logo

Proof of Leibniz's Theorem 📂Analysis

Proof of Leibniz's Theorem

Theorem

ddx(fg)=dfdxg+fdgdx \dfrac{d}{dx} (fg)=\dfrac{df}{dx}g+f\dfrac{dg}{dx}

dndxn(fg)=k=0nn!(nk)!k!dnkfdxnkdkgdxk=k=0nnCkdnkfdxnkdkgdxk=k=0n(nk)dnkfdxnkdkgdxk \begin{align*} \dfrac{d^n}{dx^n}(fg)&=\sum \limits_{k=0}^{n}\frac{n!}{(n-k)!k!}\dfrac{d^{n-k}f}{dx^{n-k}}\dfrac{d^k g}{dx^k} \\ &=\sum \limits_{k=0}^{n}{}_{n}\mathrm{C}_{k} \dfrac{d^{n-k}f}{dx^{n-k}}\dfrac{d^k g}{dx^k} \\ &=\sum \limits_{k=0}^{n} \binom{n}{k} \dfrac{d^{n-k}f}{dx^{n-k}}\dfrac{d^k g}{dx^k} \end{align*}

Description

Also known as Leibniz’s rule.

The first equation is a well-known formula, often referred to as the product rule or the rule of product for differentiation. It simply expresses the result when the product of two functions is differentiated once. More generally, the equation below represents the result when differentiated nn times. Since a polynomial can become 00 by being differentiated repeatedly, the result can be easily calculated without directly differentiating nn times.

There are also many theorems or formulas related to differentiation and integration named after Leibniz.

Proof

Let’s call DD the following differential operator.

D=ddx D=\dfrac{d}{dx}

For example, Df(x)=df(x)dxDf(x)=\dfrac{df(x)}{dx} is. If DD is used to express the differentiation of fgfg, it looks like the following.

ddx(fg)=gDf+fDg \dfrac{d}{dx}(fg)=gDf+fDg

At this time, let’s say DfD_{f} is an operator that only applies to ff, and DgD_{g} is an operator that only applies to gg. Then, the above formula is expressed as follows.

(Df+Dg)(fg)=gDf+fDg (D_{f}+D_{g})(fg)=gDf+fDg

Then you get:

ddx(fg)=(Df+Dg)(fg) \dfrac{d}{dx}(fg)=(D_{f}+D_{g})(fg)

d2dx2(fg)=D(Df+Dg)(fg) \dfrac{d^2}{dx^2}(fg)=D(D_{f}+D_{g})(fg)

At this point, since DD is a differential operator, the order of operations doesn’t matter. In other words, it means DDff=DfDfDD_{f}f=D_{f}Df. Thus, the formula is as follows.

d2dx2(fg)=D(Df+Dg)(fg)=(Df+Dg)D(fg)=(Df+Dg)(Df+Dg)(fg)=(Df+Dg)2(fg) \begin{align*} \dfrac{d^2}{dx^2}(fg) &= D(D_{f}+D_{g})(fg) \\ &= (D_{f}+D_{g})D(fg) \\ &= (D_{f}+D_{g})(D_{f}+D_{g})(fg) \\ &= (D_{f}+D_{g})^2 (fg) \end{align*}

As mentioned above, since DD allows for the commutativity of multiplication, it can be expressed as the last line. Extended to nn times of differentiation, it looks like the following.

dndxn(fg)=(Df+Dg)n(fg) \dfrac{d^n}{dx^n} (fg)=(D_{f}+D_{g})^n(fg)

Since the commutative law holds, the binomial theorem can be applied. Using the binomial theorem, you get:

dndxn(fg)=(Df+Dg)n(fg)=k=0nnCkDfnkDgk(fg)=k=0nnCkDfnkfDgkg=k=0nnCkdnkfdxnkdkgdxk=k=0nnCkf(nk)g(k) \begin{align*} \dfrac{d^n}{dx^n} (fg) &= (D_{f}+D_{g})^n(fg) \\ &= \sum \limits_{k=0} ^n {}_{n}\mathrm{C} _{k} {D_{f}}^{n-k} {D_{g}}^{k}(fg) \\ &=\sum \limits_{k=0} ^n {}_{n}\mathrm{C} _{k} {D_{f}}^{n-k} f{D_{g}}^{k}g \\ &= \sum \limits_{k=0} ^n {}_{n}\mathrm{C} _{k} \dfrac{d^{n-k}f}{dx^{n-k}} \dfrac{d^k g}{dx^k} \\ &= \sum \limits_{k=0} ^n {}_{n}\mathrm{C} _{k} f^{(n-k)} g^{(k)} \end{align*}

Example

1

  • Find d7dx7(xsinx)\dfrac{d^7}{dx^7}( x \sin x).

If xx and sinx\sin x are set respectively as gg and ff from the proof above, by Leibniz’s rule,

d7dx7(xsinx)=k=077Ckdnkdxnk(sinx)dkdxk(x) \dfrac{d^7}{dx^7}( x \sin x)=\sum \limits_{k=0} ^7 {}_{7} \mathrm{C}_{k} \dfrac{d ^{n-k} } {dx^{n-k} }(\sin x) \dfrac{d^k}{dx^k} (x)

When k2k \ge 2, since it is dkdxk(x)=0\dfrac{d^k}{dx^k}(x)=0, only the two terms k=0,1k=0,1 remain. Therefore,

d7dx7(xsinx)=7C0d7dx7(sinx)x+7C1d6dx6(sinx)=xcosx7sinx \begin{align*} \dfrac{d^7}{dx^7} ( x \sin x ) &= {}_{7} \mathrm{C} _{0} \dfrac{d^7}{dx^7}(\sin x) x + {}_{7}\mathrm{C}_{1} \dfrac{d^6}{dx^6} (\sin x) \\ &= -x \cos x -7\sin x \end{align*}

2

  • Find d10dx10(x2ex)\dfrac{d^{10}}{dx^{10}} ( x^2 e^{-x} ).

If x2x^2 and exe^{-x} are set respectively as gg and ff from the proof above, by Leibniz’s rule,

d10dx10(x2ex)=k=01010Ckd10kdx10k(ex)dkdxk(x2) \dfrac{d^{10}}{dx^{10}} (x^2 e^{-x}) = \sum \limits _{k=0} ^{10} {}_{10} \mathrm{C} _{k} \dfrac{d^{10-k}}{dx^{10-k}}(e^{-x}) \dfrac{d^k}{dx^k} ( x^2)

When k3k \ge 3, since it is dkdxk(x2)=0\dfrac{d^k}{dx^k} (x^2)=0, only the three terms that are k=0,1,2k=0,1,2 remain. Therefore,

d10dx10(x2ex)=10C0d10dx10(ex)x2+10C1d9dx9(ex)ddx(x2)+10C2d8dx8(ex)d2dx2(x2)=x2ex20xex+90ex \begin{align*} \dfrac{d^{10} } {dx^{10} } (x^2 e^{-x}) &= {}_{10} \mathrm{C}_{0} \dfrac{d^{10}}{dx^{10}} (e^{-x}) x^2 + {}_{10} \mathrm{C} _{1} \dfrac{d^9}{dx^9}(e^{-x}) \dfrac{d}{dx}(x^2) + {}_{10}\mathrm{C}_2 \dfrac{d^8}{dx^8} ( e^{-x} ) \dfrac{d^2}{dx^2} (x^2) \\ &= x^2 e^{-x} -20 x e^{-x} + 90e^{-x} \end{align*}

See Also