< x^{k}$.

It is easy to show that $x^n>0$ for each positive integer so its inverse is also positive. Thus $x^m = x^{\ell-k}<1$. We claim that this is a contradiction.

Claim 3: If $x>1$ and $n$ is a positive integer, then $x^n>1$. If we can prove this, then since $m=\ell-k$ is a positive integer we have the desired contradiction.

Proof: The base case is when $n=1$ and is trivial. Now suppose the claim holds for $n$. We shall show that also holds for $n+1$ which close the induction. $x^{n+1}= x x^n> x$ and since $x>1$ the results follows. $\Box$

So, by the claim 3 and the fact $k-\ell\ge 1$, we have that $x^m = x^{\ell-k}<1$ is a contradiction.

If $x<1$, so $1<1/x$ and by what we have said above, if $k<\ell$ we thus have $(1/x)^{1/k}\ge (1/x)^{1/\ell}$.

We claim that $1/x^{1/k} =(1/x)^{1/k} $. Let $a= x^{1/k}$ and $b = (1/x)^{1/k}$. Thus, $a^k = x$ and $b^k= 1/x=1/a^k=(1/a)^k$. Then by b) we have $b=1/a$. In other words we have $(1/x)^{1/k} = 1/x^{1/k}$ as desired.

Then $1/x^{1/k}\ge 1/x^{1/\ell}$ and thus $x^{1/\ell} \ge x^{1/k}$ which prove that is increasing.

The case when $x=1$ is trivial.

(f)

We have to show $(xy)^{1/n} = x^{1/n}y^{1/n} $. Let $a = x^{1/n}$, $b= y^{1/n}$ and $c=(xy)^{1/n}$. Then, $a^n=x,\,b^n=y,\,c^n =xy$. Thus, $c^n= xy =a^nb^n =(ab)^n$, which implies that $c=ab$, i.e., $(xy)^{1/n}=x^{1/n}y^{1/n}$ as desired.

(g)

We have to show $(x^{1/n})^{1/m}=x^{1/mn}$. Let $a =x^{1/n}$, $b= a^{1/m}$. Then $a^n = x$, $b^m = a$ so $x=a^n =(b^m)^n= b^{mn}$. Thus, $b=x^{1/mn}$, which means $(x^{1/n})^{1/m}=x^{1/mn}$ and we're done.