From Wikipedia, the free encyclopedia
I have put this 'proof' here as an interesting result I stumbled upon a while back. I have never seen it in publication (but then again I don't read too many books), nor have I had an expert check it. The algebra is watertight, but some of my assumptions may require more proof. Nonetheless it is an interesting take on a subject you probably don't give much thought too. (I recommend reading Order 2/Method 2 and Order N) Ed g2s 02:28, 5 Jul 2004 (UTC)
Some functions can be differentiated without the need for Newton's difference quotient and without having to study limits .
By constructing suitable tangent lines, all polynomials can be differentiated.
f
(
x
)
=
a
x
+
b
{\displaystyle \ f(x)=ax+b}
The gradient of this line is
a
{\displaystyle \ a}
which is therefore the derivative at any point.
f
′
(
x
)
=
a
{\displaystyle \ f^{\prime }(x)=a}
f
(
x
)
=
a
x
2
+
b
x
+
c
{\displaystyle \ f(x)=ax^{2}+bx+c}
Consider a line
g
(
x
)
{\displaystyle \ g(x)}
such that it intersects the quadratic twice at
x
=
x
1
{\displaystyle \ x=x_{1}}
(a double root). This is the tangent to the curve at this point.
Subtracting the two equations will give a quadratic with roots
x
1
,
x
1
{\displaystyle \ x_{1},x_{1}}
, meaning it will be of the form:
f
(
x
)
−
g
(
x
)
=
A
(
x
−
x
1
)
(
x
−
x
1
)
{\displaystyle \ f(x)-g(x)=A(x-x_{1})(x-x_{1})}
where A is a non-zero constant. We can choose
A
=
a
{\displaystyle \ A=a}
.
g
(
x
)
=
f
(
x
)
−
a
(
x
−
x
1
)
2
{\displaystyle \ g(x)=f(x)-a(x-x_{1})^{2}}
⇒
g
(
x
)
=
a
x
2
+
b
x
+
c
−
a
x
2
+
2
a
x
1
x
−
a
x
1
2
{\displaystyle \Rightarrow g(x)=ax^{2}+bx+c-ax^{2}+2ax_{1}x-ax_{1}^{2}}
⇒
g
(
x
)
=
(
2
a
x
1
+
b
)
x
+
c
−
a
x
1
2
{\displaystyle \Rightarrow g(x)=(2ax_{1}+b)x+c-ax_{1}^{2}}
The gradient of this line is
2
a
x
1
+
b
{\displaystyle \ 2ax_{1}+b}
. This is also the rate of change of the function
f
(
x
)
{\displaystyle \ f(x)}
at the point
x
=
x
1
{\displaystyle \ x=x_{1}}
.
f
′
(
x
1
)
=
g
′
(
x
1
)
=
2
a
x
1
+
b
{\displaystyle \ f^{\prime }(x_{1})=g^{\prime }(x_{1})=2ax_{1}+b}
To conclude:
f
(
x
)
=
a
x
2
+
b
x
+
c
{\displaystyle \ f(x)=ax^{2}+bx+c}
⇒
f
′
(
x
)
=
2
a
x
+
b
{\displaystyle \Rightarrow f^{\prime }(x)=2ax+b}
f
(
x
)
=
a
x
2
+
b
x
+
c
{\displaystyle \ f(x)=ax^{2}+bx+c}
Let us imagine a tangent line that intersects the above quadratic once and only once where
x
=
x
1
{\displaystyle \ x=x_{1}}
. This tangent line will be of the form:
g
(
x
)
=
m
x
+
n
{\displaystyle \ g(x)=mx+n}
The difference of the two functions:
f
(
x
)
−
g
(
x
)
=
a
x
2
+
(
b
−
m
)
x
+
(
c
−
n
)
{\displaystyle \ f(x)-g(x)=ax^{2}+(b-m)x+(c-n)}
will be a quadratic that has a double root at
x
=
x
1
{\displaystyle \ x=x_{1}}
f
(
x
1
)
−
g
(
x
1
)
=
a
x
1
2
+
(
b
−
m
)
x
1
+
(
c
−
n
)
=
0
{\displaystyle \ f(x_{1})-g(x_{1})=ax_{1}^{2}+(b-m)x_{1}+(c-n)=0}
⇒
c
=
n
−
(
b
−
m
)
x
1
−
a
x
1
2
(
∗
)
{\displaystyle \Rightarrow c=n-(b-m)x_{1}-ax_{1}^{2}\ \ \ (*)}
using the quadratic equation:
x
=
−
b
±
b
2
−
4
a
c
2
a
=
x
1
{\displaystyle x={\frac {-b\pm {\sqrt {b^{2}-4ac}}}{2a}}=x_{1}}
As
x
1
{\displaystyle \ x_{1}}
is a double root, the discriminant of the equation is equal to zero.
b
2
−
4
a
c
=
0
{\displaystyle \ b^{2}-4ac=0}
⇒
(
b
−
m
)
2
−
4
a
(
c
−
n
)
=
0
{\displaystyle \Rightarrow (b-m)^{2}-4a(c-n)=0}
Substituting in
c
{\displaystyle \ c}
from
(
∗
)
{\displaystyle \ (*)}
:
m
2
−
2
b
m
+
b
2
−
4
a
(
n
−
(
b
−
m
)
x
1
−
a
x
1
2
−
n
)
=
0
{\displaystyle m^{2}-2bm+b^{2}-4a(n-(b-m)x_{1}-ax_{1}^{2}-n)=0}
⇒
m
2
−
(
4
a
x
1
+
2
b
)
m
+
(
b
2
−
4
a
2
x
1
2
)
{\displaystyle \Rightarrow m^{2}-(4ax_{1}+2b)m+(b^{2}-4a^{2}x_{1}^{2})}
Solving this as a quadratic equation in
m
{\displaystyle \ m}
:
m
=
(
4
a
x
1
+
2
b
)
±
(
4
a
x
1
+
2
b
)
2
−
4
(
b
2
−
4
a
2
x
1
2
)
2
{\displaystyle m={\frac {(4ax_{1}+2b)\pm {\sqrt {(4ax_{1}+2b)^{2}-4(b^{2}-4a^{2}x_{1}^{2})}}}{2}}}
m
=
2
a
x
1
+
b
{\displaystyle \ m=2ax_{1}+b}
Which gradient of the quadratic at
x
=
x
1
{\displaystyle \ x=x_{1}}
. To conclude:
f
(
x
)
=
a
x
2
+
b
x
+
c
{\displaystyle \ f(x)=ax^{2}+bx+c}
⇒
f
′
(
x
)
=
2
a
x
+
b
{\displaystyle \Rightarrow f^{\prime }(x)=2ax+b}
f
(
x
)
=
a
x
3
+
b
x
2
+
c
x
+
d
{\displaystyle \ f(x)=ax^{3}+bx^{2}+cx+d}
Consider a quadratic
g
(
x
)
{\displaystyle \ g(x)}
such that it intersects the cubic twice at
x
=
x
1
{\displaystyle \ x=x_{1}}
(a double root) and once at
x
=
x
2
{\displaystyle \ x=x_{2}}
. You can think of this quadratic as a tangent curve .
Subtracting the two equations will give a cubic with roots
x
1
,
x
1
,
x
2
{\displaystyle \ x_{1},x_{1},x_{2}}
, meaning it will be of the form:
f
(
x
)
−
g
(
x
)
=
A
(
x
−
x
1
)
(
x
−
x
1
)
(
x
−
x
2
)
{\displaystyle \ f(x)-g(x)=A(x-x_{1})(x-x_{1})(x-x_{2})}
where A is a non-zero constant. We can choose
A
=
a
{\displaystyle \ A=a}
.
g
(
x
)
=
f
(
x
)
−
a
(
x
−
x
1
)
2
(
x
−
x
2
)
{\displaystyle \ g(x)=f(x)-a(x-x_{1})^{2}(x-x_{2})}
⇒
g
(
x
)
=
(
a
(
2
x
1
+
x
2
)
+
b
)
x
2
+
(
c
−
a
x
1
(
x
1
+
2
x
2
)
)
x
+
a
x
1
2
x
2
+
d
{\displaystyle \Rightarrow g(x)=(a(2x_{1}+x_{2})+b)x^{2}+(c-ax_{1}(x_{1}+2x_{2}))x+ax_{1}^{2}x_{2}+d}
Using the result for Order 2 polynomials:
g
′
(
x
)
=
2
(
a
(
2
x
1
+
x
2
)
+
b
)
x
−
a
x
1
(
x
1
+
2
x
2
)
+
c
{\displaystyle \ g^{\prime }(x)=2(a(2x_{1}+x_{2})+b)x-ax_{1}(x_{1}+2x_{2})+c}
As
f
(
x
)
{\displaystyle \ f(x)}
and
g
(
x
)
{\displaystyle \ g(x)}
share a tangent at
x
=
x
1
{\displaystyle \ x=x_{1}}
f
′
(
x
1
)
=
g
′
(
x
1
)
=
2
(
a
(
2
x
1
+
x
2
)
+
b
)
x
1
−
a
x
1
(
x
1
+
2
x
2
)
+
c
{\displaystyle \ f^{\prime }(x_{1})=g^{\prime }(x_{1})=2(a(2x_{1}+x_{2})+b)x_{1}-ax_{1}(x_{1}+2x_{2})+c}
f
′
(
x
1
)
=
3
a
x
1
2
+
2
b
x
1
+
c
{\displaystyle \ f^{\prime }(x_{1})=3ax_{1}^{2}+2bx_{1}+c}
To conclude:
f
(
x
)
=
a
x
3
+
b
x
2
+
c
x
+
d
{\displaystyle \ f(x)=ax^{3}+bx^{2}+cx+d}
⇒
f
′
(
x
)
=
3
a
x
2
+
2
b
x
+
c
{\displaystyle \Rightarrow f^{\prime }(x)=3ax^{2}+2bx+c}
To prove differentiation for a general polynomial an inductive proof can be used.
Assume true for
n
=
k
{\displaystyle \ n=k}
:
d
d
x
∑
r
=
0
k
a
r
x
r
=
∑
r
=
0
k
r
a
r
x
r
−
1
{\displaystyle {\frac {d}{dx}}\sum _{r=0}^{k}a_{r}x^{r}=\sum _{r=0}^{k}ra_{r}x^{r-1}}
If we can show it is true for
n
=
k
+
1
{\displaystyle \ n=k+1}
using the above assumption then we will have shown it true for all
n
∈
N
{\displaystyle \ n\in \mathbb {N} }
as we've shown it true for the first few cases.
We need to show that:
f
(
x
)
=
A
x
k
+
1
+
∑
r
=
0
k
a
r
x
r
{\displaystyle \ f(x)=Ax^{k+1}+\sum _{r=0}^{k}a_{r}x^{r}}
differentiates to:
f
′
(
x
)
=
A
(
k
+
1
)
x
k
+
∑
r
=
0
k
r
a
r
x
r
−
1
{\displaystyle \ f^{\prime }(x)=A(k+1)x^{k}+\sum _{r=0}^{k}ra_{r}x^{r-1}}
Consider a polynomial
g
(
x
)
{\displaystyle \ g(x)}
order
k
{\displaystyle \ k}
such that it intersects
f
(
x
)
{\displaystyle \ f(x)}
twice at
x
=
x
1
{\displaystyle \ x=x_{1}}
(a double root) and for simplicity the remaining
k
−
1
{\displaystyle \ k-1}
intersections are all at
x
=
x
2
{\displaystyle \ x=x_{2}}
.
Subtracting the two equations will give the following result.
f
(
x
)
−
g
(
x
)
=
A
(
x
−
x
1
)
2
(
x
−
x
2
)
k
−
1
{\displaystyle \ f(x)-g(x)=A(x-x_{1})^{2}(x-x_{2})^{k-1}}
⇒
g
(
x
)
=
A
x
k
+
1
−
A
(
x
−
x
1
)
2
(
x
−
x
2
)
k
−
1
+
∑
r
=
0
k
a
r
x
r
{\displaystyle \Rightarrow g(x)=Ax^{k+1}-A(x-x_{1})^{2}(x-x_{2})^{k-1}+\sum _{r=0}^{k}a_{r}x^{r}}
Using the binomial theorem :
g
(
x
)
=
A
x
k
+
1
−
A
(
x
2
−
2
x
1
x
+
x
1
2
)
∑
r
=
0
k
−
1
(
k
−
1
C
r
(
−
x
2
)
k
−
1
−
r
x
r
)
+
∑
r
=
0
k
a
r
x
r
{\displaystyle g(x)=Ax^{k+1}-A(x^{2}-2x_{1}x+x_{1}^{2})\sum _{r=0}^{k-1}(^{k-1}C_{r}(-x_{2})^{k-1-r}x^{r})+\sum _{r=0}^{k}a_{r}x^{r}}
Taking out the last term of the sum (where
r
=
k
−
1
{\displaystyle \ r=k-1}
)
g
(
x
)
=
A
x
k
+
1
−
A
(
x
2
−
2
x
1
x
+
x
1
2
)
x
k
−
1
−
∑
r
=
0
k
−
2
(
k
−
1
C
r
(
−
x
2
)
k
−
1
−
r
x
r
⋅
A
(
x
2
−
2
x
1
x
+
x
1
2
)
)
+
∑
r
=
0
k
a
r
x
r
{\displaystyle g(x)=Ax^{k+1}-A(x^{2}-2x_{1}x+x_{1}^{2})x^{k-1}-\sum _{r=0}^{k-2}\ \left(\ ^{k-1}C_{r}(-x_{2})^{k-1-r}x^{r}\cdot A(x^{2}-2x_{1}x+x_{1}^{2})\right)+\sum _{r=0}^{k}a_{r}x^{r}}
⇒
g
(
x
)
=
2
A
x
1
x
k
−
A
x
1
2
x
k
−
1
−
A
∑
r
=
0
k
−
2
(
k
−
1
C
r
(
−
x
2
)
k
−
1
−
r
(
x
r
+
2
−
2
x
1
x
r
+
1
+
x
1
2
x
r
)
)
+
∑
r
=
0
k
a
r
x
r
{\displaystyle \Rightarrow g(x)=2Ax_{1}x^{k}-Ax_{1}^{2}x^{k-1}-A\sum _{r=0}^{k-2}\ \left(\ ^{k-1}C_{r}(-x_{2})^{k-1-r}(x^{r+2}-2x_{1}x^{r+1}+x_{1}^{2}x^{r})\right)+\sum _{r=0}^{k}a_{r}x^{r}}
As all the terms are now of order
k
{\displaystyle \ k}
or less we can differentiate the expression using the assumption above:
g
′
(
x
)
=
2
A
k
x
1
x
k
−
1
−
A
(
k
−
1
)
x
1
2
x
k
−
2
−
A
∑
r
=
0
k
−
2
(
k
−
1
C
r
(
−
x
2
)
k
−
1
−
r
(
(
r
+
2
)
x
r
+
1
−
2
(
r
+
1
)
x
1
x
r
+
r
x
1
2
x
r
−
1
)
)
+
∑
r
=
0
k
r
a
r
x
r
−
1
{\displaystyle g^{\prime }(x)=2Akx_{1}x^{k-1}-A(k-1)x_{1}^{2}x^{k-2}-A\sum _{r=0}^{k-2}\ \left(\ ^{k-1}C_{r}(-x_{2})^{k-1-r}\left((r+2)x^{r+1}-2(r+1)x_{1}x^{r}+rx_{1}^{2}x^{r-1}\right)\ \right)+\sum _{r=0}^{k}ra_{r}x^{r-1}}
As
f
(
x
)
{\displaystyle \ f(x)}
and
g
(
x
)
{\displaystyle \ g(x)}
share a tangent at
x
=
x
1
{\displaystyle \ x=x_{1}}
f
′
(
x
1
)
=
g
′
(
x
1
)
=
2
A
k
x
k
−
A
k
x
k
+
A
x
k
−
A
∑
r
=
0
k
−
2
(
k
−
1
C
r
(
−
x
2
)
k
−
1
−
r
(
r
+
2
−
2
(
r
+
1
)
+
r
)
x
r
+
2
)
+
∑
r
=
0
k
r
a
r
x
r
−
1
{\displaystyle f^{\prime }(x_{1})=g^{\prime }(x_{1})=2Akx^{k}-Akx^{k}+Ax^{k}-A\sum _{r=0}^{k-2}\ \left(\ ^{k-1}C_{r}(-x_{2})^{k-1-r}(r+2-2(r+1)+r)x^{r+2}\right)+\sum _{r=0}^{k}ra_{r}x^{r-1}}
As
(
r
+
2
−
2
(
r
+
1
)
+
r
)
=
0
{\displaystyle \ (r+2-2(r+1)+r)=0}
the entire sum
A
∑
r
=
0
k
−
2
(
.
.
.
)
{\displaystyle \ A\sum _{r=0}^{k-2}\ \left(...\right)}
cancels out giving:
f
′
(
x
)
=
A
(
k
+
1
)
x
k
+
∑
r
=
0
k
r
a
r
x
r
−
1
{\displaystyle \ f^{\prime }(x)=A(k+1)x^{k}+\sum _{r=0}^{k}ra_{r}x^{r-1}}
By induction we have proved the result
d
d
x
∑
r
=
0
n
a
r
x
r
=
∑
r
=
0
n
r
a
r
x
r
−
1
{\displaystyle {\frac {d}{dx}}\sum _{r=0}^{n}a_{r}x^{r}=\sum _{r=0}^{n}ra_{r}x^{r-1}}
is true for all
n
∈
N
{\displaystyle \ n\in \mathbb {N} }
.