Claim: $$\sum_{k=0}^{n} \frac{x^k}{k!} < e^x, \qquad \forall x > 0$$


Let $f(x)=e^x$. Then the Taylor series gives

$$f(x) = \sum_{k=0}^{\infty} \frac{x^k}{k!}$$

$$= \sum_{k=0}^{n} \frac{x^k}{k!} + \sum_{k=n+1}^{\infty} \frac{x^k}{k!}$$

Now, because $x>0$, $$\sum_{n+1}^{\infty} \frac{x^k}{k!} > 0$$

$$\implies f(x) > f(x) - \sum_{n+1}^{\infty} \frac{x^k}{k!}$$

$$\implies \sum_{k=0}^{n} \frac{x^k}{k!} = f(x) - \sum_{n+1}^{\infty} \frac{x^k}{k!} < f(x)$$


I would appreciate constructive criticism on this proof.

  • $\begingroup$ all fine. You showed that this is just a partial (and not the full) sum. $\endgroup$ – Andreas Nov 3 '17 at 15:09
  • 1
    $\begingroup$ The inequality should be reversed: $$f(x) < f(x) - \sum_{n+1}^{\infty} \frac{x^k}{k!}$$. $\endgroup$ – Cm7F7Bb Nov 3 '17 at 15:10
  • $\begingroup$ It can help: math.stackexchange.com/questions/2228083/… $\endgroup$ – Anıl B.C.T. Nov 3 '17 at 15:12
  • $\begingroup$ @Cm7F7Bb, thank you. Typo! $\endgroup$ – AndyDufresne Nov 3 '17 at 15:15
  • $\begingroup$ What is $e$ and $e^x$ for the purposes of this task? Can it be that $e=\lim(1+\frac1n)^n$ , $e^x$ is simply the exponential function for that basis, and the equality to the exponential power series is not previously known? $\endgroup$ – Lutz Lehmann Nov 3 '17 at 16:20

You are assuming $e^x = \sum_{k=0}^{\infty}\frac{x^k}{k!}$. In this case you can observe that the sequence $$f_n(x) := \sum_{k=0}^n \frac{x^k}{k!}, \quad x > 0$$ is strictly increasing, as $f_{n+1}(x)$ is $f_n(x)$ plus a positive number (as $x$ is positive). This implies that $(f_n(x))_{n=0}^{\infty}$ admits a limit, which is $e^x$. Using the definition of limit and that $(f_n(x))_{n=0}^{\infty}$ is an increasing sequence, if you fix $\varepsilon > 0$ then you can find $N \in \mathbb{N}$ such that $$0 < e^x-f_n(x) < \varepsilon \text{ if } n>N$$ and this implies in particular that $$f_n(x) < e^x.$$

Suppose you don't know $e^x = \sum_k \frac{x^k}{k!}.$ Consider $$g_n(x):=e^x - \sum_{k=0}^n \frac{x^k}{k!}, \quad x \geq 0.$$ Observe that $g_n(0) = 0$ for any $n \in \mathbb{N}$. Assume by contradiction that there is an $N \in \mathbb{N}$ such that $$e^x \leq \sum_{k=0}^N \frac{x^k}{k!}$$ and further, that this is the least natural number for which this happens, that is $$e^x > \sum_{k=0}^{N-1} \frac{x^k}{k!}.$$ This is equivalent to saying that \begin{equation} g_N(x) \leq 0 \end{equation} and $$g_{N-1}(x) > 0.$$ But one can compute directly $$g_N'(x) = g_{N-1}(x) > 0.$$ This fact and $g_n(0) = 0$ for every $n \in \mathbb{N}$ yield $g_N(x) > 0$ if $x > 0$, contradiction with $g_N(x) \leq 0$.

| cite | improve this answer | |

You argument is fine, but there also is a shortcut. By exploiting Taylor's theorem with integral remainder we have:

$$ e^x = \sum_{k=0}^{n}\frac{x^k}{k!}+\frac{1}{n!}\int_{0}^{x} e^t (x-t)^n\,dt\tag{A} $$ and it is pretty obvious that the last integral is positive if $x> 0$, since it is the integral of a continuous and positive function.

| cite | improve this answer | |

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.