Euler's number¶
Part 4 · Sequences · Chapter 3 · Lecture notes, volume 4 (PDF) · Exercises, volume 4 (PDF) · Exercises: Computing limits of sequences · Slides (PDF)
Prerequisites
-
From this module, «Sequences and limits of sequences»: monotone sequences.
-
From this module, «Factorials, binomial coefficients and triangle inequality»: Newton's binomial formula.
Objectives
At the end of this chapter you will be able to:
-
prove that \((1 + 1/n)^n\) is increasing and bounded, and define \(e\) as its limit;
-
compute limits related to \(e\);
-
use \(e\) in compound interest problems.
1. Euler's number \(e\) (Napier's constant)¶
Theorem 1
The sequence
is convergent.
Proof
We will prove that the sequence \(\{a_n\}\) is non-decreasing (\(a_n \ge a_{n-1}, \forall n \in \N, n\ge 2\)) and bounded (\(m \le a_n \le M, \forall n \in \N, n \ge 1\)); hence it is convergent by the monotone sequence theorem. The proof is split into three parts: (1) the monotonicity of \(\{a_n\}\) and its lower bound; (2) the monotonicity and the upper bound of an auxiliary sequence \(\{b_n\}\); (3) the synthesis of the two results.
Part (1). To prove that \(\{a_n\}\) is non-decreasing, we study, for \(n \ge 2\), the ratio:
where for the “\(\ge\)” we applied Bernoulli's inequality:
Hence we have
i.e., \(a_{n} \ge a_{n-1}\) and the sequence is non-decreasing.
To prove that \(\{a_n\}\) is bounded from below, we observe that, since \(a_1 = 2\) and the sequence is non-decreasing, it follows that \(a_n \ge 2, \forall n \ge 1\). □
Proof
Part (2). Consider the auxiliary sequence
To prove that \(\{b_n\}\) is decreasing, we study, for \(n \ge 2\), the ratio:
where for the “\(\le\)” we applied Bernoulli's inequality:
and for the “\(<\)” we applied the inequality
Hence we have proved that
and the sequence \(\{b_n\}\) is decreasing.
Since \(b_1=4\) and the sequence is decreasing, we therefore obtain
that is, \(\{b_n\}\) is bounded from above. □
Proof
Part (3). The two sequences are related by
and consequently
The sequence \(\{a_n\}\) is therefore non-decreasing and bounded and, by the monotone sequence theorem, it is convergent. □
Example 1: Graphs of the sequences \(\{a_n\}\) and \(\{b_n\}\)
-
The limit of the sequence \(a_n\) just studied is an irrational number that is very important in mathematics. This limit is denoted by the letter \(e\) (Euler's number, or Napier's constant) and its decimal representation begins as follows:
\[ 2. 7182818284 \dots \]By definition, we have:
\[\begin{equation} e= \lim_{n \rr \ip } \left( 1 + \frac{1}{n}\right)^n \label{nepero} \end{equation}\] -
This number is very often used as the base of logarithms, which, when this base is used, are called natural or Napierian logarithms (after the mathematician John Napier) and are denoted simply by the symbol \(\log\) (or \(\ln\)) without indicating the base.
-
Consequently, we have:
\[\begin{equation*} \lim_{n \rr \ip } n \: \log \left( 1 + \frac{1}{n}\right) = \lim_{n \rr \ip } \log \left( 1 + \frac{1}{n}\right)^n = \log e =1 \end{equation*}\]
Example 2: Computing limits with the sequence tending to \(e\)
Since:
Theorem 2
Let \(\{c_n\}\) be any divergent sequence (to \(\ip\) or \(\im\)); then
Proof
Recall that, given \(x \in \R\), its integer part, denoted by \(\lfloor x \rfloor\), is the largest integer not exceeding \(x\).
If \(c_n \rr \ip\), since \(\lfloor c_n \rfloor > c_n -1\), by comparison we also have \(\lfloor c_n \rfloor \rr \ip\). Hence, by \(\eqref{nepero}\) and by the definition of limit we have
Using:
we obtain
and also
hence the first limit of the theorem follows from the comparison theorem.
If instead \(c_n \rr \im\), then the sequence \(d_n = -c_n \rr \ip\), and we have
Since \(d_n-1 \rr \ip\), the claim follows from the previous case.
The second limit of the theorem is proved analogously. □
This theorem is useful for computing limits involving the indeterminate form \(1^{\infty}\)
Example 3: Computing limits with the sequence tending to \(e\)
We compute
since:
We rewrite the sequence as follows:
and we consider the denominator
Summarizing, we have:
Alternative method:
since
Example 4: Computing limits with the sequence tending to \(e\)
Since:
Example 5: Computing limits with the sequence tending to \(e\)
If \(\alpha = 0\) the result is immediate, since the sequence is constant:
If instead \(\alpha \neq 0\) (so that we can divide by \(\alpha\)), we have:
where we used the previous theorem with \(c_n = \frac{n}{\alpha}\), a sequence divergent to \(\ip\) if \(\alpha>0\) and to \(\im\) if \(\alpha<0\).
For example, with \(\alpha=2\) we have:
Try it — the interactive graph below shows what you have just read: move the sliders.
2. Euler's number in finance¶
-
Suppose we own a capital of unit value and an annual interest rate \(t\). If the interest is paid annually, after one year the capital owned will be
\[ 1 + t \cdot 1 = 1 + t \] -
If instead the interest is paid monthly, we will have
-
after the first month, a capital equal to
\[ 1+ \frac{t}{12} \cdot 1= \underbrace{1 + \frac{t}{12}}_{\alpha} \] -
after the second month, a capital equal to
\[ \underbrace{1 + \frac{t}{12}}_{\alpha} + \frac{t}{12} \underbrace{\left( 1 + \frac{t}{12} \right)}_{\alpha} = \underbrace{\left(1 + \frac{t}{12}\right)^2}_{\beta} \] -
after the third month, a capital equal to
\[ \underbrace{\left(1 + \frac{t}{12}\right)^2}_{\beta} + \frac{t}{12} \: \underbrace{\left(1 + \frac{t}{12}\right)^2}_{\beta} = \left(1 + \frac{t}{12}\right)^3 \] -
at the end of the year we will have a capital equal to
\[ \left(1 + \frac{t}{12}\right)^{12} \]
-
-
If the interest is computed every \(n\)-th of a year, at the end we will have a capital equal to
\[ \left(1 + \frac{t}{n}\right)^{n} \] -
For \(t = 1\) (a 100% return) we obtain exactly the sequence that defines \(e\)
Hence, even if the interest is paid an infinite number of times per year, the capital does not grow to infinity but tends to \(e\), since:
\[ \lim_{n \rr \ip } \left( 1 + \frac{1}{n}\right)^n= e \]
3. In summary¶
In summary
-
\((1 + 1/n)^n\) is increasing and bounded, so it converges: its limit is \(e \approx 2.718\).
-
\(e\) is the base of the natural exponential and logarithm.
-
Interest paid \(n\) times a year at rate \(t\) gives the factor \((1 + t/n)^n\), which tends to \(e^t\).
Exercises on this chapter: Computing limits of sequences