Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

So I'm very close to finishing a proof of the exponential function in terms of differential equations. For this next step, I have to show the following.

For $n \ge 0$ define $E_n (t)$ recursively according to $$\begin{gather} E_0 (t) = 1, \\ E_{n+1}(t) = 1 + \int _0 ^t E_n(s)ds. \end{gather}$$

Show the following: $\displaystyle E(t)= \sum_{n=0}^{+\infty} \frac{t^n}{n!}$.

Any help would be appreciated!

share|cite|improve this question
Do you know that $E_n(t)$ is a convergent sequence?? – Babak S. Jan 23 '13 at 7:14
Or another way of putting this (sort of an intermediate step) is show that E_n(t) = 1 + t + t^2/2! + ... + t^n/n!. – Bryce Jan 23 '13 at 7:16
yes, the sequence of partial sums converges according to the ratio test – Bryce Jan 23 '13 at 7:16
I just had a TA help me. I was told E_0(t)=E_0(s)=1, which allows me to substitute this into my integral and get 1+t for E_1. Although I don't know why they're equivalent, this gives me the right solution – Bryce Jan 23 '13 at 7:22
Problems involving recursion just beg for induction proofs... – Antonio Vargas Jan 23 '13 at 7:50

We can do the problem by using Laplace transformation and also convolution of two functions.

As you noted, the sequence $E_{n+1}(t)=1+\int_0^{t}E_n(k)dk$ is convergent. So if $E_n(t)\to E(t)$ when $n\to\infty$ then we can have: $$E(t)=1+\int_0^{t}E(k)dk$$ According to what we knoe from convolution of two functions, we can write the latter identity as: $$E(t)=1+\int_0^{t}E(k)dk=1+1*E(t)$$ so we are led to the following ODE: $$E(t)=1+1*E(t)$$ Now take the Laplace of both sides and let $L(E(t))=Y$: $$Y=L(1+1*E(t))=\frac{1}{s}+\frac{1}{s}\times Y$$ So $Y=\frac{1}{s-1}$. This means that $E(t)=L^{-1}(\frac{1}{s-1})=\exp(t)$ .

Edit: Or, again we have : $$E(t)=1+\int_0^{t}E(k)dk$$ Differential from both sides, so: $$E'(t)=E(t)$$ a first order differential equation. It is separable and by solving it, you get $$\frac{d(E(t))}{E(t)}=dt$$ and so we have $E(t)=\exp(t)+C$. If you satisfy te condition you were given you will get $c=0$.

share|cite|improve this answer
Oh man, haha. We haven't learned Laplace transformations just yet. This is my first differential equations class. Also, thank you Arthur for editing it! – Bryce Jan 23 '13 at 7:43
@Bryce: I added some notes. See them and I hope you get the answer. – Babak S. Jan 23 '13 at 8:59
I don't know why this has gone without an upvote for so long! +1 and well-deserved! – amWhy Feb 10 '13 at 0:20
@amWhy: Thanks so much for your kind manner. :-) – Babak S. Feb 10 '13 at 3:30

We could use induction: For $E_1(t)$ we have $E_{1}(t)=1+t=\sum\limits_{n=0}^{1}\frac{t^n}{n!}$. Now assume that for some positive integer $k$ we have $$E_k(t)=\sum\limits_{n=0}^{k}\frac{t^n}{n!}$$ Then $$E_{k+1}(t)=1+\int\limits_{0}^{t} \sum\limits_{n=0}^{k}\frac{s^n}{n!}ds=1+ 1+\sum\limits_{n=0}^{k}\int\limits_{0}^{t} \frac{s^n}{n!}ds=1+\sum\limits_{n=0}^{k}\frac{t^{n+1}}{(n+1)n!}=1+\sum\limits_{n=1}^{k+1}\frac{t^n}{n!}=\sum\limits_{n=0}^{k+1}\frac{t^n}{n!}$$ By induction we now have $E_{j}(t)=\sum\limits_{n=0}^{j}\frac{t^n}{n!}$ and thus $E(t)=\sum\limits_{n=0}^{\infty}\frac{t^n}{n!}$.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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