1
$\begingroup$

Consider the two Bernoulli (thank you @ErikSatie) inequalities for $x>-1$: $$ 0<\alpha<1 \quad \to\quad \left(1+x\right)^\alpha\le 1+\alpha x,\tag{1} $$ $$ \alpha<0\lor 1<\alpha \quad \to\quad \left(1+x\right)^\alpha\ge 1+\alpha x.\tag{2} $$

It remains me to prove that the inequality is strict for every $x\ne 0$ (keeping $x>-1$)

It seams to me that this point is more difficult than the inequalities themselves. I can't use derivation.

In the following picture:

enter image description here

I plotted different cases of $\alpha\in\{-1,-0.1,0.1,0.5,0.9,1.1,2\}$, and with a different color the asymptotic cases $\alpha=0$ (yellow) and $\alpha=1$ (green).

What I have to do is to prove that in $(-1,\infty)$ there is only a single solution for $$ \left(1+x\right)^\alpha = 1+\alpha x,\quad \alpha\in (-\infty,0)\cup(0,1)\cup (1,\infty).\tag{3} $$

EDIT 1 The case (1) derives directly for the corresponding aspect for the arithmetic-geometric inequality (which is used to prove it).

$\endgroup$
1
  • $\begingroup$ It's Bernoulli's inequality ! $\endgroup$
    – Erik Satie
    Jan 3 at 13:33
1
$\begingroup$

Consider the function $f(x) = (1+x)^{\alpha}-1-\alpha x$. We know that $f(0)=0$, and we essentially wish to show that for $\alpha\neq 0,1$, $x=0$ is the only zero of $f(x)$.

Firstly, we can easily see that $f(x)$ is twice differentiable on $\mathbb{R}/\{0\}$ with first derivative $f'(x) = \alpha(1+x)^{\alpha-1}-\alpha$ and second derivative $f''(x)=\alpha(\alpha-1)(1+x)^{\alpha-2}$.

From this, we can see that $f'(0)=0$, and $f''(x)\neq0$ except at $\alpha=0,1$ where it is zero throughout.

Now, if there exists another $x_0\neq 0$ at which $f(x_0)=0$, then by Rolle's theorem, we would have a $c$ between $x_0$ and $0$ such that $f'(c)=0$.

But, then as $f'(0)=0$, we can again apply Rolle's theorem to get a $d$ between $c$ and $0$ at which $f''(d)=0$. This is impossible, unless $\alpha=0,1$.

Hence, the Bernoulli inequality is strict for every $x\neq 0$ for $\alpha\neq 0,1$

$\endgroup$
1
  • $\begingroup$ Your answer is useful, but I look explicitly to a solution without differentiation $\endgroup$
    – Wyatt
    Jan 26 at 13:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.