5.5 Monotonicity and Extremum
1. Increasing - Decreasing Functions
Consider the following graph of $y = f(x)$:
Let us make some observations from the graph:
- The domain of the graph is the interval $[a, b]$. The points $x=a$ and $x=b$ are called endpoints.
- We state that we have a local max (or just max) at the points: $x=c_1, x=c_3, x=b$.
- We state that we have a local min (or just min) at the points: $x=a, x=c_2, x=c_4$.
Notice that $x=c_5$ is not a turning point (neither max nor min), as near $f(c_5)$ you can find smaller as well as larger values.
Tracking the flow of the curve:
- The function is increasing (goes up) in the interval $(a, c_1)$
- The function is decreasing (goes down) in the interval $(c_1, c_2)$
- The function is increasing (goes up) in the interval $(c_2, c_3)$
- The function is decreasing (goes down) in the interval $(c_3, c_4)$
- The function is increasing (goes up) in the interval $(c_4, b)$
Remember that a positive gradient means that the function is increasing (goes up), and a negative gradient implies the function is decreasing (goes down). But we know that derivative = gradient.
If $f'(x) > 0$ then $f$ is increasing ($\nearrow$)
If $f'(x) < 0$ then $f$ is decreasing ($\searrow$)
Notice: The increasing or decreasing behavior of a function is known as its monotonicity!
2. Turning Points: MAX - MIN
How can we isolate the turning points (max or min) of a function?
First, the endpoints are extreme values (see $a$ and $b$ above). As far as the interior points are concerned, observe that the gradient at any turning point is 0 (the tangent lines at those values are horizontal!).
PROPOSITION
If $f(x)$ has a turning point (max or min) at some interior point $c$ and $f'(c)$ exists, then:
Notice that $f'(x) = 0$ at $c_1, c_2, c_3, c_4$, and $c_5$. We verify a local max at $c_1, c_3$ and a local min at $c_2, c_4$. However, $c_5$ is not a turning point. Hence, the inverse proposition is not true.
Therefore, apart from the endpoints, the possible turning points (max/min) are restricted to the following:
- points $x$ where $f'(x) = 0$ (called stationary points)
- points where $f'(x)$ does not exist
All these boundaries are classified as critical points. Here we deal only with stationary points (where $f'(x) = 0$). To verify whether a stationary point $x=c$ acts as a turning point (max or min), we execute the following test.
FIRST DERIVATIVE TEST for c
Check the sign of $f'(x)$ to verify if the curve is increasing or decreasing just before and after $c$:
| $x$ | $c$ | ||
|---|---|---|---|
| $f'(x)$ | + | 0 | $-$ |
| Conclusion for $f$ | $\nearrow$ | max | $\searrow$ |
| $x$ | $c$ | ||
|---|---|---|---|
| $f'(x)$ | $-$ | 0 | + |
| Conclusion for $f$ | $\searrow$ | min | $\nearrow$ |
If the sign does not alter, we possess neither a max nor a min.
METHODOLOGY
Given $y = f(x)$:
- Step 1: we find $f'(x)$
- Step 2: we solve $f'(x) = 0$ (say the resulting roots are $a, b, c$)
- Step 3: we construct a table as follows to perform the first derivative test:
| $x$ | $a$ | $b$ | $c$ | ||||
|---|---|---|---|---|---|---|---|
| $f'(x)$ | + | 0 | $-$ | 0 | + | 0 | + |
| Conclusion for $f$ | $\nearrow$ | max | $\searrow$ | min | $\nearrow$ | nothing | $\nearrow$ |
EXAMPLE 1
Consider $f(x) = \dfrac{1}{3}x^3 - 2x^2 + 3x + 5$
We solve $x^2 - 4x + 3 = 0$
The solutions are $x=1$ and $x=3$
We construct the table:
| $x$ | 1 | 3 | |||
|---|---|---|---|---|---|
| $f'(x) = x^2 - 4x + 3$ | + | 0 | $-$ | 0 | + |
| Conclusion for $f$ | $\nearrow$ | max | $\searrow$ | min | $\nearrow$ |
we have a max at $x=1$ [and the max value of $f$ is $f(1) = 6.33$]
we have a min at $x=3$ [and the min value of $f$ is $f(3) = 5$]
An alternative mechanism to verify if a stationary point correlates to a max or a min is the following:
SECOND DERIVATIVE TEST for c
Find $f''(x)$ (if it exists!)
- If $f''(c) > 0$ then $c$ establishes a min
- If $f''(c) < 0$ then $c$ establishes a max
- If $f''(c) = 0$ we do not attain a conclusive answer. We revert back to the first derivative test.
EXAMPLE 2
Consider again $f(x) = \dfrac{1}{3}x^3 - 2x^2 + 3x + 5$
We differentiate to find $f''(x) = 2x - 4$.
For $x=1$, $f''(1) = -2 < 0$, so we confirm a max at $x=1$.
For $x=3$, $f''(3) = 2 > 0$, so we confirm a min at $x=3$.
EXAMPLE 3
Consider $f(x) = (x-1)^4$
There is one stationary point at $x=1$.
Evaluating $f''(x) = 12(x-1)^2$.
For $x=1$, $f''(1) = 0$ (neither positive nor negative). Thus, we cannot conclude if it represents a max or a min.
The table of signs gives:
| $x$ | 1 | ||
|---|---|---|---|
| $f'(x) = 4(x-1)^3$ | $-$ | 0 | + |
| Conclusion for $f(x)$ | $\searrow$ | min | $\nearrow$ |