Sample: Higher Order
Differentiation - An
Introduction
A perspective and journey on nth differentiation
By Joshua Koshy
TABLE OF CONTENTS
Page 3 - Introduction
Page 4 - Exponential Functions
Page 5 - Polynomial Functions
Page 6 - Trigonometric Functions
Page 7 - Logarithmic Functions
Page 8 - Conclusion
Introduction
Computing derivatives is a simple process - it usually
involves a rule such as the power, chain, or product rule to
differentiate a function, or finding the instantaneous rate of
change (d/dx). Occasionally, you may be forced to use the limit
definition. However, when asked about the 40th derivative of a
function, what is the most efficient method to find it? Regular
differentiation does not suffice - it simply takes too long. This
needs a more efficient method - like functions of derivative
higher orders. These are functions that give a general formula to
find the higher order derivative of any given function. Notation
for a higher order derivative is f(n)(x), where (n) is the number
of derivatives (primes) computed. Let’s start with exponential
functions.
Chapter 1 - Exponential Functions
Given f(x) = ex, what is the f(n)(x) of this function? Well, since
the derivative of ex is itself, this means any higher order
derivative is also ex. Therefore, f(n)(x) = ex. Now, for another
function, let’s use f(x) = 2x. We know the d/dx of an exponential
ax is ax(lna). If we do this multiple times, it is just repeated with
more (lna)s. So, f(n)(x) = 2x(ln2)n. Let’s do one more example.
Given f(x) = 22x, by the chain rule, f(1)(x) = 2(22x)(ln2) and f(2)(x)
= 22(22x)(ln2)2. This happens by the chain rule. Observing this
pattern, we know that the rule is f(n)(x) = 22x(2)n(ln2)n.
Chapter 2 - Polynomial Functions
What is the 6th derivative of x6? We can use prior logic to find
the function of higher order. I’m going to introduce a rule to
you: Given f(x) = xb, f(n)(x) = C0b!nxb-n. This is Koshy's
definition of polynomial nth differentiation. C0 is the starting
coefficient, b!n is the exponent term factorial. Here is an
example: 2!2 = 2(1). 2!1 = 2. 2!3 = 2(1)(0). 2!4 = 2(1)(0)(-1). The
subscript indicates how many numbers down you are
multiplying, starting with the number in front. It is similar to a
factorial, but it does not stop at 1. It keeps going. It should also
be noted that N!0=1, N!-1 = 1/N, N!-2 =1/N(N+1), etc - The
reverse factorial. It counts up instead of down. Let’s keep going.
The term xb-n means it is x raised to the power of the beginning
power minus the number of derivatives processed. This formula
gives the higher order derivative of any function xb. There can
be other terms and coefficients - the rule still applies. Now, to
find the 6th derivative of x6, let’s use the formula. C0 = 1, 6!6 =
6!, which is 720. And, xb-n = x6-6, which is 1. Therefore,
multiplying these three yields 720, which is indeed the 6th
derivative of this function. This gives rise to a new theory - if
finding the nth derivative of xn, that means the derivative is a
constant. That also means the (n+1)th derivative is 0.
Chapter 3 - Trigonometric Functions
For sine and cosine, the nth derivative is relatively trivial. For
sine, it is f(n)(x) = sin(x+nℼ/2). This is due to how the 4
combinations of sine derivatives, which are sine, cosine, -sine,
and -cosine, are linear translations of the original sin function.
Refer to the table below:
cos(x) sin(x+ℼ/2)
-sin(x) sin(x+ℼ)
-cos(x) sin(x+3ℼ/2)
See the pattern? Now it becomes fairly obvious why this is the
nth differential for sine. It is a similar story for cosine.
-sin(x) cos(x+ℼ/2)
-cos(x) cos(x+ℼ)
sin(x) cos(x+3ℼ/2)
Therefore, the nth differential for cosine is f(n)(x) = cos(x+nℼ/2).
The other trig functions are nightmares. Do not attempt to find
the nth differential for any of the other trig functions. You will
faint.
Chapter 4 - Logarithmic Functions
Let’s start with the natural logarithm, which is often easier to
work with in terms of differentiating and integrating than other
logarithms due to its nature. Here it is, in all of its glory!
(Ignoring flaws with integration): f(n)(x) = (-1)!n-1x-n. This is
derived from taking the nth differential of 1/x, which is the
derivative of ln(x), and adjusting the n values to fit ln(x).
However, you cannot use this for integration. It will yield the
wrong answer. Now, for log(x), aka the common logarithm,
assuming base 10. f(n)(x) = 1/ln10 (-1)!n-1x-n). The only difference
is the addition of 1/ln10, which originates from differentiating
log(x). It can be re-written as ln(x)/ln10, which when
differentiating, yields 1/xln10. Factor the 1/ln10 and
differentiate as normal.
Conclusion
Thank you so much for taking the time to
read this paper. My name is Joshua
Koshy, an incoming sophomore at
Mountain Lakes High School and an
aspiring astrophysicist. My mission is to
fulfill the world with the intricacies of
mathematics and geometry as a whole.
Written 31 July, 2023
