Let be \( f : x \longmapsto f(x) \) a function of class \( \mathcal{C}^n \) and \( f^{(n)} \) its \( n \)-th derivative .
The Taylor-Young formula tells us that any function \( f \), centered at \( x = a \), can be written as a Taylor expansions \( (T_{n,a}) \) with a remainder \((R_{n,a})\), such as:
Here is the decomposed form:
Where the notation \( o\bigl((x - a)^n\bigr) \) represents the Landau remainder (a little rate), meaning that the remainder is negligible compared to \( (x - a)^n \) as \( x \to a \):
Furthermore, setting down \( (x = a + h) \), we do obtain a new form of this formula:
With the Taylor-Laplace's formula, the remainder is quantified by an integral.
Thus,
Recap of the main Taylor series
Another notation used to characterize the remainder of an Taylor expansions is the Landau notation \(o(x^n)\).
If a function \( f(x) \) is negligible compared to another function \( g(x) \) near a certain point \( a \), we can write it as:
It means that:
In our specific case, we study Taylor expansions at the neighbourhood of \(( a = 0 ) \), so:
$$ condition $$ |
$$ function $$ |
$$ Taylor \ series \ at \ 0 \ : T_n(0) $$ |
$$ \equiv T_n(0) $$ |
|---|---|---|---|
|
$$ \forall x \in \mathbb{R}, \ \forall \alpha \in \hspace{0.04em} \mathbb{N}^*, $$
|
$$ (1+x)^{\alpha}$$
$$ (Newton's \ binomial) $$
|
$$ 1 + \alpha x + \binom{\alpha}{2}x^2 + \binom{\alpha}{3}x^3 \ ... \ + \binom{\alpha}{\alpha}x^{\alpha} + o(x^{\alpha})$$
|
$$ \sum_{p = 0}^{\alpha} \binom{\alpha}{p} x^p + o(x^{\alpha}) $$ |
|
$$ \forall x \in \Bigl[ \mathbb{R} \hspace{0.2em} \backslash \bigl \{ -1 \bigr \} \Bigr], $$
|
$$ \frac{1}{1+x}$$ |
$$ 1 - x + x^2 - x^3 + \ ... \ + (-1)^n x^n + o(x^n)$$ |
$$ \sum_{k=0}^n (-1)^k x^k + o(x^n) $$ |
|
$$ \forall x \in \Bigl[ \mathbb{R} \hspace{0.2em} \backslash \bigl \{ 1 \bigr \} \Bigr], $$
|
$$ \frac{1}{1-x}$$ |
$$ 1 + x + x^2 + x^3 + \ ... \ + x^n + o(x^n)$$ |
$$ \sum_{k=0}^n x^k + o(x^n) $$ |
|
$$ \forall x \in [-1, \hspace{0.1em} + \infty[, $$
|
$$ \sqrt{1+x}$$ |
$$ 1 + \frac{1}{2}x - \frac{1}{8}x^2 + \frac{1}{4}x^3 - \frac{15}{16}x^4 + \ ... \ + o(x^{4})$$
|
$$ $$ |
|
$$ \forall x \in \hspace{0.04em} ]-1, \hspace{0.1em} + \infty[, $$
|
$$ 1 \over \sqrt{1+x}$$ |
$$ 1 - \frac{1}{2}x + \frac{3}{4}x^2 - \frac{15}{8}x^3 + \frac{105}{16}x^4 + \ ... \ + o(x^{4})$$
|
$$ $$ |
|
$$ \forall x \in \mathbb{R}, $$
|
$$ e^x $$ |
$$ 1 + x + \frac{x^2}{2!} + \frac{x^3}{3!} + \ ... \ + \frac{x^n}{n!} + o(x^n)$$ |
$$ \sum_{k=0}^n \frac{x^k}{k!} + o(x^n) $$ |
|
$$ \forall x \in \hspace{0.04em} ]-1, \hspace{0.1em} + \infty[, $$
|
$$ \ln(1+x) $$ |
$$ x - \frac{x^2}{2} + \frac{x^3}{3} - \frac{x^4}{4} + \ ... \ + \frac{ (-1)^{n-1} }{n} x^n + o(x^{n})$$
|
$$ \sum_{k=1}^n \frac{ (-1)^{k-1} }{k} x^k + o(x^{n}) $$ |
|
$$ \forall x \in \hspace{0.04em} ]1, \hspace{0.1em} + \infty[, $$
|
$$ \ln(1-x) $$ |
$$ -x - \frac{x^2}{2} - \frac{x^3}{3} - \frac{x^4}{4} - \ ... \ - \frac{x^n}{n} + o(x^{n})$$
|
$$ \sum_{k=1}^n -\frac{ x^k }{k} + o(x^{n}) $$ |
|
$$ \forall x \in \mathbb{R}, $$
|
$$ \sin(x) $$ |
$$ x - \frac{x^3}{3!} + \frac{x^5}{5!} - \frac{x^7}{7!} + \ ... \ + (-1)^n \frac{x^{2n+1}}{(2n+1)!} + o(x^{2n+2})$$
|
$$ \sum_{k=0}^n (-1)^k \frac{x^{2k+1}}{(2k+1)!}+ o(x^{2n+2}) $$ |
|
$$ \forall x \in \mathbb{R}, $$
|
$$ \cos(x) $$ |
$$ 1 - \frac{x^2}{2!} + \frac{x^4}{4!} - \frac{x^6}{6!} + \ ... \ + (-1)^{n} \frac{x^{2n}}{(2n)!} + o(x^{2n+1})$$
|
$$ \sum_{k=0}^n (-1)^{k} \frac{x^{2k}}{(2k)!} + o(x^{2n+1}) $$ |
|
$$\forall k \in \mathbb{Z}, \enspace \forall x \in \biggl[ \mathbb{R} \hspace{0.2em} \backslash \hspace{0.2em} \Bigl \{ \frac{\pi}{2} + k\pi \Bigr \} \biggr],$$
|
$$ \tan(x) $$ |
$$ x + \frac{1}{3}x^3 + \frac{2}{15}x^4 + \frac{17}{315}x^6 + \ ... \ + o(x^{6})$$ |
$$ $$ |
|
$$ \forall x \in \mathbb{R}, $$
|
$$ \operatorname{Arctan}(x) $$ |
$$ x - \frac{x^3}{3} + \frac{x^5}{5} - \frac{x^7}{7} + \ ... \ + (-1)^n \frac{x^{2n+1}}{(2n+1)} + o(x^{2n+2})$$
|
$$ \sum_{k=0}^n (-1)^k \frac{x^{2k+1}}{(2k+1)}+ o(x^{2n+2}) $$ |
|
$$ \forall x \in \mathbb{R}, $$
|
$$ \sinh(x) $$ |
$$ x + \frac{x^3}{3!} + \frac{x^5}{5!} + \frac{x^7}{7!} + \ ... \ + \frac{x^{2n+1}}{(2n+1)!} + o(x^{2n+2})$$
|
$$ \sum_{k=0}^n \frac{x^{2k+1}}{(2k+1)!}+ o(x^{2n+2}) $$ |
|
$$ \forall x \in \mathbb{R}, $$
|
$$ \cosh(x) $$ |
$$ 1 + \frac{x^2}{2!} + \frac{x^4}{4!} + \frac{x^6}{6!} + \ ... \ + \frac{x^{2n}}{(2n)!} + o(x^{2n+1})$$
|
$$ \sum_{k=0}^n \frac{x^{2k}}{(2k)!} + o(x^{2n+1}) $$ |
|
$$ \forall x \in \mathbb{R}, $$
|
$$ \tanh(x) $$ |
$$ 1 + \frac{x^2}{2!} + \frac{x^4}{4!} + \frac{x^6}{6!} + \ ... \ + \frac{x^{2n}}{(2n)!} + o(x^{2n+1})$$
|
$$ \sum_{k=0}^n \frac{x^{2k}}{(2k)!} + o(x^{2n+1}) $$ |
|
$$ \forall x \in \hspace{0.04em} ]-1, \hspace{0.1em} 1[, $$
|
$$ \tanh(x) $$ |
$$ x - \frac{1}{3}x^3 + \frac{2}{15}x^4 - \frac{17}{315}x^6 + \ ... \ + o(x^{6})$$ |
$$ $$ |
|
$$\forall k \in \mathbb{Z}, \enspace \forall x \in \biggl[ \mathbb{R} \hspace{0.2em} \backslash \hspace{0.2em} \Bigl \{ \frac{\pi}{2} + k\pi \Bigr \} \biggr], $$
|
$$ csc\left(\frac{\pi}{2} + x \right) = \sec(x) $$ |
$$ 1 + \frac{x^2}{2!} + \frac{x^4}{4!} + 61 \frac{x^6}{6!} + \ ... \ + o(x^{6})$$ |
$$ $$ |
|
$$\forall k \in \mathbb{Z}, \enspace \forall x \in \biggl[ \mathbb{R} \hspace{0.2em} \backslash \hspace{0.2em} \Bigl \{ \frac{\pi}{2} + k\pi \Bigr \} \biggr], $$
|
$$ cot\left(\frac{\pi}{2} + x \right) = -\tan(x)$$ |
$$ -x -\frac{1}{3}x^3 -\frac{2}{15}x^4 -\frac{17}{315}x^6 + \ ... \ + o(x^{6})$$ |
$$ $$ |
Proofs
Taylor-Young formula
-
Building the formula by successive integrations
Starting from the main equation of the fundamental theorem of calculus :
$$ f(x) = f(a) + \int_a^x f'(t)dt $$So,
$$ f(x) = f(a) - \int_a^x -f'(t)dt $$Performing an integration by parts , with a wise choice for \( u \) and \( v' \), we do have:
$$ \Biggl \{ \begin{gather*} u(t) = f'(t) \\ v'(t) = -dt \end{gather*} $$$$ \Biggl \{ \begin{gather*} u'(t) = f^{(2)}(t)dt \\ v(t) = x-t \end{gather*} $$$$ f(x) = f(a) - \Biggl( \Bigl[f'(t)(x-t)\Bigr]_a^x - \int_a^x f^{(2)}(t)(x-t) \hspace{0.2em} dt \Biggr) $$$$ f(x) = f(a) + f'(a)(x-a) + \int_a^x f^{(2)}(t)(x-t)dt $$Then we do it again with:
$$ \Biggl \{ \begin{gather*} u(t) = f^{(2)}(t) \\ v'(t) = (x-t)dt \end{gather*} $$$$ \Biggl \{ \begin{gather*} u'(t) = f^{(3)}(t)dt \\ v(t) = -\frac{(x-t)^2}{2} \end{gather*} $$$$ f(x) = f(a) + f'(a)(x-a) + \Biggl[ -\frac{f^{(2)}(t)}{2} (x-t)^2 \Biggr]_a^x + \int_a^x \frac{f^{(3)}(t)}{2}(x-t)^2dt $$$$ f(x) = f(a) + f'(a)(x-a) + \frac{f^{(2)}(a)}{2} (x-a)^2 + \int_a^x \frac{f^{(3)}(t)}{2}(x-t)^2dt$$And so on...
$$ f(x) = f(a) + f'(a)(x-a) + \frac{f^{(2)}(a)}{2} (x-a)^2 + \Biggl[ \frac{f^{(3)}(t)}{6} (x-t)^3 \Biggr]_a^x - \int_a^x \frac{f^{(4)}(t)}{6}(x-t)^3dt $$$$ f(x) = f(a) + f'(a)(x-a) + \frac{f^{(2)}(a)}{2} (x-a)^2 + \frac{f^{(3)}(a)}{6} (x-a)^3 + \int_a^x \frac{f^{(4)}(t)}{6}(x-t)^3dt \qquad (2) $$ -
From the Integral Remainder to Landau Notation
To arrive at the final form of Taylor-Young's formula, we must study the asymptotic behavior of this integral remainder as \(x\) approaches \(a\). Let us denote this remainder as \(R_3(x)\):
$$R_3(x) = \int_a^x \frac{f^{(4)}(t)}{6}(x-t)^3 \hspace{0.1em} dt $$The objective is to demonstrate that this remainder is negligible compared to \((x-a)^3\) in the neighborhood of \(a\), which is expressed in Landau notation as \(o\bigl((x - a)^3\bigr)\). By definition, we must therefore verify that:
$$\lim_{x \to a} \frac{R_3(x)}{(x-a)^3} = 0$$ -
Evaluating the Integral
Since the function \(f^{(4)}\) is continuous on the closed interval \(\bigl[a; x \bigr]\), it is bounded on it.
Let \(M_x\) be the maximum of its absolute value over this interval: \(M_x = \max |f^{(4)}(t)|\).
By using the triangle inequality property for integrals , we can bound our remainder from above:
$$\Bigl| R_3(x) \Bigr| = \left| \int_a^x \frac{f^{(4)}(t)}{6}(x-t)^3 \hspace{0.1em} dt \right| \leqslant \frac{M_x}{6} \left| \int_a^x (x-t)^3 \hspace{0.1em} dt \right|$$We can explicitly compute the right-hand integral with respect to the dummy variable \(t\):
$$\int_a^x (x-t)^3 \hspace{0.1em} dt = \left[ -\frac{(x-t)^4}{4} \right]_a^x = 0 - \left(-\frac{(x-a)^4}{4}\right) = \frac{(x-a)^4}{4}$$Substituting this result back into our inequality, we obtain the error bound:
$$\Bigl| R_3(x) \Bigr| \leqslant \frac{M_x}{24}(x-a)^4$$ -
Passage to the Limit
Let us now divide this expression by \((x-a)^3\) to analyze the convergence ratio:
$$\left| \frac{R_3(x)}{(x-a)^3} \right| \leqslant \frac{M_x}{24} \cdot \frac{(x-a)^4}{(x-a)^3} = \frac{M_x}{24}(x-a)$$As \(x \to a\), the term \((x-a)\) clearly tends to \(0\), while the bound \(M_x\) converges to the finite value \(|f^{(4)}(a)|\) by continuity.
By the squeeze theorem , we immediately deduce:
$$\lim_{x \to a} \frac{R_3(x)}{(x-a)^3} = 0 \quad \Longleftrightarrow \quad R_3(x) \underset{x \to a}{=} o\bigl((x-a)^3\bigr)$$The integral remainder locally vanishes into a little-\(o\), yielding the final \(3\) rd -order Taylor-Young formula:
$$f(x) = f(a) + f'(a)(x-a) + \frac{f^{(2)}(a)}{2} (x-a)^2 + \frac{f^{(3)}(a)}{6} (x-a)^3 + o\bigl((x-a)^3\bigr)$$ -
Generalization to the \(n\)-th Order
Following a strictly identical logic, by reiterating this integration by parts not \(3\) times, but \(n\) times, we can prove the general case by induction:
$$ \forall n \in \mathbb{N}, \enspace \forall f \in \hspace{0.04em} \mathcal{C}^n[D_f, \mathbb{R}], \enspace \forall (x, a) \in D_f^2, $$$$ f(x) = f(a) + f'(a)(x - a) + \frac{f^{(2)}(a)}{2!}(x - a)^2 + \ ... \ + \hspace{0.2em} \frac{f^{(n)}(a)}{n!}(x - a)^n + o\bigl((x - a)^n\bigr) \qquad \bigl(\text{Taylor-Young Formula}\bigr) $$Where the \( o\bigl((x - a)^n\bigr) \) notation represents the Landau remainder (a "little-o" term), meaning that the remainder is negligible compared to \( (x - a)^n \) as \( x \to a \):
$$ \lim_{x \to a} \frac{R_{n,a}(x)}{(x - a)^n} = 0 $$
Furthermore, by setting \( (x = a + h) \), the formula can be written in this alternative form:
$$ f(a + h) = f(a) + f'(a)h + \frac{f^{(2)}(a)}{2!}h^2 + \ ... \ + \hspace{0.2em} \frac{f^{(n)}(a)}{n!}h^n + o\bigl(h^n\bigr) \qquad \bigl(\text{Taylor-Young Formula}\bigr)^* $$
Taylor-Laplace Formula
If we take the previous equation \((2)\):
By continuing to integrate successively using the same method, the general formula follows naturally:
Which can be written as,
When seeking to obtain a global estimation or a precise error bound over an entire interval, Taylor's theorem with integral remainder (or the Taylor-Laplace's formula) is used.
Unlike Taylor-Young's formula, it requires a slightly stronger regularity hypothesis (\( f \) must be of class \( \mathcal{C}^{n+1} \)) in order to explicitly define the remainder in an integral form, known as the Laplace remainder.
Examples
-
The sinus function\(: f(x) = \sin(x) \)
-
Taylor series of order 3 at zero\(: T_3(0) \)
Let us use the previously demonstrated method to calculate a Taylor expansions of the \( \sin(x) \) function.
We firstly verify that \( \sin(x) \) can be derivated three times in a row. It is well-known that is the case.
Then, let us calculate the successive derivatives of order \(3 \), and retrieve all of these images at \( a = 0 \).
$$f(x) = \sin(x) \Longrightarrow \sin(0) = 0 $$$$f'(x) = \cos(x) \Longrightarrow \cos(0) = 1$$$$f^{(2)}(x) = -\sin(x) \Longrightarrow -\sin(0) = 0$$$$f^{(3)}(x) = -\cos(x) \Longrightarrow -\cos(0) = -1 $$Now, we apply the Taylor-Young's formula.
$$ f(x) = f(a) + f'(a)(x - a) + \frac{f^{(2)}(a)}{2!}(x - a)^2 + \ ... \ + \hspace{0.2em} \frac{f^{(n)}(a)}{n!}(x - a)^n + R_{n, a}(x) $$In our case, that would be:
$$ \sin(x) \underset{0}{=} \sin(0) + \cos(0)(x - 0) - \frac{\sin(0)}{2!}(x - 0)^2 - \frac{\cos(0)}{3!}(x - 0)^3 + R_{3, 0}(x) $$$$ \sin(x) \underset{0}{=} x -\frac{1}{6}x^3 +R_{3, 0}(x) $$ -
Framing with integral remainder
$$ \\sin(x) \underset{0}{=} x -\frac{1}{6}x^3 +R_{3, 0}(x) $$We have seen above that this remainder is worth:
$$ R_{3, 0}(x) = \int_0^x \frac{\sin^{(4)}(t)}{3!}(x-t)^3 \hspace{0.2em} dt $$But, \(\sin^{(4)}(t) = \sin(t)\). So we now have a new expression for \( R_{3, 0}(x)\):
$$ R_{3, 0}(x) = \int_0^x \frac{\sin(t)}{6}(x-t)^3 \hspace{0.2em} dt $$Let us now frame this remainder in the interval \( [-\pi, \pi]\).
$$ -\pi \leqslant t \leqslant \pi$$$$ -1 \leqslant \sin(t) \leqslant 1$$$$ \frac{-(x-t)^3}{6} \hspace{0.2em} \leqslant \hspace{0.2em} \frac{\sin(t)}{6}(x-t)^3 \hspace{0.2em} \leqslant \hspace{0.2em} \frac{-(x-t)^3}{6} $$Using the property of growth of an integral , we do have:
$$ -\int_0^x \frac{(x-t)^3}{6} \hspace{0.1em}dt \hspace{0.2em} \leqslant \hspace{0.2em} \int_0^x \frac{\sin(t)}{6}(x-t)^3 \hspace{0.1em}dt \hspace{0.2em} \leqslant \hspace{0.2em} \int_0^x \frac{(x-t)^3}{6} \hspace{0.1em}dt$$$$ -\Biggl[ -\frac{(x-t)^4}{24} \Biggr]_0^x \hspace{0.2em} \leqslant \hspace{0.2em} R_{3, 0}(x) \hspace{0.2em} \leqslant \hspace{0.2em} \Biggl[ -\frac{(x-t)^4}{24} \Biggr]_0^x $$$$ -\frac{x^4}{24} \hspace{0.2em} \leqslant \hspace{0.2em} R_{3, 0}(x) \hspace{0.2em} \leqslant \hspace{0.2em} \frac{x^4}{24} $$Yet, we know thanks to \((1)\) that:
$$ sin_{3, 0}(x) = \enspace T_{3,0}(x) + R_{3,0}(x) $$This leads us to a framing for \( sin_{3, 0}(x)\) :
$$ T_{3, 0}(x) -\frac{x^4}{24} \hspace{0.2em} \leqslant \hspace{0.2em} T_{3, 0}(x) + R_{3, 0}(x) \hspace{0.2em} \leqslant \hspace{0.2em} T_{3, 0}(x) + \frac{x^4}{24} $$$$ x -\frac{1}{6}x^3 -\frac{x^4}{24} \hspace{0.2em} \leqslant \hspace{0.2em} \sin_{3, 0}(x) \hspace{0.2em} \leqslant \hspace{0.2em} x -\frac{1}{6}x^3 + \frac{x^4}{24} $$ -
Taylor series of order n at zero\(: T_n(0) \)
Performing an Taylor expansions of order \(n\) at \((x=0)\) for the \(\sin(x)\) function, we obtain:
$$ \sin(x) = x -\frac{x^3 }{3!}+ \frac{x^5}{5!} - \frac{x^7}{7!} + \ ... \ + \hspace{0.2em} (-1)^{n} \frac{x^{2n+1}}{(2n+1)!} + R_{n, 0}(x) $$Moreover, the remainder of this Taylor expansions is worth:
$$ R_{n, 0}(x) = \int_0^x \frac{\sin^{(2n+2)}(t)}{(2n+1)!}(x-t)^{2n+1} \hspace{0.2em} dt $$$$ R_{n, 0}(x) = \int_0^x \frac{(-1)^{n+1}\sin(t)}{(2n+1)!}(x-t)^{2n+1} \hspace{0.2em} dt $$With the Taylor-Lagrange's inequality , we can frame this remainder.
The Taylor-Lagrange's inequality tell us that:
Let \( f \) be a function of class \( \mathcal{C}^{n+1} \) on an interval \( I \) of \( \mathbb{R} \) (with \( n \in \mathbb{N} \)).
We name \( M_{n+1} \) a majorant which is worth the absolute value of the \( (n+1) \)-th derivative of \( f \) on this interval:$$ \forall t \in I, \quad \left| f^{(n+1)}(t) \right| \leqslant M_{n+1} $$Then, for any couple of real numbers \( (a, x) \in I^2 \), the Taylor-Laplace remainder \( R_{n,a}(x) \) verifies the following inequality:
$$ \Bigl| R_{n,a}(x) \Bigr| \leqslant M_{n+1} \frac{|x - a|^{n+1}}{(n+1)!} $$By applying triangle inequality property for integrals on the previous expression:
$$ \Bigl|R_{n, 0}(x) \Bigr| = \Bigl| \int_0^x \frac{(-1)^{n+1}\sin(t)}{(2n+1)!}(x-t)^{2n+1} \hspace{0.1em} dt \Bigr| \leqslant \int_0^x \frac{\bigl|(-1)^{n+1}\sin(t)\bigr|}{(2n+1)!}\bigl|(x-t)^{2n+1}\bigr| \hspace{0.1em} dt $$But, for any real number \( t \), we know that \( \bigl|\sin(t)\bigr| \leqslant 1 \). We can thus set a major upper bound \( M = 1 \) and take it out ou of the integral:
$$ \Bigl|R_{n, 0}(x) \Bigr| \leqslant \frac{1}{(2n+1)!} \int_0^x (x-t)^{2n+1} \hspace{0.1em} dt $$The right member integration gives directly:
$$ \int_0^x (x-t)^{2n+1} \hspace{0.1em} dt = \Biggl[ -\frac{(x-t)^{2n+2}}{2n+2} \Biggr]_0^x = 0 - \left( -\frac{x^{2n+2}}{2n+2} \right) = \frac{x^{2n+2}}{2n+2} $$Combining these results, we definitely find the classic the Taylor-Lagrange's inequality :
$$ \Bigl|R_{n, 0}(x) \Bigr| \leqslant \frac{1}{(2n+1)!} \cdot \frac{x^{2n+2}}{2n+2} \ \Longrightarrow \ \Bigl|R_{n, 0}(x) \Bigr| \leqslant \frac{x^{2n+2}}{(2n+2)!} $$Taking this equality to the limit when \( n \to +\infty\) we obtain:
$$ \lim_{n \to +\infty} \enspace \Bigl |R_{n, 0}(x) \Bigr| \hspace{0.2em} \leqslant \hspace{0.2em} \lim_{n \to +\infty} \enspace \frac{x^{2n+2}}{(2n+2)!} $$By compared growth of the limits , the factorial function outweighs the power of x function :
$$ \lim_{n \to +\infty} \enspace \Bigl |R_{n, 0}(x) \Bigr| \hspace{0.2em} \leqslant 0 $$In the end, with the squeeze theorem :
$$ \lim_{n \to +\infty} \enspace R_{n, 0}(x) = 0 $$So as a result, a Taylor expansions of the \(\sin(x) \) function is worth:
$$ \sin(x) = x -\frac{x^3 }{3!}+ \frac{x^5}{5!} - \frac{x^7}{7!} + \ ... \ + \hspace{0.2em} (-1)^{n} \frac{x^{2n+1}}{(2n+1)!} + R_{n, 0}(x) \qquad (\text{avec} \enspace \lim_{n \to +\infty} \enspace R_{n, 0}(x) = 0) $$
-
Back to top