5

Problem.

Show that $$e^x > 1 + x , \ \ x \neq 0$$

My attempt.

Using Mean Value Theorem: $$f'(c) = \frac{f(b) - f(a)}{b-a}$$

$$\Rightarrow e^c = \frac{e^x - 1}{x}$$ $$\Rightarrow xe^c = e^x -1$$ $$\Rightarrow x+1 = \frac{e^x}{e^c}$$

I'm not feeling very comfortable with these calculations...Am I on the right track? Can I even choose $e^o$ in my initial calculation as it says that $x \neq 0 $ ?

user3200098
  • 1,227
  • 2
    Depends on your definition of $e^x$. – user2345215 Feb 28 '14 at 17:46
  • Just use the fact that $e^c > 1$ (when $x > 0$; you'll need to consider $x < 0$ separately). [Before that: what is the fact you've used to write $f'(c)$ equal to that ratio? What is $c$?] – ShreevatsaR Feb 28 '14 at 17:48
  • The second line equation in your aligned equations is wrong, check it. – kjetil b halvorsen Feb 28 '14 at 17:50
  • Is the first line of your attempt an appeal to the Mean Value Theorem? This should be stated explicitly (with words and sentences). Don't just throw a bunch of equations at the reader and expect him to figure out for himself what the point of them is. – hmakholm left over Monica Feb 28 '14 at 17:51
  • 1
    Henning, if the reader does not immediately realize that is the Mean Value Theorem, then the reader cannot answer my question. But I added it, since you asked. – user3200098 Feb 28 '14 at 17:53
  • 2
    @user3200098: If this is homework, then your task is to learn writing proofs in the style expected by mathematicians. That style does not accept writing just a series of equations and leaving it to the reader to figure out what the significance of your equations are. It is your job to explain what you're doing with the equation (with actual words and sentences), and attempts to wiggle out of that by insulting the reader will just earn you ridicule. – hmakholm left over Monica Feb 28 '14 at 17:56
  • @kjetilbhalvorsen What is wrong with my 2nd equation? I am using x = 0 in the mean value theorem, but I'm not sure if I am allowed to do that. – user3200098 Feb 28 '14 at 17:57
  • @HenningMakholm I didn't mean to insult anyone and I am sorry if you feel insulted. I retract my previous statement. – user3200098 Feb 28 '14 at 17:58
  • @ShreevatsaR I'm having a hard time considering x<0. Can you give a hint? – user3200098 Feb 28 '14 at 18:20
  • There's a problem that $e^0 \ngtr 1 + 0$. You probably meant to say $e^x \ge 1 + x$. – NovaDenizen Mar 01 '14 at 02:05
  • @NovaDenizen: The problem says $x \neq 0$. – ShreevatsaR Mar 01 '14 at 16:14
  • @user3200098: I have added an answer elaborating the proof for both $x>0$ and $x<0$. – ShreevatsaR Mar 01 '14 at 16:14

10 Answers10

2

A geometric way to prove the inequality is to use the convexity of the function $f=\exp$ (since its second derivative is $e^x\ge0$) then the line tangent at $x=0$ with equation: $$y=f'(0)x+f(0)=x+1$$ is below the curve of the function hence $$e^x> x+1,\quad x\ne0$$

2

(1) Assume $x>0$. By MVT, there is a $c \in (0,x)$ so that $e^c = \frac{e^x - 1}{x}$. Then $e^x = xe^c + 1$. But since $c \in (0,x)$, we have $e^c > 1$, and $e^x = xe^c + 1 > x+1$.

(1) Assume $x<0$. By MVT, there is a $c \in (x,0)$ so that $e^c = \frac{1-e^x}{-x}$. Then $e^x = xe^c + 1$. But since $c \in (x,0)$, we have $0 < e^c < 1$, and $e^x = xe^c + 1 > x+1$ (since $x<0$).

2

The Mean Value Theorem is a valid approach. Here is another approach; hopefully, it will make you more comfortable to have two.

The integer version of Bernoulli's Inequality, proven at the end of this answer, is sufficient to prove that, for $n_0\gt\max(-x,1)$ and $x\ne0$, we have $$ 1+x\lt\left(1+\frac x{n_0}\right)^{n_0}\le\lim_{n\to\infty}\left(1+\frac xn\right)^n=e^x $$

robjohn
  • 345,667
2

for $x>0$, $e^x>1$ so $$\int_0^b e^xdx>\int_0^b dx=b$$ $$e^b-1>b$$ $$e^b>b+1$$ for $x<0$, $e^x<1$ so $$\int_a^0 e^xdx<\int_a^0dx,a<0$$ $$1-e^a<-a$$ $$1+a<e^a$$ so for any non zero number $c, e^c>c+1$ seeing as $e^0=1$

0

Enough to show that for $x> 0$ we have $(1-x) < e^{-x}$, or $$(1-x) e^x < 1$$ But we have $$(1-x)\cdot e^x = 1 + \sum_{n\ge 1} \left(\frac{1}{n!} - \frac{1}{(n-1)!}\right) x^n < 1$$

orangeskid
  • 53,909
0

The answer can be solved using the Taylor Series of $e^x$, which is $\displaystyle \sum_{n=0}^{\infty} \frac{x^n}{n!} = 1 + x + \frac{x^2}{2!} + \frac{x^3}{3!} + \cdots$. Note that if $x > 0$, each and every term in the series will be greater than $0$, and therefore $e^x > 1+x$.

2012ssohn
  • 3,827
  • 2
    Its also true for $x<0$. Your example is not a counterexample. – J.R. Feb 28 '14 at 17:52
  • Even for $x < 0$, we have $e^x > 1 + x$. And Taylor series is overkill (unless the exponential function has not been defined that way). – ShreevatsaR Feb 28 '14 at 17:53
0

Let's consider 2 cases:

a) $x > 0$: Then $(e^x - 1)/x = e^c > 1$ for some $c > 0$ then the conclusion follows.

b) $x < 0$: Then we need to prove: $-x > 1 - e^x$ or $1 > \frac{(1 - e^x)}{(0 - x)}$. Apply MVT again we have $\text{RHS} = e^c$ for some $c \,\text{in}\, (x, 0)$ so $\text{RHS} = e^c < e^0 = 1$ and the conclusion follows.

collapsar
  • 344
DeepSea
  • 77,651
0

Let $$f(x) = e^x - x - 1$$

Therfore you have $$f'(x) = e^x-1 $$

If you have a look at $f'(x)$ you will see that it is always greater than $0$ for $x>0$ . The value of $f(0)=0$ . After that the function is monotonically increasing and will always be $>0$ . Therefore from the above we can conculde that for $x>0$ , $f(x)>0$ . Hence $e^x-x-1>0$ , $e^x>x+1$

abkds
  • 2,210
0

Let $f$ be the exponential function.

Suppose $x > 0$. Then by the mean-value theorem, there exists some $c$, with $0 < c < x$, such that $$f'(c) = \dfrac{f(x) - f(0)}{x - 0}$$ $$e^c = \dfrac{e^x - 1}{x}.$$ As $c > 0$, we have $e^c > 1$, and therefore $\dfrac{e^x - 1}{x} > 1$. Multiplying both sides by the positive number $x$ gives $e^x - 1 > x$.


Suppose $x < 0$. Then by the mean-value theorem, there exists some $c$, with $x < c < 0$, such that $$f'(c) = \dfrac{f(0) - f(x)}{0 - x}$$ $$e^c = \dfrac{1 - e^x}{-x} = \dfrac{e^x - 1}{x}.$$

As $c < 0$, we have $e^c < 1$, and therefore $\dfrac{e^x - 1}{x} < 1$. Multiplying both sides by the negative number $x$ reverses the inequality, giving $e^x - 1 > x$ again, as required.

ShreevatsaR
  • 41,374
0

Your approach using the Mean Value Theorem is OK, but you made a mistake in the final line. It would correctly read $x+{1\over e^c}={e^x\over e^c}$ instead of $x+1$ on the left hand side. But you really should have stopped at the second line, $e^c={e^x-1\over x}$ for some $c$ between $0$ and $x$, and then reasoned as follows:

If $x\gt0$, then $c\gt0$. In this case $(e^x-1)/x=e^c\gt1$, hence $e^x-1\gt x$, so that $e^x\gt1+x$.

If $x\lt0$, then $c\lt0$. In this case $(e^x-1)/x=e^c\lt1$, hence (and here you have to remember than multiplying by a negative number reverses the direction of the inequality!) $e^x-1\gt x$, so that $e^x\gt1+x$.

Barry Cipra
  • 79,832