Analytical Applications of Differentiation

Study Sheet

Analytical Applications of Differentiation

Using ff' and ff'' to analyze and sketch functions

The Mean Value Theorem & Rolle's Theorem

Concept
Mean Value Theorem (MVT)

If ff is continuous on the closed interval [a,b][a,b] and differentiable on the open interval (a,b)(a,b), then there exists at least one cc in (a,b)(a,b) with

f(c)=f(b)f(a)ba.f'(c)=\frac{f(b)-f(a)}{b-a}.

That is, the instantaneous rate at some interior point equals the average rate of change over [a,b][a,b]: some tangent line is parallel to the secant line.

Concept
Rolle's Theorem (special case)

If ff is continuous on [a,b][a,b], differentiable on (a,b)(a,b), and f(a)=f(b)f(a)=f(b), then there is a cc in (a,b)(a,b) with f(c)=0f'(c)=0.

Reminder — The Mean Value Theorem:f(c)=f(b)f(a)bafor some c(a,b)f'(c)=\frac{f(b)-f(a)}{b-a} \quad\text{for some } c \in (a,b)
Example
Applying the MVT

Let f(x)=x2f(x)=x^2 on [1,3][1,3]. It is a polynomial, so continuous and differentiable everywhere. The average rate is

f(3)f(1)31=912=4.\frac{f(3)-f(1)}{3-1}=\frac{9-1}{2}=4.

Set f(c)=2c=4f'(c)=2c=4, so c=2c=2, which lies in (1,3)(1,3). ✓

Example
Applying Rolle's Theorem

Let f(x)=x22xf(x)=x^2-2x on [0,2][0,2]. Then f(0)=0=f(2)f(0)=0=f(2), and ff is a polynomial, so Rolle applies. Solving f(c)=2c2=0f'(c)=2c-2=0 gives c=1c=1 in (0,2)(0,2). ✓

Tip

Tip: Always verify the hypotheses first. For f(x)=xf(x)=|x| on [1,1][-1,1] we have f(1)=f(1)f(-1)=f(1), but ff is not differentiable at 00, so Rolle's Theorem does not apply.

Increasing/Decreasing & the First Derivative Test

Concept
Monotonicity and local extrema

On an interval where f(x)>0f'(x)>0, ff is increasing; where f(x)<0f'(x)<0, ff is decreasing. First Derivative Test: at a critical number cc (where f(c)=0f'(c)=0 or ff' is undefined):

  • [leftmargin=5mm,itemsep=1pt]
  • ff' changes ++\to- at cc \Rightarrow local maximum;
  • ff' changes +-\to+ at cc \Rightarrow local minimum;
  • no sign change \Rightarrow neither.
Example
Sign chart for f(x)=x33xf(x)=x^3-3x

f(x)=3x23=3(x1)(x+1)f'(x)=3x^2-3=3(x-1)(x+1), so critical numbers are x=1,1x=-1,\,1.

Local max at x=1x=-1 with f(1)=2f(-1)=2; local min at x=1x=1 with f(1)=2f(1)=-2.

Tip

Tip: Critical numbers only name candidates. A sign change of ff' is what confirms an extremum. Test one point in each interval.

Concavity, the Second Derivative Test & Inflection Points

Concept
Concavity and inflection

Where f(x)>0f''(x)>0, ff is concave up (\smile); where f(x)<0f''(x)<0, ff is concave down (\frown). An inflection point is where concavity changes---so ff'' must change sign there. Second Derivative Test: if f(c)=0f'(c)=0, then f(c)>0f''(c)>0\Rightarrow local min, f(c)<0f''(c)<0\Rightarrow local max, and f(c)=0f''(c)=0 is inconclusive.

Example
Concavity of f(x)=x33xf(x)=x^3-3x

f(x)=6xf''(x)=6x. Then f<0f''<0 for x<0x<0 (concave down) and f>0f''>0 for x>0x>0 (concave up); the concavity changes at x=0x=0, so (0,0)(0,0) is an inflection point. Using the Second Derivative Test at the critical numbers: f(1)=6<0f''(-1)=-6<0 (local max) and f(1)=6>0f''(1)=6>0 (local min), matching the First Derivative Test.

Example
f=0f''=0 does not guarantee inflection

For f(x)=x4f(x)=x^4, f(x)=12x2f''(x)=12x^2, which is 00 at x=0x=0 but never changes sign (f0f''\ge 0). So x=0x=0 is not an inflection point---ff is concave up throughout.

Tip

Tip: An inflection point requires an actual sign change of ff'', not merely f=0f''=0.

The Extreme Value Theorem & Absolute Extrema (Candidates' Test)

Concept
Extreme Value Theorem (EVT)

If ff is continuous on a closed interval [a,b][a,b], then ff attains an absolute maximum and an absolute minimum on [a,b][a,b]. Candidates' Test: the absolute extrema occur at (i) critical numbers in (a,b)(a,b) or (ii) the endpoints. Evaluate ff at every candidate and compare.

Example
Candidates' Test for f(x)=x24x+3f(x)=x^2-4x+3 on [0,3][0,3]

f(x)=2x4=0x=2f'(x)=2x-4=0\Rightarrow x=2 (in the interval). Evaluate all candidates:

f(0)=3,f(2)=1,f(3)=0.f(0)=3,\qquad f(2)=-1,\qquad f(3)=0.

Absolute maximum 33 at x=0x=0; absolute minimum 1-1 at x=2x=2.

Tip

Tip: On a closed interval you compare numbers, not signs. Never forget the endpoints---they are candidates too.

Optimization (Applied Max/Min)

Concept
A reliable setup
  • [leftmargin=6mm,itemsep=1pt]
  • Draw a picture; name variables.
  • Write the quantity to optimize (the objective).
  • Use a constraint to reduce to one variable; state the domain.
  • Differentiate, find critical numbers, and justify the extremum (1st/2nd derivative test or endpoints).
Example
Largest open-top box from a 12×1212\times 12 sheet

Cut squares of side xx from each corner and fold. The volume is

V(x)=x(122x)2,0<x<6.V(x)=x(12-2x)^2,\qquad 0<x<6.

Then V(x)=(122x)2+x2(122x)(2)=(122x)(126x)V'(x)=(12-2x)^2+x\cdot 2(12-2x)(-2)=(12-2x)(12-6x). Setting V=0V'=0: x=6x=6 (gives V=0V=0) or x=2x=2. Since VV' changes ++\to- at x=2x=2, it is a maximum:

V(2)=2(124)2=264=128 cubic units.V(2)=2(12-4)^2=2\cdot 64=128 \text{ cubic units.}
Tip

Tip: Always state the domain from the physical situation, then justify that your critical number gives the max or min---do not just assert it.

Curve Sketching & Connecting ff, ff', ff''

Example
Sketch of f(x)=x33xf(x)=x^3-3x

Combining the work above: increasing on (,1)(-\infty,-1) and (1,)(1,\infty), decreasing on (1,1)(-1,1); local max (1,2)(-1,2), local min (1,2)(1,-2); concave down for x<0x<0, up for x>0x>0; inflection (0,0)(0,0).

Example
The graphs of ff, ff', and ff'' together

For f(x)=13x3xf(x)=\tfrac13x^3-x we get f(x)=x21f'(x)=x^2-1 and f(x)=2xf''(x)=2x. Notice: ff has a local max where ff' crosses zero going ++\to- (x=1x=-1); ff has an inflection where ff'' crosses zero and ff' bottoms out (x=0x=0).

Tip

Tip: When reading a graph of ff': the height of ff' tells the slope of ff; the zeros of ff' mark candidate extrema of ff; where ff' is increasing, ff is concave up.

Going Deeper: Advanced Analysis

Concept
1. A proof-flavored MVT application

Claim: If f(x)=0f'(x)=0 for every xx in an interval II, then ff is constant on II. Why: Pick any two points a<ba<b in II. By the MVT there is a cc with

f(b)f(a)ba=f(c)=0,\frac{f(b)-f(a)}{b-a}=f'(c)=0,

so f(b)=f(a)f(b)=f(a). Since a,ba,b were arbitrary, ff takes one value throughout II. This is the theorem behind “antiderivatives differ by a constant.”

Example
2. MVT as an inequality machine

Show sinasinbab|\sin a-\sin b|\le|a-b| for all a,ba,b. Apply the MVT to g(x)=sinxg(x)=\sin x on [a,b][a,b]: there is a cc with

sinbsinaba=cosc.\frac{\sin b-\sin a}{b-a}=\cos c.

Taking absolute values, sinbsinaba=cosc1\dfrac{|\sin b-\sin a|}{|b-a|}=|\cos c|\le 1, so sinasinbab|\sin a-\sin b|\le|a-b|.

Example
3. Reading ff from a graph of ff'

Suppose the graph shown is ff' (not ff), with f(x)=x21f'(x)=x^2-1.

Then ff is increasing where f>0f'>0: on (,1)(-\infty,-1) and (1,)(1,\infty); decreasing on (1,1)(-1,1). So ff has a local max at x=1x=-1 and a local min at x=1x=1. Since ff' is decreasing for x<0x<0 and increasing for x>0x>0, ff is concave down then concave up, with an inflection point at x=0x=0.

Example
4. Inflection from a sign change of ff''

For f(x)=x44x3f(x)=x^4-4x^3, f(x)=12x224x=12x(x2)f''(x)=12x^2-24x=12x(x-2). Sign of ff'': positive for x<0x<0, negative on (0,2)(0,2), positive for x>2x>2. Two sign changes give two inflection points, at x=0x=0 and x=2x=2. Contrast with f(x)=x4f(x)=x^4, where f=12x20f''=12x^2\ge0 never changes sign---no inflection despite f(0)=0f''(0)=0.

Concept
5. Why the Candidates' Test works

The EVT guarantees a continuous ff attains its max and min on [a,b][a,b]. Wherever such an extremum occurs at an interior point where ff is differentiable, Fermat's Theorem forces f=0f'=0 there. So an absolute extremum can only hide at (i) an interior point with f=0f'=0 or ff' undefined, or (ii) an endpoint. Checking that finite list of candidates is therefore guaranteed to find the answer.

Example
6. Harder optimization with a constraint

Minimize the surface area of a closed cylinder of fixed volume VV. Constraint: πr2h=Vh=Vπr2\pi r^2 h=V\Rightarrow h=\dfrac{V}{\pi r^2}. Objective:

S=2πr2+2πrh=2πr2+2Vr.S=2\pi r^2+2\pi r h=2\pi r^2+\frac{2V}{r}.

Then S(r)=4πr2Vr2=0r3=V2πr=V2π3S'(r)=4\pi r-\dfrac{2V}{r^2}=0\Rightarrow r^3=\dfrac{V}{2\pi}\Rightarrow r=\sqrt[3]{\dfrac{V}{2\pi}}. Since S=4π+4Vr3>0S''=4\pi+\dfrac{4V}{r^3}>0, this is a minimum. Substituting back gives h=2rh=2r: the most economical can is as tall as it is wide.

Example
7. Optimization by distance-squared

Closest point on y=1xy=\tfrac1x (x>0x>0) to the origin. Minimize D2=x2+1x2D^2=x^2+\dfrac{1}{x^2} (minimizing D2D^2 minimizes DD). Then d(D2)dx=2x2x3=0x4=1x=1\dfrac{d(D^2)}{dx}=2x-\dfrac{2}{x^3}=0\Rightarrow x^4=1\Rightarrow x=1. The point is (1,1)(1,1) with distance 2\sqrt2.

Tip

Big picture: ff' controls slope (rising/falling and where it turns); ff'' controls bending (concavity and inflection). Every result in this topic---MVT, the derivative tests, EVT, optimization, curve sketching---is a way of turning information about ff' and ff'' into conclusions about ff.

Formulas, Proofs & Tips

Tip
Critical points and concavity
f(c)=0 or undefinedc critical;f>0concave up, f<0concave downf'(c)=0 \text{ or undefined} \Rightarrow c \text{ critical};\qquad f''>0 \Rightarrow \text{concave up},\ f''<0 \Rightarrow \text{concave down}

What it means. Maxima and minima hide where the slope is zero; the second derivative says which one it is.

Example. f(x)=x33xf(x)=x^3-3x: f(x)=3x23=0f'(x)=3x^2-3=0 at x=±1x=\pm1, the critical points.

Why it works. At a smooth peak or valley the tangent is horizontal, so f=0f'=0. If f>0f''>0 the slope is increasing, so the curve bends upward and a horizontal tangent must be a minimum; f<0f''<0 reverses it.

Tip. f=0f'=0 alone is not enough — y=x3y=x^3 has f(0)=0f'(0)=0 but no extremum. Check a sign change of ff', or use ff''.

Tip
The Mean Value Theorem
f(c)=f(b)f(a)bafor some c(a,b)f'(c)=\frac{f(b)-f(a)}{b-a} \quad\text{for some } c \in (a,b)

What it means. Somewhere in the interval the instantaneous rate equals the average rate.

Example. For f(x)=x2f(x)=x^2 on [0,2][0,2]: f(c)=4020=2f'(c)=\tfrac{4-0}{2-0}=2, which occurs at c=1c=1.

Why it works. Tilt the graph so the endpoints are level (subtract the secant line). The resulting function has equal endpoint values, so it has an interior max or min, where the derivative is 00 — untilting gives f(c)f'(c) equal to the secant slope.

Tip. Needs ff continuous on [a,b][a,b] and differentiable on (a,b)(a,b). Rolle's Theorem is the special case f(a)=f(b)f(a)=f(b).