Review for Test 3

Test 3 is April 26th and will cover the two projects on derivatives.

We will use functions provided by the MTH229 package:

using MTH229

Derivatives

We discussed three ways to find derivatives in Julia:

We can wrap this in a function, but perhaps it is best to just compute it. For example, for the function $f(x) = x - \log(x)$, we have the derivative at $2$ is approximately:

f(x) = x - log(x)
c = 2
h = 1e-6
(f(c+h) - f(c)) / h  # basically 1/2, but an approximation
0.5000001253030462

Different values of h give different approximations and there is an art to selecting the right size of h, but for most purposes, 1e-6 will be reasonable.

f(x) = x - log(x)
c = 2
f'(c)
0.5

Second derivatives are found by using 2 "primes":

f''(c)
0.25

We can use f' just like any other function gets used. For example, to find out when $f'(x) = 0$ in the interval $[1/10, 10]$ we have:

fzeros(f', 1/10, 10)
1-element Array{Float64,1}:
 1.0

At $c=2$, we can see the symbolic derivative is also $1/2$ with diff(f)(2).


The project introduced the tangent function to describe the tangent line to $f$ and $c$. We can plot this as follows:

f(x) = x^5 - x - 1
plot(f, 0, 2)
c = 2
plot!(tangent(f, c), 0, 2)

We can ask questions about the tangent line. For example:

When will the tangent line cross the $x$ axis? Clearly it is between 0 and 2:

fzeros(tangent(f, c), 0, 2)
1-element Array{Real,1}:
 1.63291

Or where will the tangent line cross the $y$ axis? This is the value at $0$:

tl = tangent(f, c)
tl(0)
-129.0

First and second derivatives

This rough table describes some relationships we explored in one of the projects:

f$+$$\nearrow$$\smile$
f'.$+$$\nearrow$
f''..$+$

It helps us understand two "tests":

The first derivative test states that for a differentiable function $f(x)$ with a critical point at $c$ then if $f′(x)$ changes sign from + to − at c then $f(c)$ is a local maximum and if it changes sign from − to + then $f(c)$ is a local minimum

The second derivative test states that if $c$ is a critical point of $f(x)$ and $f″(c)>0$ then $f(c)$ is a local minimum and if $f″(c)<0$ then $f(c)$ is a local maximum.

Where a critical point is a value in the domain of $f(x)$ where the derivative is $0$ or undefined. In contrast, an inflection point is a point where $f''(x)$ is $0$ or undefined and the second derivative changes sign at this point.

To find these we used fzero and fzeros. The function fzero(f, a, b) will find a zero of $f(x)$ within a bracketing interval $[a,b]$. The function fzeros(f, a, b) will try to find the zeros of $f$ between $[a,b]$. (It may miss zeros that do not cross the $x$ axis!)

For example, if $f(x) = \cos(x) - \log(x)$ on $(0,10)$ we have the critical points found with:

f(x) = cos(x) - log(x)
fzeros(f', 0, 10)
3-element Array{Float64,1}:
 3.43683
 6.11902
 9.5299 

(The function is differentiable on this interval.)

The inflection points are given by

fzeros(f'', 0, 10)
2-element Array{Float64,1}:
 4.7566
 7.8377

(which requires verifying the second derivative changes sign at these two points.)

This plotif function was helpful in graphical explorations:

For example, using the second derivative for f shows that indeed these two points above are inflection points

plotif(f, f'', 0, 10)

The hardest problems here deal with the case where a function is described in terms of its derivative (or second derivative). For example, suppose $f'(x) = x^3 - x + 1$. What can be said about where $f$ is increasing? Concave up?

Well, $f$ is increasing when $f'(x) \geq 0$ (basically), so we identify where $f$ is 0.

fp(x) = x^3 - x + 1
cps = fzeros(fp, -5, 5)
1-element Array{Real,1}:
 -1.32472

We can graph or check otherwise that to the left of the lone zero $fp < 0$ and to the right $fp > 0$. So $f(x)$ will increase on $( -1.32472, \infty)$.

Now, $f$ will be concave up when $f''(x) \geq 0$. We can investigate this with:

fzeros(fp', -5, 5)
2-element Array{Float64,1}:
 -0.57735
  0.57735

We see two possible inflection points. A graph will show that $f''(x) < 0$ in $(-0.57735, 0.57735)$ and is otherwise positive. So $f(x)$ is concave up except on the interval $(-0.57735, 0.57735)$.