4.1 Maximum and Minimum Values goo.

gl/8FNtRj 1

Chapter 4 Applications of Differentiation

4.1 Maximum and Minimum Values, page 276
Definition 1 (page 276). Let c be a number in the domain D of a function f . Then f (c) is
the
oBLCfVT2JF0

(1) absolute maximum value (絕對極大值) of f on D if f (c) ≥ f (x) for all x in D.

(2) absolute minimum value (絕對極小值) of f on D if f (c) ≤ f (x) for all x in D.

 Absolute maximum (or minimum) 有時候也稱為 global maximum (or minimum).

 所有的絕對極大值、 絕對極小值統稱為函數 f 的極值 (extreme values).
y y

x x

Figure 1: Absolute maximum value and absolute minimum value of f .

 判斷絕對極值時, 必須定義域內 「所有」 (for all) 的點都要做比較。

Definition 2 (page 276). The number f (c) is a

(1) local maximum value (局部極大值) of f on D if f (c) ≥ f (x) when x is near c.

(2) local minimum value (局部極小值) of f on D if f (c) ≤ f (x) when x is near c.

We say that something is true near c, we mean that it is true on “some open interval
containing c.”
y y

x x

Figure 2: Local maximum value and local minimum value of f .

 局部極值的定義中, 「附近」 (near) 這個詞很重要。

 2 4.1 Maximum and Minimum Values goo.gl/8FNtRj

Example 3. State the absolute (and local) maximum (and minimum) values of the function
y = f (x).
efbompPs1hs
y
y = f (x)

1
x
1

Figure 3: Find absolute (and local) maximum (and minimum) values of the function.

Solution.

(a) Absolute maximum:

(b) Local maximum:

(c) Absolute minimum:

(d) Local minimum:

Theorem 4 (The Extreme Value Theorem, 極值定理, page 278). If f is continuous on a closed
interval [a, b], then f attains an absolute maximum value f (c) and an absolute minimum value
f (d) at some numbers c and d in [a, b].

 極值定理的條件是 「閉區間」 上的 「連續函數」。

 極值定理的結論只告知 「存在性」。
Example 5. Give examples that if f is not continuous, or f is continuous on (a, b), the
Extreme Value Theorem does not hold. Plot a continuous function that it attains maximum
values and minimum values at more than one number.

y y y

x
x x

Figure 4: Study the Extreme Value Theorem.

4.1 Maximum and Minimum Values goo.gl/8FNtRj 3

Theorem 6 (Fermat’s Theorem 費馬定理, page 279). If f has a local maximum or minimum
at c, and if f ′ (c) exists, then f ′ (c) = 0.
X3SRnYvweSQ
 條件 “f ′ (c) exists” 很重要。 反例: 。
 一般而言, 費馬定理的反敘述不對, 例如: 。

Proof. Here we prove the local maximum case. Since f (c) ≥ f (x) if x is sufficiently close
to c, this implies that if h is sufficiently close to 0, with h being positive or negative, then
f (c) ≥ f (c + h), or equivalently, f (c + h) − f (c) ≤ 0.
If h > 0, we have f (c+h)−f
h
(c)
≤ 0. Since f ′ (c) exists, we get

f (c + h) − f (c)
f ′ (c) = lim =
h→0 h
f (c+h)−f (c)
If h < 0, we have h ≥ 0. Since f ′ (c) exists, we get

f (c + h) − f (c)
f ′ (c) = lim =
h→0 h
Hence f ′ (c) = 0.

Definition 7 (page 280). A critical number (臨界點) of a function f is a number c in the

domain of f such that f ′ (c) = 0 or f ′ (c) does not exist.
TIa3VmnB5ek
 臨界點 (critical numbers) 是在函數 f 的 「定義域」 內。
3 3 8
Example 8 (page 280). Find the critical numbers of f (x) = x 5 (4 − x) = 4x 5 − x 5 .

Solution. We compute

f ′ (x) =

Therefore the critical numbers are .

Theorem 9 (page 280). If f has a local maximum or minimum at c, then c is a critical

number of f .

 此定理等價於“若函數 f 不存在臨界點, 則 f 沒有局部極大值, 也沒有局部極小值。”

The Closed Interval Method (page 280). To find the absolute maximum and minimum
values of a piecewise continuous function on a closed interval [a, b]:

(1) Find the values of f at the critical numbers of f in (a, b).

(2) Find the values of f at the endpoints of the interval, that is, f (a) and f (b).

(3) The largest and smallest of the values from (1) and (2) are absolute maximum value
and absolute minimum value, respectively.

 求絕對極值的方法: 找出所有臨界點與端點; 比較那些點的函數值之大小。

 4 4.2 The Mean Value Theorem goo.gl/ZUU6e5

4.2 The Mean Value Theorem, page 287

Question 1. A highway from Taipei to Kaohsiung is 330 km and the speed limit is 110 km/h.
Man A drove the car on the high way from Taipei at 9 : 00 AM to Kaohsiung at 11 : 59 AM.
3R2E7ilyfck
Did he exceed the speed limit?

Theorem 2 (Rolle’s Theorem, page 287). Let f be a function that satisfies the following three
hypotheses:

(1) f is continuous on the closed interval [a, b].

(2) f is differentiable on the open interval (a, b).

(3) f (a) = f (b).

Then there is a number c in (a, b) such that f ′ (c) = 0.

Figure 1: Rolle’s Theorem.

Proof. There are three cases.

(I) f (x) = k, a constant. We have f ′ (x) = 0, so the number c can be taken to be any
number in (a, b).

(II) f (x) > f (a) for some x in (a, b). By the , f has a maximum
somewhere in [a, b]. Since f (a) = f (b), it must attain this maximum value at a number
c in the open interval (a, b). Then f has a at c, and f is differentiable
at c. By ′
, we know f (c) = 0.

(III) f (x) < f (a) for some x in (a, b). By the , f has a minimum
value in [a, b], and since f (a) = f (b), it attains this local minimum value at a number
c ∈ (a, b). By , f ′ (c) = 0.

 定理條件, 函數 f (x) 必須在 「閉區間連續」。

 定理條件, 函數 f (x) 必須在開區間 「可微分」 (每一個點)。
 定理結論只告知 「存在性」。
4.2 The Mean Value Theorem goo.gl/ZUU6e5 5

Example 3. Give examples that each condition in Rolle’s Theorem is required.

Solution. x6lRgE7n8o4

y y y

x x x

Figure 2: Study Rolle’s Theorem.

Example 4 (page 287). Prove that x3 + x − 1 = 0 has exactly one real root.

Solution.

Theorem 5 (The Mean Value Theorem, 平均值定理, page 288). Let f be a function that
satisfies the following hypotheses:
ZtR_UIq3vxQ

(1) f is continuous on the closed interval [a, b].

(2) f is differentiable on the open interval (a, b).

Then there is a number c in (a, b) such that

f (b) − f (a)
f ′ (c) = or equivalently, f (b) − f (a) = f ′ (c)(b − a).
b−a

Figure 3: The Mean Value Theorem.

 6 4.2 The Mean Value Theorem goo.gl/ZUU6e5

Proof of Mean Value Theorem. Define a new function

h(x) =

We will verify that h(x) satisfies the three hypotheses of Rolle’s Theorem.

(1) The function h is continuous on [a, b]: It is the sum of f and a first-degree polynomial,
both of which are continuous.

(2) The function h is differentiable on (a, b): Both f and the first-degree polynomial are
differentiable. In fact, we have

h′ (x) =

(3) h(a) = h(b) = 0:

h(a) =

h(b) =

By , there is a number c ∈ (a, b) such that h′ (c) = 0. Therefore,

 均值定理也是要注意 「連續」、「可微」、「存在」。
 為什麼這個定理要叫做 「平均值定理」?
Theorem 6. If f ′ (x) = 0 for all x in an interval (a, b), then f is constant on (a, b).

58cSDtk3T9M Proof. Let x1 and x2 be any two numbers in (a, b) with x1 < x2 . Since f is differentiable
on (a, b), it must be differentiable on (x1 , x2 ) and continuous on [x1 , x2 ]. By applying the
to f on the interval [x1 , x2 ], we get a number c such that x1 < c < x2
and

Therefore f has the same value at any two numbers x1 and x2 in (a, b). So f (x) is constant
on (a, b).

 Theorem 6 提供一個刻劃常數函數的方法。
4.2 The Mean Value Theorem goo.gl/ZUU6e5 7

Corollary 7. If f ′ (x) = g ′ (x) for all x in an interval (a, b), then f − g is constant on (a, b);
that is, f (x) = g(x) + c where c is a constant.

Proof. Let F (x)

x
 注意 f (x) = |x| 與 g(x) = 1, 它在 x ∈ (−1, 1) 當中並不滿足推論當中的條件。
|x−y|
Example 8. Show that tan x2 − tan y2 ≥ 2 for any x, y ∈ (−π, π).

Solution. If x = y, the inequality holds. If x 6= y, without loss of generality, we assume WjWwBQAaW3Q

−π < x < y < π. Consider the function f (t) = tan 2t , then

• f (t) is .

• f (t) is .

By the , there is a number c ∈ (x, y) such that f (x) − f (y) = f ′ (c)(x −

y), which implies |f (x) − f (y)| = |f ′ (c)||x − y|. Since f ′ (t) = , we have |f ′ (c)| =
. So |f (x) − f (y)| ≥ 12 |x − y|, which means

x y |x − y|
tan − tan ≥ .
2 2 2
Theorem 9 (Cauchy’s Mean Value Theorem, (柯西均值定理) Appendix F, A45). Suppose
that the functions f and g are continuous on [a, b] and differentiable on (a, b), and g′ (x) 6= 0
for all x in (a, b). Then there is a number c ∈ (a, b) such that biygd5TQf38

f ′ (c) f (b) − f (a)

′
= .
g (c) g(b) − g(a)

Proof. The key point is to find a new function F (x) and apply the Mean Value Theorem.

F (x) =
 8 4.4 Indeterminate Forms and l’Hospital’s Rule goo.gl/QwjMrw

4.4 Indeterminate Forms and l’Hospital’s Rule, page

304
In this section, we want to introduce a new method to deal with the limit such as

ln x x2
PwmK2Ha2SAA
lim , or lim .
x→∞ x − 1 x→∞ ex

Definition 1 (page 304–305).

f (x)
(1) If we have a limit of the form lim , where both f (x) → 0 and g(x) → 0 as x → a,
x→a g(x)
0
it is called an indeterminate form of type 0 (零分之零的不定型).
f (x)
(2) If we have a limit of the form lim , where both f (x) → ∞ (or −∞) and g(x) → ∞
x→a g(x)
∞
(or −∞) as x → a, it is called an indeterminate form of type ∞ (無限大分之無限大的不
定型).

L’ Hospital’s Rule (page 305). Suppose f and g are differentiable and g′ (x) 6= 0 on an open
interval I that contains a (except possibly at a). Suppose that a limit has an indeterminate
form of type 00 or ∞
∞ . Then

f (x) f ′ (x)
lim = lim ′
x→a g(x) x→a g (x)

if the limit on the right side exists (or is ∞ or −∞).

(x)
lim fg (x)
′
 務必檢查定理的條件: (1) 是否為不定型; (2) 檢查 x→a 是否存在。 ′

 定理也適用於單邊極限。
(k)
(x) L (x) L L (x)
lim fg(x) = lim fg (x) = · · · = lim fg (x)
′
 可以串聯至有限次可微分函數。 x→a x→a x→a
= M.
′ (k)

Example 2.
t−ln(1+t)
aBKa8sOdzCM
(a) Find lim+ t2 .
t→0
√
t−ln(1+t)
(b) Use (a) to find lim+ t .
t→0

Solution.
4.4 Indeterminate Forms and l’Hospital’s Rule goo.gl/QwjMrw 9

Indeterminate Products, page 308

Definition 3 (page 305). If we have a limit of the form lim f (x)g(x), where lim f (x) = 0
x→a x→a
and lim g(x) = ∞ (or −∞) as x → a, it is called an indeterminate form of type 0 · ∞. (零乘
x→a VdDFNTkTC1k
以無限大的不定型)

We can deal with it by writing the product f g as a quotient:

f g
fg = or fg = ,
1/g 1/f
0 ∞
and this converts the given limit into an indeterminate form of type 0 or ∞.

Example 4 (page 308). Evaluate lim+ x ln x.

x→0

Solution.

 如何把函數分配至分子或分母是一門學問。

Indeterminate Differences, page 309

Definition 5 (page 305). If lim f (x) = ∞ and lim g(x) = ∞, then the limit
x→a x→a

lim (f (x) − g(x)) wbQEYnzfK1Q

x→a

is called an indeterminate form of type ∞ − ∞ (無限大減無限大的不定型).

We can try to convert the difference into a quotient (for instance, by using a common
denominator (通分), or rationalization (有理化), or factoring out a common factor (提公因式))
so that we have an indeterminate form of type 00 or ∞
∞.

Example 6. Find the limit lim x12 − sin12 x .

x→0

Solution.

 隨時觀察並恰當整理函數, 不要盲目使用 l’Hospital Rule。

 10 4.4 Indeterminate Forms and l’Hospital’s Rule goo.gl/QwjMrw

Indeterminate Powers, page 310

Definition 7 (page 310). Several indeterminate forms arise from the limit

h2ifbDam0Xw lim (f (x))g(x)

x→a

(1) lim f (x) = 0 and lim g(x) = 0: type 00 .

x→a x→a

(2) lim f (x) = ∞ and lim g(x) = 0: type ∞0 .

x→a x→a

(3) lim f (x) = 1 and lim g(x) = ±∞: type 1∞ .

x→a x→a

Each of these three cases can be treated either by taking the natural logarithm: let y =
(f (x))g(x) , then ln y = g(x) ln f (x) or by writing the function as an exponential: (f (x))g(x) =
eg(x) ln f (x) . In either method we are led to the indeterminate product g(x) ln f (x), which is of
type 0 · ∞.

Example 8 (page 310). Find lim+ xx .

x→0

Solution.

 取對數算出極限值後, 記得還原。
4.4 Indeterminate Forms and l’Hospital’s Rule goo.gl/QwjMrw 11

幾個使用 l’Hospital Rule 的經驗

f ′ (x)
• 使用前務必檢查定理的條件: (1) 是否為不定型; (2) 檢查 lim 是否存在。
x→a g (x)
′

x
lim o_0cpMZYcps
x→0 1 + sin x

x − sin x
lim
x→∞ x + sin x

• 避免鬼打牆; 像是 sin x, cos x (as x → ∞) 或是 sin x1 , cos x1 , x1 , ln1x (as x → 0)。

x2 sin x1 ( 00 ),L 2x sin x1 + x2 cos x1 (− x12 ) 2x sin x1 − cos x1

lim = lim = lim . . .?
x→0 sin x x→0 cos x x→0 cos x
x ( 00 ),L′ 1
lim+ x ln x = lim+ 1 = lim = lim+ −x(ln x)2 . . . 鬼打牆
x→0 x→0
ln x
x→0+ − (ln1x)2 · 1
x
x→0

• 分子分母適時地整理、 重新分配, 或是變數變換, 有助於計算。

1 1 1
1 e− x ( 00 ),L′ e− x · x12 e− x
lim 1 = lim+ 2 = lim+ = lim+ 3 . . . 鬼打牆
x→0+ x2 e x x→0 x x→0 2x x→0 2x

1 ( ∞ ),L
1 x2 ∞ −2 x13 2
lim+ 1 = lim+ 1 = lim+ 1
= lim+ 1 . . . 情況變好, 再用一次
x→0 x2 e x x→0 e x x→0 e ·
x − x12 x→0 xe x
1
lim+ 1
x→0 x2 e x

tan x − x
lim
x→0 x − sin x
tan 3x
lim+ √
x→0 1 − cos 2x
• l’Hospital Rule 某種程度是“大絕”, 但並非萬能。
(sin x)ex
lim
x→∞ (x + sin x)e2x

• 記得其他求極限的方法, 像是 Squeeze Theorem, definition of derivative, 也很好用。(之後還

會介紹用積分方法求極限, 下學期會介紹使用泰勒展式法求極限。)
sin x 1 sin x 1
lim = lim x sin = lim = lim x sin =
x→0 x x→∞ x x→∞ x x→0 x

1 x
 
1 ln(1 + x)
lim 1 + = lim (1 + x) x = lim =
x→∞ x x→0 x→0 x

1 x
 
1 ln(1 + x)
lim 1 + = lim (1 + x) x = lim =
x→0 x x→∞ x→∞ x
 12 4.4 Indeterminate Forms and l’Hospital’s Rule goo.gl/QwjMrw

Appendix
Proof of l’Hospital’s Rule (Appendix A46)
f ′ (x)
We are assuming that lim f (x) = 0 and lim g(x) = 0. Let lim = L. Define
x→a g (x)
EH5tz1psN8c ′
x→a x→a
( (
f (x) if x 6= a g(x) if x 6= a
F (x) = , G(x) = .
0 if x = 1 0 if x = 1

Then both F and G are continuous on I since f and g are continuous on {x ∈ I|x 6= a} and

lim F (x) = lim f (x) = 0 = F (a), lim G(x) = lim g(x) = 0 = G(a).
x→a x→a x→a x→a

Furthermore, F and G are differentiable on (a, x) (or (x, a)) since F ′ = f ′ and G′ = g′ . Since
G′ 6= 0, by the Cauchy’s Mean Value Theorem, there is a number y such that a < y < x (or
x < y < a) and

F ′ (y) F (x) − F (a) F (x)

′
= = .
G (y) G(x) − G(a) G(x)

Hence
f (x) F (x) F ′ (y) f ′ (y)
lim+ = lim+ = lim+ ′ = lim+ ′ = L,
x→a g(x) x→a G(x) y→a G (y) y→a g (y)
F ′ (y) f ′ (y)
 
f (x) F (x)
and lim = lim− = lim− ′ = lim− ′ = L.
x→a− g(x) x→a G(x) y→a G (y) y→a g (y)

Therefore,

f (x)
lim = L.
x→a g(x)

This proves l’Hospital’s Rule for the case where a is finite.

If a is infinite, we let t = x1 . Then t → 0+ as x → ∞, so we have

f ( 1t ) L f ′ ( 1t ) · − t12 f ′ ( 1t )


f (x) f ′ (x)
lim = lim+ 1 = lim+ ′ 1 = lim = lim .
t→0 g ( ) · − 12 t→0+ g ′ ( 1 ) x→∞ g ′ (x)


x→∞ g(x) t→0 g( )
t t t t
4.4 Indeterminate Forms and l’Hospital’s Rule goo.gl/QwjMrw 13

L’Hospital Rule 與極限法則的合併使用

回想極限的四則運算法則與羅必達法則:

Limit Laws (page 99). Suppose that lim p(x) and lim q(x) exist. Then
x→a x→a

p(x) lim p(x)

lim = x→a if lim q(x) 6= 0.
x→a q(x) lim q(x) x→a
x→a

L’ Hospital’s Rule (page 302). Suppose f and g are differentiable and g′ (x) 6= 0 on an open
interval I that contains a (except possibly at a). Suppose that

lim f (x) = 0 and lim g(x) = 0

x→a x→a

or that

lim f (x) = ±∞ and lim g(x) = ±∞

x→a x→a

0 ∞
(In other words, we have an indeterminate form of type 0 or ∞ .) Then

f (x) f ′ (x)
lim = lim ′
x→a g(x) x→a g (x)

if the limit on the right side exists (or is ∞ or −∞).

這兩個定理可以搭配起來靈活運用, 比方說:

Theorem 9. Suppose that p(x), q(x) satisfy the assumptions of Limit Laws with lim p(x) 6= 0,
x→a
and suppose f (x), g(x) satisfy the assumptions of L’Hospital Rule. Then

p(x)f (x) p(x) f ′ (x)

lim = lim · lim ′ .
x→a q(x)g(x) x→a q(x) x→a g (x)

上述定理要強調的是說: 雖然分子 p(x)f (x) 整體看取極限是 0, 分母 q(x)g(x) 整體看取極限是

(x)
0, 於是 p(x)f 0
q(x)g(x) 是 0 的不定型, 但是當分子與分母各自只有一部分是不定型 (只有 g 是 0 ), 而
f 0

p, q 具有非零的極限, 那麼就可以把 p, q 的極限抽出來, 考慮剩下的不定型之極限, 再相乘。

(x) ′
(x)
理由很簡單, 因為 lim fg(x) = lim fg′ (x) = L, 所以下式的第一個等式合法 (乘法法則)
x→a x→a

p(x)f (x) p(x) f (x) p(x) f ′ (x)

lim = lim · lim = lim · lim ′ .
x→a q(x)g(x) x→a q(x) x→a g(x) x→a q(x) x→a g (x)

定理一就是大大化簡計算的一個方法, 因為有很多人一看到 00 就不加思索的地上下微分(典型

的“看見黑影就開槍”), 雖然還是可以算出答案, 但是過程中可能會添上很多計算上的麻煩。
 14 4.3 How Derivatives Affect the Shape of a Graph goo.gl/11C7Ku

4.3 How Derivatives Affect the Shape of a Graph,

page 293
Increasing/Decreasing Test (page 293).

8H-9SUDxCqo
(a) If f ′ (x) > 0 on an interval, then f is increasing on that interval.

(b) If f ′ (x) < 0 on an interval, then f is decreasing on that interval.

Proof.

(a) Let x1 < x2 . By the , there is c ∈ (x1 , x2 ) such that

(b) Let x1 < x2 . By the , there is c ∈ (x1 , x2 ) such that

Example 1. Find where the function f (x) = 3x4 − 4x3 − 12x2 + 5 is increasing and where it
is decreasing.

Solution. We compute f ′ (x) =

Solutions of f ′ (x) = 0 are . Hence
f (x) is increasing on ; f (x) is decreasing on .

The First Derivative Test (page 294). Suppose that c is a critical number of a continuous
function f .
ehahVquKqMk
(a) If f ′ changes from positive to negative at c, then f has a local maximum at c.

(b) If f ′ changes from negative to positive at c, then f has a local minimum at c.

(c) If f ′ does not change sign at c (for example, if f ′ is positive on both side of c or negative
on both sides), then f has no local maximum or minimum at c.

y y y

x x x

Figure 1: The First Derivative Test.

4.3 How Derivatives Affect the Shape of a Graph goo.gl/11C7Ku 15

Example 2. Find the local minimum and maximum values of the function f (x) = 3x4 −
4x3 − 12x2 + 5 in Example 1.

Solution.

x −1 0 2
f 0 5 −27
f′

Hence f has local maximum ; f has local minimum .

Definition 3 (page 296). If the graph f lies above all of it tangents on an interval I, then
it is called concave upward (凹口朝上) on I. If the graph f lies below all of it tangents on an
interval I, then it is called concave downward (凹口朝下) on I. 9-0b-5pEFvM

 有些教科書或文獻使用凸函數 (convex) 取代凹口向上 (concave upward)。

y y

x x

Figure 2: Concave upward and concave downward.

Concavity Test (page 296).

(a) If f ′′ (x) > 0 for all x in I, then the graph of f is concave upward on I.

(b) If f ′′ (x) < 0 for all x in I, then the graph of f is concave downward on I.

Proof of (a). Since f ′′ (x) > 0 in I, we know that f ′ (x) is increasing in I. Given x0 ∈ I, the
tangent line equation to the graph of f (x) at (x0 , f (x0 )) is

y − f (x0 ) = f ′ (x0 )(x − x0 ) ⇒ y = f ′ (x0 )(x − x0 ) + f (x0 ).

We will show that f (x) ≥ f ′ (x0 )(x − x0 ) + f (x0 ) for all x ∈ I.

Consider the function

F (x) = f (x) − f ′ (x0 )(x − x0 ) − f (x0 ) for x ∈ I.

First, we know that F (x0 ) = 0. Next, we compute F ′ (x) = f ′ (x) − f ′ (x0 ), which implies
F ′ (x0 ) = f ′ (x0 ) − f ′ (x0 ) = 0. Since F ′ (x) < 0 for x < x0 and F ′ (x) > 0 for x > x0 , we know
that F (x0 ) is a local (and hence absolute) minimum at x = x0 in I. That means F (x) ≥ 0
for all x ∈ I, thus f (x) ≥ f ′ (x0 )(x − x0 ) + f (x0 ) for all x ∈ I.
 16 4.3 How Derivatives Affect the Shape of a Graph goo.gl/11C7Ku

Definition 4 (page 297). A point P on a curve y = f (x) is called an inflection point (反曲
點) if f is continuous there and the curve changes from concave upward to concave downward
505tEz7OtPc
or from concave downward to concave upward at P .

y y

x x

Figure 3: Inflection points.

Example 5. Find the concave upward and downward intervals, and inflection points of the
function f (x) = 3x4 − 4x3 − 12x2 + 5 in Example 1. Sketch the graph of f .

Solution. We compute

f ′′ (x) = .

So

x −1 x1 0 x2 2
f 0 f (x1 ) 5 f (x2 ) −27
f′ − 0 + 0 − 0 +
f ′′

The points of inflections are .

f is concave upward on .

f is concave downward on .

Figure 4: The graph of f (x) = 3x4 − 4x3 − 12x2 + 5.

4.3 How Derivatives Affect the Shape of a Graph goo.gl/11C7Ku 17

The Second Derivative Test (page 297). Suppose f ′′ is continuous near c.

(a) If f ′ (c) = 0 and f ′′ (c) > 0, then f has a local minimum at c. EU_LBaZWe4k

(b) If f ′ (c) = 0 and f ′′ (c) < 0, then f has a local maximum at c.

sin x
Example 6. Show that f (x) = x is decreasing on (0, π2 ).

Solution.

 比較 Section 2.3, 那時候證明了 | sin x| ≤ |x|。

Example 7. Classify all cubic functions f (x) = ax3 + bx2 + cx + d.

Solution. jFFuiBKJP4M
 18 4.5 Summary of Curve Sketching goo.gl/gJhoUh

4.5 Summary of Curve Sketching, page 315

Definition 1 (page 320). If

bA12vd64GN8 lim (f (x) − (mx + b)) = 0,

x→∞

where m 6= 0, then the line y = mx + b is called a slant asymptote (斜漸近線).

Proposition 2. The graph of f (x) has a slant asymptote if and only if

f (x)
lim = m 6= 0 and lim (f (x) − mx) = b.
x→∞ x x→∞

Proof. When x > 0,

f (x) f (x) − (mx + b) b f (x)

= + m + ⇒ lim = 0 + m + 0 = m.
x x x x→∞ x
lim (f (x) − mx) = lim (f (x) − (mx + b) + b)
x→∞ x→∞

= lim (f (x) − (mx + b)) + lim b = 0 + b = b.

x→∞ x→∞

Conversely, we have

lim (f (x) − (mx + b)) = lim ((f (x) − mx) − b)

x→∞ x→∞

= lim (f (x) − mx) − lim b = b − b = 0.

x→∞ x→∞

 函數圖形當 x → −∞ 也可能存在斜漸近線, 定義與其等價條件都改成 x→−∞

lim 。

 兩個極限都要存在才能稱函數有斜漸近線。 例如 f (x) = ln x 沒有斜漸近線。

Guidelines for sketching a curve

定 Domain: the set of x for which f (x) is defined.

FBWCqsr2iyk
交 Intercepts: y-intercept f (0), x-intercepts: let y = 0 and solve for x.

對 Symmetry: even function, odd function, periodic function.

漸 Asymptotes: horizontal asymptotes, vertical asymptotes, slant asymptotes.

一 Intervals of increase or decrease: use the Increasing/Decreasing test.

極 Local maximum and minimum values: find the critical numbers of f (f ′ (c) = 0 or
f ′ (c) does not exist.)

二 Concavity and points of inflection: compute f ′′ (x) and use the Concavity Test.

圖 Sketch the Curve: use the information in items 1–7, draw the graph.
4.5 Summary of Curve Sketching goo.gl/gJhoUh 19

2x2
Example 1 (page 317). Sketch the curve y = f (x) = x2 −1 .

Solution. eI0fbbIgRy0

A. The domain is .

B. The x- and y-intercept are both .

C. Since , the function f is .

D. Since

2x2
lim = ,
x→±∞ x2 − 1

the line is a . The denominator is 0 when . we

compute the following limits:

2x2 2x2
lim+ = lim− =
x→1 x2 − 1 x→1 x2 − 1
2x2 2x2
lim + 2 = lim − 2 =
x→−1 x − 1 x→−1 x − 1

Therefore the lines and are vertical asymptotes.

E. Direct computation gives

y′ = .

Since f ′ (x) > 0 when and f ′ (x) < 0 when , f is

increasing on and decreasing on .

F. The only critical number is . Since f ′ changes from positive to negative at 0,

f (0) = 0 is a by the First Derivative Test.

G. Direct computation gives

f ′′ (x) = .

We know f ′′ (x) > 0 on and f ′′ (x) < 0 on . Thus the curve is concave
upward on the interval and concave downward on . It has
no point of inflection since .

H. Using this information to sketch the curve. (畫在右上角)

 20 4.5 Summary of Curve Sketching goo.gl/gJhoUh

2
Example 2 (page 317). Sketch the curve y = f (x) = √x .
x+1

3W1Yt_jtLQg
Solution.

A. The domain is .

B. The x- and y-intercept are both .

C. Symmetry: None.

D. Since

x2
lim √ = ,
x→∞ x+1

there is no horizontal asymptote. Since

x2
lim + √ = ,
x→−1 x+1

the line is a vertical asymptotes.

E. Direct computation gives

y′ = .

We see that f ′ (x) = 0 when , so the only critical number is . Since f ′ (x) > 0
when and f ′ (x) < 0 when , f is increasing on and decreasing
on .

F. Since f ′ (0) = 0 and f ′ changes from negative to positive at 0, f (0) = 0 is a

by the First Derivative Test.

G. Direct computation gives

f ′′ (x) = .

Since the numerator is always , we know f ′′ (x) > 0 for all x in the domain of
f , which means f is concave upward on and there is no point of inflection.

H. Using this information to sketch the curve. (畫在右上角)

4.5 Summary of Curve Sketching goo.gl/gJhoUh 21

Example 3 (page 318). Sketch the curve y = f (x) = xex .

Solution. 1Qp4US5GRcA

A. The domain is .

B. The x- and y-intercept are both .

C. Symmetry: None.

D. Since

lim xex = ,
x→∞

there is no horizontal asymptote. By the l’Hospital Rule, we have

x
lim xex = lim = ,
x→−∞ x→−∞ e−x

so the is a horizontal asymptote.

E. Direct computation gives

y′ = .

Since f ′ (x) > 0 when and f ′ (x) < 0 when , f is increasing on

and decreasing on .

F. Since f ′ (−1) = 0 and f ′ changes from negative to positive at x = −1, f (−1) = −e−1 is
a by the First Derivative Test.

G. Direct computation gives

f ′′ (x) = .

Since f ′′ (x) > 0 if and f ′′ (x) < 0 if , f is concave upward on

and concave downward on . The inflection point is .

H. Using this information to sketch the curve.

 22 4.5 Summary of Curve Sketching goo.gl/gJhoUh

cos x
Example 4 (page 319). Sketch the curve y = f (x) = 2+sin x .

wbkisiSCwIU
Solution.

A. The domain is .

B. The x-intercepts are and y-intercept is .

C. Symmetry: f is neither even nor odd. Since f (x + 2π) = f (x) for all x, f is
and has period . Thus, the following steps we only consider 0 ≤ x ≤ 2π and then
extend the curve by translation.

D. Asymptotes: None.

E. Direct computation gives

y′ = .

Thus f ′ (x) > 0 when . So f is increasing

on and decreasing on .

F. From part E and First Derivative Test, the local minimum value is and
local maximum value is .

G. Direct computation gives

f ′′ (x) =

Since f ′′ (x) > 0 if , f is concave upward on and concave downward

on . The inflection point is .

H. Using this information to sketch the curve.

4.5 Summary of Curve Sketching goo.gl/gJhoUh 23

Example 5 (page 319). Sketch the curve y = f (x) = ln(4 − x2 ).

Solution. NwfDfOYvDcc

A. The domain is .

B. The y-intercept is f (0) = ln 4. To find the x-intercept, we set ln(4 − x2 ) = 0, so we

have . Therefore the x-intercepts are .

C. Since f (−x) = f (x), f is and the curve is symmetric about the .

D. Since

lim ln(4 − x2 ) = , lim ln(4 − x2 ) = ,

x→−2+ x→2−

the lines are vertical asymptotes.

E. Direct computation gives

y′ = .

Since f ′ (x) > 0 when and f ′ (x) < 0 when , f is increasing on

and decreasing on .

F. The only critical number is . Since f ′ changes from positive to negative at 0,

f (0) = ln 4 is a by the First Derivative Test.

G. Direct computation gives

f ′′ (x) = .

Since f ′′ (x) < 0 for all x, the curve is on and has no inflection
point.

H. Using this information to sketch the curve.

 24 4.5 Summary of Curve Sketching goo.gl/gJhoUh

x3
Example 6 (page 320). Sketch the curve y = f (x) = x2 +1 .

aoV6RJZFQ2Y
Solution.

A. The domain is .

B. The x- and y-intercept are both .

C. Since , the function f is .

D. Since x2 + 1 is never 0, there is no vertical asymptote. Since f (x) → ∞ as x → ∞ and

f (x) → −∞ as x → −∞, there is no horizontal asymptote. Long division gives

x3
f (x) = = ,
x2 + 1
x
f (x) − x = − = .
x2 +1

So the line is a .

E. Direct computation gives

y′ = .

Since f ′ (x) > 0 when , f is increasing on .

F. Although f ′ (0) = 0, f ′ does not change sign at 0, so there is or

.

G. Direct computation gives

f ′′ (x) = .

Since f ′′ (x) = 0 when , we set up the following chart.

The points of inflection are .

H. Using this information to sketch the curve.

4.7 Optimization Problems goo.gl/gNJP2d 25

4.7 Optimization Problems, page 330

Steps in solving optimization problems
1. Understand the problem: What is the unknown? What are the given quantities?
What are the given conditions?

2. Draw a diagram: In most problems it is useful to draw a diagram and identify the
given and required quantities on the diagram.

3. Introduce notation: Assign a symbol to the quantity that is to be maximized or

minimized (call it Q for now). Also select symbols a, b, c, . . . , x, y for other known
quantities and label the diagram with these symbols.

4. Express Q in terms of some of the other symbols.

5. If Q has been expressed as a function of more than one variable, use the given informa-
tion to find relationships among these variables. Then use these equations to eliminate
all but one of the variables. Thus we get Q = f (x).

6. Use the methods of Section 4.1 and 4.3 to find the absolute maximum or minimum
value of f .

Example 1 (Snell’s Law, 斯乃爾定律, page 268). Let v1 be the velocity of light in air and v2
the velocity of light in water. According to Fermat’s Principle, a ray of light will travel from
a point A in the air to a point B in the water by a path ACB that minimizes the time taken. kZ6gWTJvVNY

Show that
sin θ1 v1
= ,
sin θ2 v2
where θ1 (the angle of incidence) and θ2 (the angle of refraction) are known.

Solution.
 26 4.7 Optimization Problems goo.gl/gNJP2d

Example 2 (搬家俱, page 268). A steel pipe is being carried down a hallway a m wide. At
the end of the hall there is a right-angled turn into a narrower hallway b m wide. What is the
eQpoM26CaM4
length of the longest pipe that can be carried horizontally around the corner?

Solution.

Example 3 (看畫、 教室的風水, page 269). A painting in an art gallery has height h and is
hung so that its lower edge is a distance d above the eye of an observer. How far from the wall
t2MdXHcMYZI should the observer stand to get the best view? (In other words, where should the observer
stand so as to maximize the angle θ subtended at his eye by the painting?)

Solution.
4.7 Optimization Problems goo.gl/gNJP2d 27

Example 4. A right circular cone is inscribed in a sphere of radius r. Find the largest
possible volume of such a cone. In this case, what is the height and radius of the cone?
hbKrEiTBiqA

Solution.

Example 5 (折紙問題, page 269). The upper right-hand corner of a piece of paper, 30 cm by
20 cm, is folded over to the bottom edge. How would you fold it so as to minimize the length
of the fold? In other words, how would you choose x to minimize y? FzTOC_OCRC0

Solution.
 28 4.7 Optimization Problems goo.gl/gNJP2d

Example 6.

N-KwPosBmfE
(a) Find the point (denote P ) on the line y = x2 that is closest to the point Q(3, 0).

(b) Show that the line P Q is orthogonal to the tangent line of y = x2 at P .

Solution.

Example 7 (電影的極佳位置).

CzCCC_7mTJg
4.9 Antiderivative goo.gl/q3mCPB 29

4.9 Antiderivative, page 350

Definition 1 (page 350). A function F is called an antiderivative (反導函數) of f on an
interval I if F ′ (x) = f (x) for all x in I.
3ES4KjEu5gU

Theorem 2 (page 351). If F is an antiderivative of f on an interval I, then the most general

antiderivative of f on I is

F (x) + C,

where C is an arbitrary constant.

Proof. If F and G are any two antiderivative of f , then F ′ (x) = f (x) = G′ (x). Form the
corollary of the Mean Value Theorem (Section 4.2 Corollary 8), we know G(x) − F (x) = C,
where C is a constant. So G(x) = F (x) + C.

This is a table of antidifferentiation formulas. We use the notation F ′ (x) = f (x) and
G′ (x) = g(x).
gwp9cLFo8Ac

Function Particular antiderivative Function Particular antiderivative

cf (x) cF (x) sec2 x tan x
f (x) + g(x) F (x) + G(x) sec x tan x sec x
xn+1 √ 1
xn (n 6= −1) n+1 1−x2
sin−1 x
1 1
x ln |x| 1+x2 tan−1 x
ex ex cosh x sinh x
cos x sin x sinh x cosh x
sin x − cos x

Example 3. Find the most general antiderivative of the function. (Let F (x) is the antideriva-
tive of the function f (x).)

(1) f (x) = e2 F (x) =

(2) f (x) = x(2 − x)2

F (x) =

(3) f (x) = xπ − x3.14 F (x) =

2+x2
(4) f (x) = 1+x2

F (x) =
 30 4.9 Antiderivative goo.gl/q3mCPB

Example 4. Find f (x).

nb2j0fedk9w
(1) f ′ (x) = 2 cos x + sec2 x, − π2 < x < π2 , f ( π3 ) = 4.

(2) f ′′ (x) = 2ex + 3 sin x, f (0) = 0, f (π) = 0.

Solution.

 以上方程式稱為 「帶有初始條件的微分方程式」 (ordinary differential equation)。