Real Analysis 1

Prof. Omar Hirzallah

Department of Mathematics
The Hashemite University
 Course Title: Real Analysis (1) Course Number: 110101211
Pre-requisite: 1101011151 Credit Hours: 3
Designation: Compulsory Instructor: Prof. Omar Hirzallah
Text Book: Introduction to Real Analysis, R.G. Bartle and D.R. Sherbert, John Wiley and Sons,
Inc., 2000, 4th Edition
Table of Contents

Topics Page

Chapter 2: The real Numbers 1

2.1 The algebraic and Order Properties of the Real Numbers. 1
2.2 The Absolute Value and the Real Line 4
2.3 The Completeness Property of ℝ 10
2.4 Applications of the Supremum Property 16
2.5 Intervals 22
Chapter 3: Sequences and Series 26
3.1 Sequences and their Limits 26
3.2 Limit Theorems 34
3.3 Monotone Sequences 43
3.4 Subsequences and the Bolzano-Weierstrass Theorem 48
3.5 The Cauchy Criterion 55
3.6 Properly Divergent Sequences 61
3.7 Introduction to Infinite Series 68
Chapter 4: Limits 75
4.1 Limits of Functions 75
4.2 Limit Theorems 81
4.3 Some Extensions of the Limit Concept 90
Chapter 5: Continuous Functions 94
5.1 Continuous Functions 94
5.2 Combinations of Continuous Functions 100
5.3 Continuous Functions on Intervals 103
5.4 Uniform Continuity 108
5.5 Continuity and Gages ---
5.6 Monotone and Inverse Functions 117
Chapter 6: Differentiation 120
6.1 The Derivative 120
6.2 The Mean Value Theorem 123
6.3 L’Hospital’s Rules 127
 1

Chapter 2: The real Numbers

Section 2.1: The Algebraic and Order Properties of ℝ

Notations 2.1.1:
(1) ℝ: The set of all real numbers ⇒ ℝ = (−∞, ∞) = {𝑥: −∞ < 𝑥 < ∞}
(2) ℚ: The set of rational numbers.
(3) ℕ: The set of all natural numbers, that is ℕ = {1,2,3, … }
The Trichotomy Property 2.1.2 (‫)الخاصية الثالثية على األعداد الحقيقية‬:
If 𝑎 ∈ ℝ, then 𝑎 > 0, 𝑎 = 0, or 𝑎 < 0.

Note that: (1) 𝑎 > 0 or 𝑎 = 0 ⇔ 𝑎 ≥ 0

(2) 𝑎 < 0 or 𝑎 = 0 ⇔ 𝑎 ≤ 0
So, the Trichotomy property can be written as:
If 𝑎 ∈ ℝ, then 𝑎 ≥ 0 or 𝑎 < 0
𝑎 ‫يستخدم هذا األسلوب لتقليل الحاالت المحتملة للعدد الحقيقي‬
Or }
‫لتصبح حالتين بدالً من ثالث حاالت‬
If 𝑎 ∈ ℝ, then 𝑎 > 0 or 𝑎 ≤ 0

Theorem 2.1.3:
(1) If 𝑎 ∈ ℝ such that 0 ≤ 𝑎 ≤ 𝜀, ∀𝜀 > 0, then 𝑎 = 0.
(2) If 𝑎 ∈ ℝ such that 0 ≤ 𝑎 < 𝜀, ∀𝜀 > 0, then 𝑎 = 0.
(3) If 𝑎 ∈ ℝ such that −𝜀 ≤ 𝑎 ≤ 𝜀, ∀𝜀 > 0, then 𝑎 = 0.
(4) If 𝑎 ∈ ℝ such that −𝜀 < 𝑎 < 𝜀, ∀𝜀 > 0, then 𝑎 = 0.
Proof: We prove part (1). The proofs of the other parts (2)&(3) are similar.
Suppose that 0 ≤ 𝑎 ≤ 𝜀, ∀𝜀 > 0. The proof is by contradiction: Assume that 𝑎 ≠ 0.
Since 0 ≤ 𝑎 ≤ 𝜀, ∀𝜀 > 0 ⇒ 𝑎 ≥ 0……① and 𝑎 ≤ 𝜀, ∀𝜀 > 0……..②
𝑎 𝑎
Since 𝑎 ≠ 0, then ① implies that 𝑎 > 0. So, in ② take 𝜀 = 2 we have 𝑎 ≤ 2
1
⇒ 1 ≤ 2 which is impossible ⇒ We have a contradiction ⇒ Our assumption is false

⇒ 𝑎 = 0.
 2

Example 2.1.4:
1
(1) If 𝑎 ∈ ℝ such that 0 ≤ 𝑎 ≤ 𝑛 , ∀𝑛 ∈ ℕ, then 𝑎 = 0.
1
(2) If 𝑎 ∈ ℝ such that 0 ≤ 𝑎 < 𝑛 , ∀𝑛 ∈ ℕ, then 𝑎 = 0.
1 1
(3) If 𝑎 ∈ ℝ such that − 𝑛 < 𝑎 < 𝑛 , ∀𝑛 ∈ ℕ, then 𝑎 = 0.

Solution: We prove part (1). The proofs of parts (2)&(3) are exercises.
Let 𝜀 > 0 be given.
1 1
Since 𝜀 > 0 and lim = 0, then ∃𝑛0 ∈ ℕ such that 𝑛 < 𝜀
𝑛→∞ 𝑛 0

1 1
But 0 ≤ 𝑎 ≤ 𝑛 , ∀𝑛 ∈ ℕ ⇒ 0 ≤ 𝑎 ≤ 𝑛 , since 𝑛0 ∈ ℕ
0

1
But 𝑛 < 𝜀 ⇒ 0 ≤ 𝑎 < 𝜀 and since 𝜀 is arbitrary we have 0 ≤ 𝑎 < 𝜀, ∀𝜀 > 0
0

Theorem 3 ⇒ 𝑎 = 0.
Corollary 2.1.5:
(1) If 𝑎, 𝑏 ∈ ℝ such that 𝑎 ≤ 𝑏 + 𝜀, ∀𝜀 > 0, then 𝑎 ≤ 𝑏.
(2) If 𝑎, 𝑏 ∈ ℝ such that 𝑎 < 𝑏 + 𝜀, ∀𝜀 > 0, then 𝑎 ≤ 𝑏.
1
(3) If 𝑎, 𝑏 ∈ ℝ such that 𝑎 ≤ 𝑏 + 𝑛 , ∀𝑛 ∈ ℕ, then 𝑎 ≤ 𝑏.
1
(4) If 𝑎, 𝑏 ∈ ℝ such that 𝑎 < 𝑏 + 𝑛 , ∀𝑛 ∈ ℕ, then 𝑎 ≤ 𝑏.

Proof: We prove part (1). The proofs of the other parts are similar.
Suppose that 𝑎 ≤ 𝑏 + 𝜀, ∀𝜀 > 0. The proof is by contradiction:
Assume that 𝑎 > 𝑏. (we must have a contradiction)
𝑎 > 𝑏 ⇒ 𝑎 − 𝑏 > 0 ⇒ 𝑎 − 𝑏 ≥ 0……..①
𝑎 ≤ 𝑏 + 𝜀, ∀𝜀 > 0 ⇒ 𝑎 − 𝑏 ≤ 𝜀, ∀𝜀 > 0……..②
①&② ⇒ 0≤ 𝑎 − 𝑏 ≤ 𝜀, ∀𝜀 > 0
Theorem 3 ⇒ 𝑎 − 𝑏 = 0 ⇒ 𝑎 = 𝑏 which is a contradiction ⇒ Our assumption is false
⇒ 𝑎 ≤ 𝑏.

Definition 2.1.6:

(1) The geometric mean of two positive real numbers 𝑎 and 𝑏 is √𝑎𝑏.
𝑎+𝑏
(2) The arithmetic mean of two real numbers 𝑎 and 𝑏 is .
2
 3

The Arithmetic-Geometric Mean Inequality 2.1.7:

𝑎+𝑏
If 𝑎, 𝑏 be positive real numbers, then √𝑎𝑏 ≤ with equality iff 𝑎 = 𝑏.
2

Proof: (𝑎 − 𝑏)2 ≥ 0 ⇔ ⏟
𝑎2 − 2𝑎𝑏 + 𝑏 2 ≥ 0 ⇔ 𝑎2 + 2𝑎𝑏 + 𝑏 2 ≥ 4𝑎𝑏 ⇔ ⏟
(𝑎 + 𝑏)2 ≥ 4𝑎𝑏
4𝑎𝑏 ‫اجمع للطرفين‬ ‫نأخذ الجذر التربيعي للطرفين‬

𝑎+𝑏
⇔ √(𝑎 + 𝑏)2 ≥ √4𝑎𝑏 ⇔ 𝑎 + 𝑏 ≥ 2√𝑎𝑏 ⇔ √𝑎𝑏 ≤ 2
𝑎+𝑏
For Equality: √𝑎𝑏 = iff (𝑎 − 𝑏)2 = 0 iff 𝑎 = 𝑏.
2

Bernoulli Inequality 2.1.8: If 𝑥 ∈ ℝ such that 𝑥 ≥ −1, then (1 + 𝑥)𝑛 ≥ 1 + 𝑛𝑥, 𝑛 ∈ ℕ.

Proof: The proof is by induction:
Step 1: We show that the inequality is true at 𝑛 = 1:
𝐿𝐻𝑆 = (1 + 𝑥)𝑛 = (1 + 𝑥)1 = 1 + 𝑥 } ⇒ 𝐿𝐻𝑆 = 𝑅𝐻𝑆
𝑅𝐻𝑆 = 1 + 𝑛𝑥 = 1 + 𝑥
⇒ 𝐿𝐻𝑆 ≥ 𝑅𝐻𝑆 ⇒ The inequality holds when 𝑛 = 1
Step 2: Suppose that the inequality is true at 𝑛 = 𝑘, that is
(1 + 𝑥)𝑘 ≥ 1 + 𝑘𝑥 … … …①
Step 3: At 𝑛 = 𝑘 + 1 we have
𝐿𝐻𝑆 = (1 + 𝑥)𝑘+1
= (1 + 𝑥)𝑘 (1 + 𝑥)
≥ (1 + 𝑘𝑥)(1 + 𝑥) by Step 2 (see the relation ①)
= 1 + 𝑥 + 𝑘𝑥 + 𝑘𝑥 2
= 1 + (1 + 𝑘)𝑥 + 𝑘𝑥 2
≥ 1 + (1 + 𝑘)𝑥 (since 𝑘𝑥 2 ≥ 0)
= 𝑅𝐻𝑆
The inequality holds at 𝑛 = 𝑘 + 1.
∴ The inequality holds ∀𝑛 ∈ ℕ.

Exercises 2.1.9 (Pages 30-32): 10,12,14,15,16,17,20

 4

Section 2.2: The Absolute Value and the Real Line

Definition 2.2.1: The absolute value of a real number 𝑎 ∈ ℝ, written as |𝑎|, is defined by
𝑎 , 𝑎≥0
|𝑎| = { }
−𝑎 , 𝑎 < 0
𝑥 , 𝑥≥0
As a function: Let 𝑓(𝑥) = |𝑥|, 𝑥 ∈ ℝ ⇒ 𝑓(𝑥) = { }
−𝑥 , 𝑥 < 0

The graph of 𝑓(𝑥) = |𝑥|

Remark 2.2.2:
(1) |−2| = 2, |2| = 2, |0| = 0
(2) |𝑎| = |−𝑎|, ∀𝑎 ∈ ℝ.
(3) |𝑥| ≥ 0, ∀𝑥 ∈ ℝ.
(4) |𝑥 6 + 1| = 𝑥 6 + 1, ∀𝑥 ∈ ℝ (since 𝑥 6 + 1 > 0, ∀𝑥 ∈ ℝ)
𝑥 2 − 2𝑥 − 3 , 𝑥 2 − 2𝑥 − 3 ≥ 0
(5) |𝑥 2 − 2𝑥 − 3| = { }
−(𝑥 2 − 2𝑥 − 3) , 𝑥 2 − 2𝑥 − 3 < 0
𝑥 2 − 2𝑥 − 3 = 0 ⇒ (𝑥 − 3)(𝑥 + 1) = 0 ⇒ 𝑥 = −1,3

𝑥 2 − 2𝑥 − 3 , 𝑥 ≤ −1 or 𝑥 ≥ 3
|𝑥 2 − 2𝑥 − 3| = { }
3 + 2𝑥 − 𝑥 2 , −1 < 𝑥 < 3
Theorem 2.2.3: Let 𝑎, 𝑏 ∈ ℝ. Then
𝑎 |𝑎|
(1) |𝑎𝑏| = |𝑎| |𝑏| (2) |𝑏| = |𝑏|, where 𝑏 ≠ 0
Proof: We prove part (1). The proof of part (2) is similar.
① ② ③ ④

(1) We have 4 cases: ⏞

𝑎 ≥ 0, 𝑏 ≥ 0 , ⏞
𝑎 ≥ 0, 𝑏 < 0 , ⏞
𝑎 < 0, 𝑏 ≥ 0 , ⏞
𝑎 < 0, 𝑏 < 0
|𝑎| = 𝑎
Case 1: 𝑎 ≥ 0, 𝑏 ≥ 0 ⇒ 𝑎𝑏 ≥ 0: { |𝑏| = 𝑏 } ⇒ |𝑎𝑏| = 𝑎𝑏 = |𝑎| |𝑏|
|𝑎𝑏| = 𝑎𝑏
|𝑎| = 𝑎
Case 2: 𝑎 ≥ 0, 𝑏 < 0 ⇒ 𝑎𝑏 ≤ 0: { |𝑏| = −𝑏 } ⇒ |𝑎𝑏| = 𝑎(−𝑏) = |𝑎| |𝑏|
|𝑎𝑏| = −𝑎𝑏
|𝑎| = −𝑎
Case 3: 𝑎 < 0, 𝑏 ≥ 0 ⇒ 𝑎𝑏 ≤ 0: { |𝑏| = 𝑏 } ⇒ |𝑎𝑏| = (−𝑎)𝑏 = |𝑎| |𝑏|
|𝑎𝑏| = −𝑎𝑏
|𝑎| = −𝑎
Case 4: 𝑎 < 0, 𝑏 < 0 ⇒ 𝑎𝑏 > 0: { |𝑏| = −𝑏 } ⇒ |𝑎𝑏| = 𝑎𝑏 = (−𝑎)(−𝑏) = |𝑎| |𝑏|
|𝑎𝑏| = 𝑎𝑏
Remark 2.2.4:
(1) If 𝑎 ∈ ℝ, then 𝑎2 = |𝑎|2. To prove this observe that (‫)الحظ أن‬:
𝑎2 = |𝑎2 | =
⏟ |𝑎𝑎| = |𝑎| |𝑎|
⏟ = |𝑎|2
since 𝑎2 ≥0 Theorem 2.2.3 part (1)

(2) From part (1): 𝑎2 = |𝑎|2 ⇒ √𝑎2 = √|𝑎|2 = |𝑎|

 5

√16 ⏟ = ±3 :‫ بعض الطالب يخطئون في حسابات الجذر التربيعي فمثال يكتبون‬

= ±4 , √9 ⏟
‫هذا خطأ‬ ‫هذا خطأ‬
‫ تذكر أن 𝑥√ هو اقتران وبالتالي كل عنصر في المجال له‬
‫صورة واحدة فقط وهذه الصورة يجب أن تكون قيمتها أكبر‬
‫من أو تساوي صفر وال يجوز أن تكون سالبة‬
:‫بناءا ً على ذلك فإن‬ 
√9 = |3| = 3, √16 = |4| = 4

Recall the following: Let (1) ⏟

𝑃∧𝑇 ⇔𝑃
𝑃: proposition 𝑃 and 𝑇
𝑇: true proposition (tautology) (2) ⏟
𝑃∨𝐹 ⇔𝑃
𝐹: false proposition (contradiction) 𝑃 or 𝐹

Theorem 2.2.5: Let 𝑎 ∈ ℝ such that 𝑎 > 0. Then

(1) |𝑥| ≤ 𝑎 iff ⏟
−𝑎 ≤ 𝑥 ≤ 𝑎 iff 𝑥 ∈ [−𝑎, 𝑎] (2) |𝑥| < 𝑎 iff ⏟
−𝑎 < 𝑥 < 𝑎 iff 𝑥 ∈ (−𝑎, 𝑎)
means means
𝑥≥−𝑎 and 𝑥≤𝑎 𝑥>−𝑎 and 𝑥<𝑎

(3) |𝑥| ≥ 𝑎 iff 𝑥 ≥ 𝑎 or 𝑥 ≤ −𝑎 (4) |𝑥| > 𝑎 iff 𝑥 > 𝑎 or 𝑥 < −𝑎

iff 𝑥 ∈ (−∞, −𝑎] ∪ [𝑎, ∞) iff 𝑥 ∈ (−∞, −𝑎) ∪ (𝑎, ∞)

Proof: We prove parts (1) and (3). The proofs of parts (2) and (4) are similar.
(1) Since 𝑥 ∈ ℝ ⇒ We have two cases: 𝑥 ≥ 0 or 𝑥 < 0
Case 1: Let 𝑥 ≥ 0 ⇒ |𝑥| = 𝑥. Now,
|𝑥| ≤ 𝑎 ⇔ 𝑥 ≤ 𝑎 ⇔ 𝑥 ⏟≤ 𝑎 and ⏟ 𝑥 ≥ −𝑎 ⇔ −𝑎 ≤ 𝑥 ≤ 𝑎
proposition Tautology
Case 2: Let 𝑥 < 0 ⇒ |𝑥| = −𝑥. Now,
|𝑥| ≤ 𝑎 ⇔ −𝑥 ≤ 𝑎 ⇔ −𝑎 ≤ 𝑥 ⇔ ⏟ −𝑎 ≤ 𝑥 and 𝑥
⏟≤ 𝑎 ⇔ −𝑎 ≤ 𝑥 ≤ 𝑎
proposition Tautology
(3) Since 𝑥 ∈ ℝ ⇒ We have two cases: 𝑥 ≥ 0 or 𝑥 < 0
Case 1: Let 𝑥 ≥ 0 ⇒ |𝑥| = 𝑥. Now: |𝑥| ≥ 𝑎 ⇔ 𝑥 ≥ 𝑎 ⇔ 𝑥
⏟≥ 𝑎 or ⏟
𝑥 ≤ −𝑎
proposition cotradiction
Case 2: Let 𝑥 < 0 ⇒ |𝑥| = −𝑥. Now,
|𝑥| ≥ 𝑎 ⇔ −𝑥 ≥ 𝑎 ⇔ 𝑥 ≤ −𝑎 ⇔ 𝑥 ⏟≥ 𝑎 or ⏟
𝑥 ≤ −𝑎
contradiction proposition
Example 2.2.6: Solve the following equations and inequalities (‫)متباينات‬
(1) |𝑥 + 1| + |𝑥 − 2| = 7 (2) |𝑥| + |𝑥 + 1| < 2
(3) |2𝑥 − 1| < 𝑥 + 1 (4) |𝑥 − 1| ≤ |𝑥|
Solution: We solve parts (1) and (2). The solution of parts (3) and (4) are exercise.
 6

(1)

(2)
 7

Theorem 2.2.7: If 𝑎 ∈ ℝ, then −|𝑎| ≤ 𝑎 ≤ |𝑎|.

Proof: We have two cases: 𝑎
⏟≥ 0 or 𝑎
⏟< 0
① ②
‫سالب‬ ‫موجب‬
⏞≤ ⏞
Case 1: Let 𝑎 ≥ 0 ⇒ |𝑎| = 𝑎 ⇒ −|𝑎| 𝑎 = |𝑎| ⇒ −|𝑎| ≤ 𝑎 ≤ |𝑎|.
‫سالب‬ ‫موجب‬

Case 2: Let 𝑎 < 0 ⇒ |𝑎| = −𝑎 ⇒ −|𝑎| = ⏞ ⏞ ⇒ −|𝑎| ≤ 𝑎 ≤ |𝑎|.

𝑎 < |𝑎|

Remark 2.2.8: Recall that if 𝑎 ∈ ℝ, then

(1) −|𝑎| ≤ 𝑎 ≤ |𝑎| ⇔ ±𝑎 ≤ |𝑎|.
⏟
Theorem 2.2.7
2 |𝑎2 | |𝑎|2
(2) 𝑎 = =
(3) 2
√𝑎 = |𝑎|.
The Triangle Tnequality 2.2.9 (‫)المتباينة المثلثية‬: If 𝑎, 𝑏 ∈ ℝ, then
(1) |𝑎 + 𝑏| ≤ |𝑎| + |𝑏| with equality iff 𝑎𝑏 ≥ 0
(2) |𝑎 − 𝑏| ≤ |𝑎| + |𝑏| with equality iff 𝑎𝑏 ≤ 0
Proof:
(1) (𝑎 + 𝑏)2 = 𝑎2 + 2𝑎𝑏 + 𝑏 2 ≤ 𝑎2 + 2|𝑎𝑏| + 𝑏 2 = |𝑎|2 + 2|𝑎| |𝑏| + |𝑏|2 = (|𝑎| + |𝑏|)2
⇒ (𝑎 + 𝑏)2 ≤ (|𝑎| + |𝑏|)2 ⇒ √(𝑎 + 𝑏)2 ≤ √(|𝑎| + |𝑏|)2
⇒ |𝑎 + 𝑏| ≤ ||𝑎| + |𝑏|| but ||𝑎| + |𝑏|| = |𝑎| + |𝑏|
⇒ |𝑎 + 𝑏| ≤ |𝑎| + |𝑏|
For equality: |𝑎 + 𝑏| = |𝑎| + |𝑏| iff |𝑎𝑏| = 𝑎𝑏 iff 𝑎𝑏 ≥ 0
(2) |𝑎 − 𝑏| = |𝑎 + (−𝑏)| ≤ |𝑎| + |−𝑏| = |𝑎| + |𝑏|
|𝑎 − 𝑏| = |𝑎| + |𝑏| iff 𝑎(−𝑏) ≥ 0 iff −𝑎𝑏 ≥ 0 iff 𝑎𝑏 ≤ 0
For equality: ⏟
by part (1)

‫) في المثلث يكون مجموع طولي أي ضلعين أكبر من أو يساوي طول الضلع الثالث‬The triangle inequality( ‫المتباينة المثلثية‬

 |𝑎 + 𝑏| ≤ |𝑎| + |𝑏| and |𝑎| ≤ |𝑎 + 𝑏| + |𝑏| and |𝑏| ≤ |𝑎 + 𝑏| + |𝑎|

 |𝑎 − 𝑏| ≤ |𝑎| + |𝑏| and |𝑎| ≤ |𝑎 − 𝑏| + |𝑏| and |𝑏| ≤ |𝑎 − 𝑏| + |𝑎|

Example 2.2.10: Are there real numbers 𝑎, 𝑏 such that:

(1) |𝑎| = 2, |𝑏| = 6 and |𝑎 + 𝑏| = 3
(2) |𝑎| = 3, |𝑏| = 8 and |𝑎 − 𝑏| = 5
Solution:
(1) We have to check 3 cases all must hold:
|𝑎 + 𝑏| ≤ |𝑎| + |𝑏| ⇒ 3 ≤ 6 + 2 is true
|𝑎| ≤ |𝑎 + 𝑏| + |𝑏| ⇒ 2 ≤ 3 + 6 is true
|𝑏| ≤ |𝑎 + 𝑏| + |𝑎| ⇒ 6 ≤ 3 + 2 is false
∴ There are no real numbers 𝑎, 𝑏 such that |𝑎| = 2, |𝑏| = 6 and |𝑎 + 𝑏| = 3
 8

(2) We have to check 3 cases all must hold:

|𝑎 − 𝑏| ≤ |𝑎| + |𝑏| ⇒ 5 ≤ 3 + 8 is true
|𝑎| ≤ |𝑎 − 𝑏| + |𝑏| ⇒ 3 ≤ 5 + 8 is true
|𝑏| ≤ |𝑎 − 𝑏| + |𝑎| ⇒ 8 ≤ 5 + 3 is true
∴ There are real numbers 𝑎, 𝑏 such that |𝑎| = 3, |𝑏| = 8 and |𝑎 − 𝑏| = 5

The Generalized Triangle Inequality 2.2.11: Let 𝑎1 , 𝑎2 , … , 𝑎𝑛 ∈ ℝ. Then

|𝑎1 + 𝑎2 + ⋯ + 𝑎𝑛 | ≤ |𝑎1 | + |𝑎2 | + ⋯ + |𝑎𝑛 |
Proof: By induction:
 For 𝑛 = 2: |𝑎1 + 𝑎2 | ≤ |𝑎1 | + |𝑎2 | is true (by the triangle inequality 2.2.9)
 For 𝑛 = 𝑘: Suppose that |𝑎1 + 𝑎2 + ⋯ + 𝑎𝑘 | ≤ |𝑎1 | + |𝑎2 | + ⋯ + |𝑎𝑘 |…….①
 For 𝑛 = 𝑘 + 1: |𝑎1 + 𝑎2 + ⋯ + 𝑎𝑘+1 | = |(𝑎1 + 𝑎2 + ⋯ + 𝑎𝑘 ) + 𝑎𝑘+1 |
≤ |𝑎1 + 𝑎2 + ⋯ + 𝑎𝑘 | + |𝑎𝑘+1 | (by the triangle inequality 9)
≤ |𝑎1 | + |𝑎2 | + ⋯ + |𝑎𝑘 | + |𝑎𝑘+1 | (by ①)
⇒ |𝑎1 + 𝑎2 + ⋯ + 𝑎𝑘+1 | ≤ |𝑎1 | + |𝑎2 | + ⋯ + |𝑎𝑘 | + |𝑎𝑘+1 |
⇒ It is true at 𝑛 = 𝑘 + 1
Theorem 2.2.12: If 𝑎, 𝑏 ∈ ℝ, then
(1) ||𝑎| − |𝑏|| ≤ |𝑎 + 𝑏|
(2) ||𝑎| − |𝑏|| ≤ |𝑎 − 𝑏|
Proof:
(1) |𝑎| = |𝑎 + 𝑏 − 𝑏| ≤ |𝑎 + 𝑏| + |𝑏| (by the triangle inequality)
⇒ |𝑎| − |𝑏| ≤ |𝑎 + 𝑏| ………..①
|𝑏| = |𝑎 + 𝑏 − 𝑎| ≤ |𝑎 + 𝑏| + |𝑎| (by the triangle inequality)
⇒ −|𝑎 + 𝑏| ≤ |𝑎| − |𝑏| ………..②
①&②: ⇒ −|𝑎 + 𝑏| ≤ |𝑎| − |𝑏| ≤ |𝑎 + 𝑏| ⇒ ||𝑎| − |𝑏|| ≤ |𝑎 + 𝑏|

(2) ||𝑎| − |𝑏|| = ||𝑎| − |−𝑏||

≤ |𝑎 + (−𝑏)| (by part (1))
= |𝑎 − 𝑏|
Definition 2.2.13: Let 𝑎 ∈ ℝ and 𝜀 > 0. The 𝜀-neighborhood of 𝑎, written as 𝑉𝜀 (𝑎), is defined as
𝑉𝜀 (𝑎) = (𝑎 − 𝜀, 𝑎 + 𝜀) = {𝑥 ∈ ℝ: 𝑎 − 𝜀 < 𝑥 < 𝑎 + 𝜀}.

Note that: 𝑎 − 𝜀 < 𝑥 < 𝑎 + 𝜀 ⇔ −𝜀 < 𝑥 − 𝑎 < 𝜀 ⇔ |𝑥 − 𝑎| < 𝜀

∴ 𝑉𝜀 (𝑎) = {𝑥 ∈ ℝ: |𝑥 − 𝑎| < 𝜀}
 9

Theorem 2.2.14: Let 𝑎 ∈ ℝ and 𝑥 ∈ 𝑉𝜀 (𝑎), ∀𝜀 > 0. Then 𝑥 = 𝑎

Proof:
𝑥 ∈ 𝑉𝜀 (𝑎), ∀𝜀 > 0 ⇒ 𝑥 ∈ {𝑦 ∈ ℝ: |𝑦 − 𝑎| < 𝜀}, ∀𝜀 > 0 ⇒ |𝑥 − 𝑎| < 𝜀, ∀𝜀 > 0
0 ≤ |𝑥 − 𝑎| < 𝜀, ∀𝜀 > 0 ⇒ |𝑥 − 𝑎| = 0 (by Theorem 2.1.3)
⇒ 𝑥 − 𝑎 = 0 ⇒ 𝑥 = 𝑎.

Recall that: If 𝐴 and 𝐵 are two sets, then 𝐴 ⊆ 𝐵 iff [𝑥 ∈ 𝐴 ⇒ 𝑥 ∈ 𝐵]

Example 2.2.15: Let 0 < 𝑎 < 1 and 0 < 𝜀 < min{𝑎, 1 − 𝑎}. Show that 𝑉𝜀 (𝑎) ⊆ (0,1)
Proof: Let 𝑥 ∈ 𝑉𝜀 (𝑎) ⇒ |𝑥 − 𝑎| < 𝜀 ⇒ |𝑥 − 𝑎| < min{𝑎, 1 − 𝑎}. So,
 |𝑥 − 𝑎| < 𝑎 ⇒ −𝑎 < 𝑥 − 𝑎 < 𝑎 ⇒ 0 < 𝑥 < 2𝑎 ⇒𝑥 > 0…….①
|𝑥 − 𝑎| < 1 − 𝑎 ⇒ −(1 − 𝑎) < 𝑥 − 𝑎 < 1 − 𝑎 ⇒ −1 + 2𝑎 < 𝑥 < 1 ⇒ 𝑥 < 1…….②
①&② ⇒ 0 < 𝑥 < 1 ⇒ 𝑥 ∈ (0,1). So, 𝑉𝜀 (𝑎) ⊆ (0,1)

Example 2.2.16: Let 𝑎, 𝑏 ∈ ℝ and 𝜀 > 0. Show if 𝑥 ∈ 𝑉𝜀 (𝑎) and 𝑦 ∈ 𝑉𝜀 (𝑏), then 𝑥 + 𝑦 ∈ 𝑉2𝜀 (𝑎 + 𝑏)
Proof: 𝑥 ∈ 𝑉𝜀 (𝑎) ⇒ |𝑥 − 𝑎| < 𝜀 and 𝑦 ∈ 𝑉𝜀 (𝑎) ⇒ |𝑦 − 𝑏| < 𝜀
|𝑥 + 𝑦 − (𝑎 + 𝑏)| = ⏟
|(𝑥 − 𝑎) + (𝑦 − 𝑏)| ≤ |𝑥 − 𝑎| + |𝑦 − 𝑏| < 𝜀 + 𝜀
by the triangle inequality

|𝑥 + 𝑦 − (𝑎 + 𝑏)| < 2𝜀 ⇒ 𝑥 + 𝑦 ∈ 𝑉2𝜀 (𝑎 + 𝑏)

Exercises 2.2.17 (Pages 35-36): 5,6,7,8,9,10,11,12,13,14,15,16,17,18,19

 10

Section 2.3: The Completeness Property of ℝ

Definition 2.3.1:
(1) A nonempty set 𝑆 is called bounded (written as bdd) above if ∃𝑢 ∈ ℝ such that
𝑥 ≤ 𝑢, ∀𝑥 ∈ 𝑆, that is 𝑆 ⊆ (−∞, 𝑢].
The real number 𝑢, if exists, is called an upper bound (written as UB)
[𝑢 is not UB of a set 𝑆] iff [∃𝑥 ∈ 𝑆 such that 𝑥 > 𝑢]

∴ 𝑆 is not bdd above ⇔ ∃ sequnce (𝑥𝑛 ) in 𝑆 such that lim 𝑥𝑛 = ∞

𝑛→∞

(2) A real number 𝑢∗ is called a supremum (or a least UB) of a set 𝑆 (written sup 𝑆 = 𝑢∗ ) if
the following two conditions hold:
𝐶1 : 𝑢∗ is an UB of 𝑆.
𝐶2 : if 𝑢 is an UB of 𝑆, then 𝑢∗ ≤ 𝑢.

(3) A nonempty set 𝑆 is called bdd below if ∃𝑣 ∈ ℝ such that 𝑥 ≥ 𝑣, ∀𝑥 ∈ 𝑆, that is

𝑆 ⊆ [𝑢, ∞).

[𝑢 is not LB of a set 𝑆] iff [∃𝑥 ∈ 𝑆 such that 𝑥 < 𝑢]

∴ 𝑆 is not bdd below ⇔ ∃ sequnce (𝑥𝑛 ) in 𝑆 such that lim 𝑥𝑛 = −∞

𝑛→∞

(4) A real number 𝑣 ∗ is called an infimum (or a greatest LB) of a set 𝑆 (written inf 𝑆 = 𝑣 ∗ ) if
the following two conditions hold:
𝐶1 : 𝑣 ∗ is a LB of 𝑆.
𝐶2 : if 𝑣 is a LB of 𝑆, then 𝑣 ∗ ≥ 𝑣.

Example 2.3.2:
1 1 1 1
(1) Consider the set 𝑆 = {𝑛 : 𝑛 ∈ ℕ} ⇒ 𝑆 = {1, 2 , 3 , 4 , … } (sketch the graph of 𝑆 on the real
line if possible).
 sup 𝑆 = 1 and the set of all UB’s of 𝑆 is [1, ∞) ⇒ 𝑆 is bdd above
 inf 𝑆 = 0 and the set of all LB’s of 𝑆 is (−∞, 0] ⇒ 𝑆 is bdd below
⇒ 𝑆 is a bdd set.
 11

(2) Consider the set (−1, ∞). (sketch the graph of 𝑆 on the real line if possible).
 sup (−1, ∞) = ∞ ⇒ sup (−1, ∞) does not exist ⇒ The interval (−1, ∞) has no UB.
⇒ The interval (−1, ∞) is not bdd above
 inf (−1, ∞) = −1 ⇒ The set of all LB’s of (−1, ∞) is the set (−∞, −1]
⇒ (−1, ∞) is bdd below ⇒ (−1, ∞) is not a bdd set (since it is not bdd above)

(3) Consider the set ℝ = (−∞, ∞) ⇒ sup ℝ = ∞ and inf ℝ = −∞

⇒ sup ℝ and inf ℝ does not exist.
⇒ The set ℝ is not bdd above and not bdd below ⇒ ℝ is not a bdd set.

(4) sup [−2,1) = 1 and inf [−2,1) = −2

sup (−2,1] = 1 and inf (−2,1] = −2
sup [−2,1] = 1 and inf [−2,1] = −2
1−𝑛 1 2 3 4
(1) Consider the set 𝑆 = {𝑛2 +1 : 𝑛 ∈ ℕ} ⇒ 𝑆 = {0, − 5 , − 10 , − 17 , − 26 , … }
1−𝑛 1−𝑛 1
⇒ sup {𝑛2 +1 : 𝑛 ∈ ℕ} = 0 and inf {𝑛2 +1 : 𝑛 ∈ ℕ} = − 5

Example 2.3.3: Find sup 𝑆 and inf 𝑆 if they exist:

(1) 𝑆 = {𝑥 ∈ ℝ: 𝑥 + 1 > 𝑥 2 }
(2) 𝑆 = {𝑥 ∈ ℝ: 𝑥 + 1 < 𝑥 2 }
Solution:
(1) We have to sketch the graph of 𝑆 on the real line. To do that we must study the
inequality 𝑥 + 1 > 𝑥 2 ⇒ −𝑥 2 + 𝑥 + 1 > 0?
−1±√1+4 1∓√5
−𝑥 2 + 𝑥 + 1 = 0 ⇒ 𝑥 = =
−2 2

1−√5 1+√5
⇒ 𝑆 = {𝑥 ∈ ℝ: 𝑥 + 1 > 𝑥 2 } = ( , )
2 2

1+√5 1−√5
⇒ sup 𝑆 = and inf 𝑆 =
2 2

(2) 𝑆 = {𝑥 ∈ ℝ: 𝑥 + 1 < 𝑥 2 } = {𝑥 ∈ ℝ: −𝑥 2 + 𝑥 + 1 < 0}

1−√5 1+√5
= (−∞, )∪( , ∞)
2 2
⇒ sup 𝑆 = ∞ and inf 𝑆 = −∞ ⇒ sup 𝑆 and inf 𝑆 does not exist
 12

Definition 2.3.4: Let 𝑓: 𝐷 → ℝ be. Then

sup 𝑓(𝑥) = sup{𝑓(𝑥): 𝑥 ∈ 𝐷} = sup (Range 𝑓)
𝑥∈𝐷

and
inf 𝑓(𝑥) = inf{𝑓(𝑥): 𝑥 ∈ 𝐷} = inf (Range 𝑓)
𝑥∈𝐷

Example 2.3.5:
(1) 𝑓(𝑥) = −𝑥 2 + 𝑥 + 1. Find sup 𝑓(𝑥), inf 𝑓(𝑥), sup 𝑓(𝑥) and inf 𝑓(𝑥).
𝑥∈ℝ 𝑥∈ℝ 𝑥∈[1,2) 𝑥∈[1,2)
3 2
(2) Let 𝑓(𝑥) = 𝑥 + 3𝑥 − 2. Find sup 𝑓(𝑥) and inf 𝑓(𝑥).
𝑥∈[−2,0.5) 𝑥∈[−2,0.5)

Solution:
(1) To find Range(𝑓) we give a sketch the graph of 𝑓(𝑥)
5
 When Domain (𝑓) = ℝ ⇒ Range(𝑓) = (−∞, 4]
5 5 5
⇒ sup 𝑓(𝑥) = sup (−∞, 4] = 4 and inf 𝑓(𝑥) = inf (−∞, 4] = −∞
𝑥∈ℝ 𝑥∈ℝ

⇒ inf 𝑓(𝑥) does not exist

𝑥∈ℝ

 When Domain (𝑓) = ℝ ⇒ Range(𝑓) = (−5,1] (why?)

⇒ sup 𝑓(𝑥) = sup(−5,1] = 1 and inf 𝑓(𝑥) = inf(−5,1] = −5
𝑥∈[1,2) 𝑥∈ℝ

(2) Domain (𝑓) = [−2,0.5) ⇒ Range (𝑓) =[−2,2]

⇒ sup 𝑓(𝑥) = 2 and inf 𝑓(𝑥) = −2
𝑥∈[−2,0.5) 𝑥∈[−2,0.5)

Remark 2.3.6: Recall that:

(1) sup 𝑆 = 𝑢∗ iff the following two conditions hold:
𝐶1 : 𝑢∗ is an UB of 𝑆.
𝐶2 : If 𝑢 is an UB of 𝑆, then 𝑢∗ ≤ 𝑢.
(2) inf 𝑆 = 𝑣 ∗ iff the following two conditions hold:
𝐶1 : 𝑣 ∗ is a LB of 𝑆.
𝐶2 : If 𝑣 is a LB of 𝑆, then 𝑣 ∗ ≥ 𝑣.
(3) The contraposition of the proposition 𝑃 ⇒ 𝑄 is ~𝑄 ⇒ ~𝑃, that is [𝑃 ⇒ 𝑄]≡[~𝑄 ⇒ ~𝑃].
 13

Theorem 2.3.7: Let 𝑆 be a set of real numbers.

(1) Let 𝑢∗ be a real number. Then sup 𝑆 = 𝑢∗ iff the following two conditions hold:
𝐶1 : 𝑢∗ is an UB of 𝑆.
𝐶2′ : if 𝑢∗ > 𝑢, then 𝑢 is not an UB of 𝑆.
(2) Let 𝑣 ∗ be a real number. Then inf 𝑆 = 𝑣 ∗ iff the following two conditions hold:
𝐶1 : 𝑣 ∗ is a LB of 𝑆.
𝐶2′ : if 𝑣 ∗ < 𝑣, then 𝑣 is not a LB of 𝑆.
Proof: Observe that 𝐶2′ is the contraposition of 𝐶2 . So, 𝐶2 ≡ 𝐶2′ .

Theorem 2.3.8: Let 𝑆 be a set of real numbers.

(1) Let 𝑢∗ be a real number. Then sup 𝑆 = 𝑢∗ iff the following two conditions hold:
𝐶1 ∶ 𝑢∗ is an UB of 𝑆.
𝐶2′′ : ∀𝜀 > 0, ∃𝑥𝜀 ∈ 𝑆 such that 𝑢∗ − 𝜀 < 𝑥𝜀
(2) Let 𝑣 ∗ be a real number. Then inf 𝑆 = 𝑣 ∗ iff the following two conditions hold:
𝐶1 ∶ 𝑣 ∗ is a LB of 𝑆.
𝐶2′′ : ∀𝜀 > 0, ∃𝑥𝜀 ∈ 𝑆 such that 𝑣 ∗ + 𝜀 > 𝑥𝜀
Proof: We prove part (1) by showing 𝐶2′ ≡ 𝐶2′′ . The proof of part (2) is similar.
Suppose that 𝐶2′ holds ⇒ If 𝑢 < 𝑢∗ , then 𝑢 is not an UB of 𝑆…….①
Given 𝜀 > 0. Let 𝑢𝜀 = 𝑢∗ − 𝜀 ⇒ 𝑢𝜀 < 𝑢∗ ⇒ 𝑢𝜀 is not an UB of 𝑆 (by ①)
⇒ ∃𝑥𝜀 ∈ 𝑆 such that 𝑢𝜀 < 𝑥𝜀 ⇒ 𝑢∗ − 𝜀 < 𝑥𝜀 ⇒ 𝐶2′′ hold. So, 𝐶2′ ⇒ 𝐶2′′ ……②

Suppose that 𝐶2′′ holds ⇒ ∀𝜀 > 0, ∃𝑥𝜀 ∈ 𝑆 such that 𝑢∗ − 𝜀 < 𝑥𝜀 …….③
Let 𝑢 < 𝑢∗ . Take 𝜀 = 𝑢∗ − 𝑢 ⇒ 𝜀 > 0 so by ② ∃𝑥𝜀 ∈ 𝑆 such that 𝑢∗ − 𝜀 < 𝑥𝜀
⇒ 𝑢∗ − (𝑢∗ − 𝑢) < 𝑥𝜀 ⇒ 𝑢 < 𝑥𝜀 ⇒ 𝑢 is not an UB of 𝑆 ⇒ 𝐶2′ holds. So, 𝐶2′′ ⇒ 𝐶2′ ……④
By ③&④ we have 𝐶2′ ⇔ 𝐶2′′ .

Theorem 2.3.9:
(1) If 𝑢0 is an UB of a set 𝑆 and 𝑢0 ∈ 𝑆, then sup 𝑆 = 𝑢0 .
(2) If 𝑣0 is a LB of a set 𝑆 and 𝑣0 ∈ 𝑆, then inf 𝑆 = 𝑣0 .
Proof:
(1) We will show that 𝑢0 satisfies 𝐶1 and 𝐶2 in the definition of supremum.
In the definition taking 𝑢∗ = 𝑢0 we have:
 𝑢0 is an UB of a set 𝑆 ⇒ 𝐶1 holds
 Let 𝑢 be an UB of 𝑆. Want to show that 𝑢 ≥ 𝑢0 :
Since 𝑢 is an UB of 𝑆 ⇒ 𝑢 ≥ 𝑥, ∀𝑥 ∈ 𝑆……①
But 𝑢0 ∈ 𝑆. So, in ② take 𝑥 = 𝑢0 ⇒ 𝑢 ≥ 𝑢0 ⇒ 𝐶2 holds
 14

⇒ 𝑢0 satisfies 𝐶1 and 𝐶2 in the definition of supremum ⇒ sup 𝑆 = 𝑢0 .

(2) Exercise.

The Completeness Property of ℝ 2.3.10:

(1) If 𝑆 is a set of real numbers with UB, then 𝑆 has a supremum.
(2) If 𝑆 is a set of real numbers with LB, then 𝑆 has an infimum.
(3) If 𝑆 is a bdd of real numbers, then 𝑆 has a supremum and infimum.

The Archimedean Property 2.3.11: If 𝑥 ∈ ℝ, then ∃𝑛 ∈ ℕ such that 𝑛 > 𝑥.

Proof: We use the proof by contradiction: Let 𝑥 ∈ ℝ.
Assume that 𝑛 ≤ 𝑥, ∀𝑛 ∈ ℕ ⇒ 𝑥 is an UB of ℕ
The Completeness Property of ℝ ⇒ ℕ has supremum so let sup ℕ = 𝑢∗ with 𝑢∗ ∈ ℝ
⇒ 𝑢∗ satisfies 𝐶2′′ in part (1) of Theorem 8 ⇒ ∀𝜀 > 0, ∃𝑛𝜀 ∈ ℕ such that 𝑢∗ − 𝜀 < 𝑛𝜀
Take 𝜀 = 1 ⇒ , ∃𝑛1 ∈ ℕ such that 𝑢∗ − 1 < 𝑛1 ⇒ 𝑢∗ < 𝑛1 + 1……….①
𝑢∗ satisfies 𝐶2 in the definition of supremum and since 𝑛1 + 1 ∈ ℕ and we have
𝑛1 + 1 ≤ 𝑢∗ ………②
①&②: ⇒ 𝑛1 + 1 ≤ 𝑢∗ < 𝑛1 + 1 ⇒ 𝑛1 + 1 < 𝑛1 + 1 which is a contradiction
⇒ Our assumption is false ⇒ ∃𝑛 ∈ ℕ such that 𝑛 > 𝑥.

Remark 2.3.12:
2
(1) ℚ: The set of rational numbers. For example: 0, −7,9, 5 , √16 are rational numbers.
1
(2) ℚ𝑐 : The set of irrational numbers. For example: √2, √3, 𝜋, are irrational numbers
√2
(3) ℤ: The set of integers ⇒ ℤ ⊂ ℚ
(4) ℝ = ℚ ∪ ℚ𝑐 and ℚ ∩ ℚ𝑐 = ∅ (the empty set)
(5) 𝑎 ∈ ℝ ⇔ 𝑎 ∈ ℚ or 𝑎 ∈ ℚ𝑐
(6) If 𝑥 ∈ ℝ, then ∃𝑚 ∈ ℤ such that 𝑚 − 1 ≤ 𝑥 < 𝑚.

The Density Theorem 2.3.13: Between any two real numbers there is a rational number
If 𝑥, 𝑦 ∈ ℝ such that 𝑥 < 𝑦, then ∃𝑟 ∈ ℚ such that 𝑥 < 𝑟 < 𝑦.
(any interval has a rational number inside it)
1
Proof: 𝑥 < 𝑦 ⇒ 𝑦 − 𝑥 > 0 ⇒ 𝑦−𝑥 > 0
1
The Archimedean Property 2.3.11 ⇒ ∃𝑛 ∈ ℕ such that 𝑛 > 𝑦−𝑥 ⇒ 𝑛𝑦 − 𝑛𝑥 > 1
 15

⇒ 𝑛𝑥 + 1 < 𝑛𝑦..…..①
Since 𝑛𝑥 ∈ ℝ ⇒ ∃𝑚 ∈ ℤ such that 𝑚 − 1 ≤ 𝑛𝑥 < 𝑚 ⇒ 𝑛𝑥 < 𝑚………..②
②: ⇒ 𝑚 − 1 ≤ 𝑛𝑥 ⇒ 𝑚 ≤ 𝑛𝑥 + 1. So, by ① we have 𝑚 < 𝑛𝑦……….③
𝑚
②&③ ⇒ 𝑛𝑥 < 𝑚 < 𝑛𝑦 ⇒ 𝑥 < < 𝑦……④
𝑛
𝑚
So, take 𝑟 = and since 𝑚, 𝑛 ∈ ℤ with 𝑛 ≠ 0 we have 𝑟 ∈ ℚ with 𝑥 < 𝑟 < 𝑦 (by ④)
𝑛

Corollary 2.3.14: Between any two real numbers there is an irrational number
If 𝑥, 𝑦 ∈ ℝ such that 𝑥 < 𝑦, then ∃𝑟 ′ ∈ ℚ𝑐 such that 𝑥 < 𝑟 ′ < 𝑦.
(any interval has an irrational number inside it)
𝑥 𝑦
Proof: 𝑥 < 𝑦 ⇒ <
√2 √2

𝑥 𝑦
Applying the density Theorem on the interval ( , ): ∃𝑟 ∈ ℚ such that
√2 √2
𝑥 𝑦
<𝑟< ⇒ 𝑥 < 𝑟√2 < 𝑦. Let 𝑟 ′ = 𝑟√2 ⇒ 𝑥 < 𝑟 ′ < 𝑦
√2 √2

Since 𝑟 ∈ ℚ and √2 ∈ ℚ𝑐 , then 𝑟√2 ∈ ℚ𝑐 ⇒ 𝑟 ′ ∈ ℚ𝑐 with 𝑥 < 𝑟 ′ < 𝑦

Exercises 2.3.15 (Pages 39-40): 4,5,9,10

 16

Section 2.4: Applications of the Supremum Property

Recall that:
(5) sup 𝑆 = 𝑢∗ if the following two conditions hold:
𝐶1 : 𝑢∗ is an UB of 𝑆.
𝐶2 : if 𝑢 is an UB of 𝑆, then 𝑢∗ ≤ 𝑢.
(6) inf 𝑆 = 𝑣 ∗ if the following two conditions hold:
𝐶1 : 𝑣 ∗ is a LB of 𝑆.
𝐶2 : if 𝑣 is a LB of 𝑆, then 𝑣 ∗ ≥ 𝑣.

Remark:
:‫ يكون البرهان باختيار طرف من أطراف العالقة ثم نقوم بتطبيق الخطوات الخمس التالية‬inf 𝑆 ‫ أو‬sup 𝑆 ‫إلثبات عالقة ما فيها‬
① ② ③ ④ ⑤
⏞𝐶1 Definition ⇒ ⏞
⇒ ⏞ ‫⏞ ⇒ خطوة اإلنتقال‬
Definition ⇒ ⏞𝐶2

Definition 2.4.1: Let 𝐴 be a subset of ℝ. Then the set −𝐴 is defined by −𝐴 = {−𝑎: 𝑎 ∈ 𝐴}

Example 2.4.2:
(1) If 𝐴 = {−2,1,4}, then −𝐴 = {2, −1, −4}
(2) −ℤ = −{… , −3, −2, −1,0,1,2,3, … } = {… ,3,2,1,0, −1, −2, −3, … } = ℤ

Very Important
Theorem 2.4.3: Let 𝐴 be a set of real numbers. Then
(1) sup(−𝐴) = − inf 𝐴, that is inf 𝐴 = − sup(−𝐴).
(2) inf (−𝐴) = − sup 𝐴, that is sup 𝐴 = − inf(−𝐴).
(3)
(4)

Proof:
(1) Let sup(−𝐴) = 𝑢∗ and inf 𝐴 = 𝑣 ∗ .
Want to show that 𝑢∗ = −𝑣 ∗ by showing that 𝑢∗ ≥ −𝑣 ∗ and 𝑢∗ ≤ −𝑣 ∗ :
Step 1: sup(−𝐴) = 𝑢∗ :
𝐶1 on −𝐴: 𝑢∗ is an UB of −𝐴 ⇒ −𝑎 ≤ 𝑢∗ , ∀𝑎 ∈ 𝐴 (by definition)
⇒ 𝑎 ≥ −𝑢∗ , ∀𝑎 ∈ 𝐴
⇒ −𝑢∗ is a LB of 𝐴 (by definition)
⇒ 𝑣 ∗ ≥ −𝑢∗ (by 𝐶2 on 𝐴)
⇒ 𝑢∗ ≥ −𝑣 ∗ ……①
Step 2: inf 𝐴 = 𝑣 ∗ :
𝐶1 on 𝐴: 𝑣 ∗ is a LB of 𝐴 ⇒ 𝑎 ≥ 𝑣 ∗ , ∀𝑎 ∈ 𝐴 (by definition)
⇒ −𝑎 ≤ −𝑣 ∗ , ∀𝑎 ∈ 𝐴
⇒ 𝑥 ≤ −𝑣 ∗ , ∀𝑥 ∈ −𝐴
⇒ −𝑣 ∗ is an UB of −𝐴 (by definition)
⇒ −𝑣 ∗ ≥ 𝑢∗ (by 𝐶2 on −𝐴)
⇒ −𝑣 ∗ ≥ 𝑢∗ ……②
①&②: ⇒ 𝑢∗ = −𝑣 ∗ ⇒ sup(−𝐴) = − inf 𝐴
(2) Let 𝐵 = −𝐴 ⇒ 𝐴 = −𝐵 ⇒ inf(−𝐴) = inf 𝐵 = − sup(−𝐵) = − sup 𝐴 (by part (1))
 17

Definition 2.4.4: Let 𝐴 be a set of real numbers and let 𝑏 ∈ ℝ. The set 𝑏𝐴 is defined by
𝑏𝐴 = {𝑏𝑎: 𝑎 ∈ 𝐴}
Example 2.4.5:
(5) If 𝐴 = {1,2,4}, then 0.1𝐴 = {0.1,0.2,0.4)}
(6) 2ℤ = 2{0, ±1, ±2, ±3, … } = {0, ±2, ±4, ±6, … } = The set of even integers.

Very Important

Theorem 2.4.6: Let 𝐴 be a set of real numbers and let 𝑏 ∈ ℝ. Then

(1) If 𝑏 > 0, then sup(𝑏𝐴) = 𝑏(sup 𝐴)
(2) If 𝑏 > 0, then inf(𝑏𝐴) = 𝑏(inf 𝐴)
(3) If 𝑏 < 0, then sup(𝑏𝐴) = 𝑏(inf 𝐴)
(4) If 𝑏 < 0, then inf(𝑏𝐴) = 𝑏(sup 𝐴)
(7)

Proof:
(1) Let 𝑏 > 0, sup(𝑏𝐴) = 𝑢∗ and sup 𝐴 = 𝑤 ∗ .
Want to show that 𝑢∗ = 𝑏𝑤 ∗ by showing that 𝑢∗ ≥ 𝑏𝑤 ∗ and 𝑢∗ ≤ 𝑏𝑤 ∗ :
Step 1: sup(𝑏𝐴) = 𝑢∗ :
𝐶1 on 𝑏𝐴: 𝑢∗ is an UB of 𝑏𝐴 ⇒ 𝑏𝑎 ≤ 𝑢∗ , ∀𝑎 ∈ 𝐴 (by definition)
𝑢∗
⇒𝑎≤ , ∀𝑎 ∈ 𝐴 (since 𝑏 > 0)
𝑏

𝑢∗
⇒ is an UB of 𝐴 (by definition)
𝑏

𝑢∗
⇒ 𝑤∗ ≤ (by 𝐶2 on 𝐴)
𝑏

⇒ 𝑢∗ ≥ 𝑏𝑤 ∗ ……①

Step 2: sup 𝐴 = 𝑤 ∗:
𝐶1 on 𝐴: 𝑤 ∗ is an UB of 𝐴 ⇒ 𝑎 ≤ 𝑤 ∗ , ∀𝑎 ∈ 𝐴 (by definition)
⇒ 𝑏𝑎 ≤ 𝑏𝑤 ∗ , ∀𝑎 ∈ 𝐴 (since 𝑏 > 0)
⇒ 𝑏𝑤 ∗ is an UB of 𝑏𝐴 (by definition)
⇒ 𝑢∗ ≤ 𝑏𝑤 ∗ (by 𝐶2 on 𝐴)
⇒ 𝑢∗ ≤ 𝑏𝑤 ∗ ……②

①&②: ⇒ 𝑢∗ = 𝑏𝑤 ∗ ⇒ sup(𝑏𝐴) = 𝑏(sup 𝐴).

(2) inf (𝑏𝐴) = −

⏟ sup (−𝑏𝐴) = −𝑏
⏟ sup (−𝐴) = −𝑏 (⏟− inf (𝐴) ) = 𝑏 inf (𝐴)
by Theorem 3 by part (1) by Theorem 3
 18

(3) Let 𝑏 < 0 and 𝑐 = −𝑏 ⇒ 𝑐 > 0. Now,

sup (𝑏𝐴) = sup (−𝑐𝐴) = ⏟

𝑐 sup (−𝐴) = 𝑐 (⏟ − inf 𝐴 ) = 𝑏 inf 𝐴
by part (1) by Theorem 3
(4) Exercise

Definition 2.4.7: Let 𝐴 be a set of real numbers and let 𝑏 ∈ ℝ. Define the set 𝐴 + 𝑏 by
𝐴 + 𝑏 = {𝑎 + 𝑏: 𝑎 ∈ 𝐴}
Example 2.4.8:
(7) If 𝐴 = {1,2,4}, then 𝐴 + 0.5 = {1 + 0.5,2 + 0.5,4 + 0.5} = {1.5,2.5,4.5}
(8) 𝐴 − 1 = 𝐴 + (−1) = {1 − 1,2 − 1,4 − 1} = {0,1,3}
Theorem 2.4.9: Let 𝐴 be a set of real numbers and let 𝑏 ∈ ℝ. Then
(1) sup(𝐴 + 𝑏) = (sup 𝐴) + 𝑏
(2) inf(𝐴 + 𝑏) = (inf 𝐴) + 𝑏
Proof:
(1) Let sup(𝐴 + 𝑏) = 𝑢∗ and sup 𝐴 = 𝑤 ∗
Want to show that 𝑢∗ = 𝑤 ∗ + 𝑏 by showing: 𝑢∗ ≥ 𝑤 ∗ + 𝑏 and 𝑢∗ ≤ 𝑤 ∗ + 𝑏:
Step 1: sup(𝐴 + 𝑏) = 𝑢∗ :
𝐶1 on 𝐴 + 𝑏: 𝑢∗ is an UB of 𝐴 + 𝑏 ⇒ 𝑎 + 𝑏 ≤ 𝑢∗ , ∀𝑎 ∈ 𝐴 (by definition)
⇒ 𝑎 ≤ 𝑢∗ − 𝑏, ∀𝑎 ∈ 𝐴
⇒ 𝑢∗ − 𝑏 is an UB of 𝐴 (by definition)
⇒ 𝑤 ∗ ≤ 𝑢∗ − 𝑏 (by 𝐶2 on 𝐴)
⇒ 𝑤 ∗ + 𝑏 ≤ 𝑢∗ ……①

Step 2: sup 𝐴 = 𝑤 ∗:
𝐶1 on 𝐴: 𝑤 ∗ is an UB of 𝐴 ⇒ 𝑎 ≤ 𝑤 ∗ , ∀𝑎 ∈ 𝐴 (by definition)
⇒ 𝑎 + 𝑏 ≤ 𝑤 ∗ + 𝑏, ∀𝑎 ∈ 𝐴
⇒ 𝑤 ∗ + 𝑏 is an UB of 𝐴 + 𝑏 (by definition)
⇒ 𝑢∗ ≤ 𝑤 ∗ + 𝑏 (by 𝐶2 on 𝐴 + 𝑏)
⇒ 𝑢∗ ≤ 𝑤 ∗ + 𝑏……②
①&②: ⇒ 𝑢∗ = 𝑤 ∗ + 𝑏 ⇒ sup(𝐴 + 𝑏) = (sup 𝐴) + 𝑏.
 19

(2) inf(𝐴 + 𝑏) = ⏟
− sup[−(𝐴 + 𝑏)] = − sup(−𝐴 − 𝑏) = ⏟
−[sup(−𝐴) − 𝑏]
by Theorem 3 by part (1)

= − sup(−𝐴) + 𝑏 = (⏟ inf 𝐴 )+𝑏

by Theorem 3

Example 2.4.10: Find

6𝑛+1 2𝑛2 +𝑛−1
(1) sup { : 𝑛 ∈ ℕ} (2) inf { : 𝑛 ∈ ℕ}
2𝑛 𝑛+1
2 2
(3) sup {5 − 𝑛2 : 𝑛 ∈ ℕ} (4) inf {5 − 𝑛2 : 𝑛 ∈ ℕ}

Solution:
6𝑛+1 1 1
(1) sup { : 𝑛 ∈ ℕ} = sup {3 + 2𝑛 : 𝑛 ∈ ℕ} = 3 + sup {2𝑛 : 𝑛 ∈ ℕ}
2𝑛

1 1 1 7
= 3 + sup { : 𝑛 ∈ ℕ} = 3 + (1) =
2 𝑛 2 2
2𝑛2 +𝑛−1
(2) inf { : 𝑛 ∈ ℕ} = inf{2𝑛 − 1: 𝑛 ∈ ℕ} = inf{2𝑛: 𝑛 ∈ ℕ} − 1
𝑛+1

= 2 inf{𝑛: 𝑛 ∈ ℕ} − 1 = 2 inf{1,2,3, … } − 1 = 2(1) − 1 = 1

2 2 1
(3) sup {5 − 𝑛2 : 𝑛 ∈ ℕ} = 5 + sup {− 𝑛2 : 𝑛 ∈ ℕ} = 5 − 2 inf {𝑛2 : 𝑛 ∈ ℕ} = 5 − 2(0) = 5
2 2 1
(4) inf {5 − 𝑛2 : 𝑛 ∈ ℕ} = 5 + inf {− 𝑛2 : 𝑛 ∈ ℕ} = 5 − 2 sup {𝑛2 : 𝑛 ∈ ℕ} = 5 − 2(1) = 3

Example 2.4.11: Find

(1) sup {√𝑛 + 1 − √𝑛 ∶ 𝑛 ∈ ℕ} (2) inf{√𝑛 + 1 − √𝑛 ∶ 𝑛 ∈ ℕ}

Solution:

:‫الحل التالي خطأ‬

(1) sup {√𝑛 + 1 − √𝑛 ∶ 𝑛 ∈ ℕ} = sup {√𝑛 + 1 ∶ 𝑛 ∈ ℕ} + sup {−√𝑛 ∶ 𝑛 ∈ ℕ}
= sup {√𝑛 + 1 ∶ 𝑛 ∈ ℕ} − inf {√𝑛 ∶ 𝑛 ∈ ℕ}
= sup {√2, √3, √4, … } − inf {1, √2, √3, √4, … }
=∞−1 =∞
(2) inf {√𝑛 + 1 − √𝑛 ∶ 𝑛 ∈ ℕ} = inf {√𝑛 + 1 ∶ 𝑛 ∈ ℕ} + inf {−√𝑛 ∶ 𝑛 ∈ ℕ}
= inf {√𝑛 + 1 ∶ 𝑛 ∈ ℕ} − sup {√𝑛 ∶ 𝑛 ∈ ℕ}
= inf {√2, √3, √4, … } − sup {1, √2, √3, √4, … }
= √2 − ∞ = −∞

1 1 1 1
(1) sup {√𝑛 + 1 − √𝑛 ∶ 𝑛 ∈ ℕ} = sup { ∶ 𝑛 ∈ ℕ} = sup { , , ,…}
√𝑛+1+√𝑛 √2+1 √3+√2 √4+√3
1
= .
√2+1
1 1 1 1
(2) inf {√𝑛 + 1 − √𝑛 ∶ 𝑛 ∈ ℕ} = inf { ∶ 𝑛 ∈ ℕ} = inf { , , ,…}
√𝑛+1+√𝑛 √2+1 √3+√2 √4+√3
= 0.
 20

Definition 2.4.12: Let 𝐴 and 𝐵 be two sets of real numbers. Define the sets 𝐴 + 𝐵 by
𝐴 + 𝐵 = {𝑎 + 𝑏: 𝑎 ∈ 𝐴, 𝑏 ∈ 𝐵}
Example 2.4.13: If 𝐴 = {−1,0} and 𝐵 = {−3,0,1}.
(1) Find 𝐴 + 𝐵 (2) Find 𝐴 + 𝐴 (3) Find 2𝐴 (4) Is 𝐴 + 𝐴 = 2𝐴
Solution:
(1) 𝐴 + 𝐵 = {−1 + (−3), −1 + 0, −1 + 1,0 + (−3), 0 + 0,0 + 1} = {−4, −1,0, −3,0,1}
⇒ 𝐴 + 𝐵 = {−4, −3, −1,0,1}
(2) 𝐴 + 𝐴 = {−1 + (−1), −1 + 0,0 + (−1), 0 + 0} = {−2, −1, −1,0} = {−2, −1,0}
(3) 2𝐴 = {−2,0}
(4) 𝐴 + 𝐴 ≠ 2𝐴
Theorem 2.4.14: Let 𝐴 and 𝐵 be two sets of real numbers. Then
(1) sup(𝐴 + 𝐵) = sup 𝐴 + sup 𝐵
(2) inf(𝐴 + 𝐵) = inf 𝐴 + inf 𝐵
(3) sup(𝐴 + 𝐴) = 2 sup 𝐴
(4) inf(𝐴 + 𝐴) = 2 inf 𝐴
Proof:
(1) Let sup(𝐴 + 𝐵) = 𝑢∗ , sup 𝐴 = 𝑣 ∗ and sup 𝐵 = 𝑤 ∗ . Want to show that
𝑢∗ = 𝑣 ∗ + 𝑤 ∗ by showing that 𝑢∗ ≥ 𝑣 ∗ + 𝑤 ∗ and 𝑢∗ ≤ 𝑣 ∗ + 𝑤 ∗ :
Step 1: sup(𝐴 + 𝐵) = 𝑢∗ :
𝐶1 on 𝐴 + 𝐵: 𝑢∗ is an UB of 𝐴 + 𝐵 ⇒ 𝑎 + 𝑏 ≤ 𝑢∗ , ∀𝑎 ∈ 𝐴, ∀𝑏 ∈ 𝐵 (by definition)
⇒ [𝑎 ≤ 𝑢∗ − 𝑏, ∀𝑎 ∈ 𝐴], ∀𝑏 ∈ 𝐵
⇒ [𝑢∗ − 𝑏 is an UB of 𝐴],∀𝑏 ∈ 𝐵 (by definition)
⇒ [𝑣 ∗ ≤ 𝑢∗ − 𝑏], ∀𝑏 ∈ 𝐵 (by 𝐶2 on 𝐴)
⇒ [𝑏 ≤ 𝑢∗ − 𝑣 ∗ ], ∀𝑏 ∈ 𝐵
⇒ 𝑢∗ − 𝑣 ∗ is an UB of 𝐵 (by definition)
⇒ 𝑤 ∗ ≤ 𝑢∗ − 𝑣 ∗ (by 𝐶2 on 𝐵)
⇒ 𝑤 ∗ + 𝑣 ∗ ≤ 𝑢∗ …….①
Step 2: sup 𝐴 = 𝑣 ∗ and sup 𝐵 = 𝑤 ∗ :
𝐶1 on 𝐴: 𝑣 ∗ is an UB of 𝐴 ⇒ 𝑎 ≤ 𝑣 ∗ , ∀𝑎 ∈ 𝐴……...②
𝐶1 on 𝐵: 𝑤 ∗ is an UB of 𝐵 ⇒ 𝑏 ≤ 𝑤 ∗ , ∀𝑏 ∈ 𝐵….....③
②+③: ⇒ 𝑎 + 𝑏 ≤ 𝑣 ∗ + 𝑤 ∗ , ∀𝑎 ∈ 𝐴, ∀𝑏 ∈ 𝐵
⇒ 𝑣 ∗ + 𝑤 ∗ is an UB of 𝐴 + 𝐵 ⇒ 𝑢∗ ≤ 𝑣 ∗ + 𝑤 ∗ ……..④
①&④: ⇒ 𝑢∗ = 𝑣 ∗ + 𝑤 ∗ ⇒ sup(𝐴 + 𝐵) = sup 𝐴 + sup 𝐵
 21

− sup[−(𝐴 + 𝐵)] = − sup[−𝐴 + (−𝐵)]

(2) inf(𝐴 + 𝐵) = ⏟
by Theorem 3

= − [sup(−𝐴)
⏟ + sup(−𝐵)] = − sup(−𝐴) − sup(−𝐵)
by part (1)
= ⏟ inf 𝐴 + ⏟ inf 𝐵
by Theorem 3 by Theorem 3

Example 2.4.15: Find

(−1)𝑛 5
(1) sup { − 𝑚+1 − 3: 𝑚, 𝑛 ∈ ℕ}
2𝑛

(−1)𝑛 5
(2) inf { − 𝑚+1 − 3: 𝑚, 𝑛 ∈ ℕ}
2𝑛

Solution:
(−1)𝑛 5 (−1)𝑛 −5
(1) sup { − 𝑚+1 − 3: 𝑚, 𝑛 ∈ ℕ} = sup { + 𝑚+1 : 𝑚, 𝑛 ∈ ℕ} − 3
2𝑛 2𝑛

(−1)𝑛 −5
= sup { : 𝑛 ∈ ℕ} + sup { : 𝑚 ∈ ℕ} − 3
2𝑛 𝑚+1

1 (−1)𝑛 1
= sup { : 𝑛 ∈ ℕ} − 5 inf { : 𝑚 ∈ ℕ} − 3
2 𝑛 𝑚+1
1 1 1 1 1 1 1
= sup {−1, , − , , … } − 5 inf { , , , … } − 3
2 2 3 4 2 3 4
1 1
= ( ) − 5(0) − 3
2 2
= −2.75
(−1)𝑛 5 (−1)𝑛 −5
(2) inf { − 𝑚+1 − 3: 𝑚, 𝑛 ∈ ℕ} = inf { + 𝑚+1 : 𝑚, 𝑛 ∈ ℕ} − 3
2𝑛 2𝑛

(−1)𝑛 −5
= inf { : 𝑛 ∈ ℕ} + inf { : 𝑚 ∈ ℕ} − 3
2𝑛 𝑚+1

1 (−1)𝑛 1
= inf { : 𝑛 ∈ ℕ} − 5 sup { : 𝑚 ∈ ℕ} − 3
2 𝑛 𝑚+1
1 1 1 1 1 1 1
= inf {−1, , − , , … } − 5 sup { , , , … } − 3
2 2 3 4 2 3 4
1 1
= (−1) − 5 ( ) − 3
2 2
= −6
Exercises 2.4.16 (Pages 44-45): 1,2,8
 22

Section 2.5: Intervals

Bounded Intervals 2.5.1: There are 4 types of bounded intervals.

If 𝑎, 𝑏 ∈ ℝ and 𝑎 < 𝑏:
 The closed interval [𝑎, 𝑏] = {𝑥 ∈ ℝ: 𝑎 ≤ 𝑥 ≤ 𝑏} ⇒ 𝑎, 𝑏 ∈ [𝑎, 𝑏]
 The open interval (𝑎, 𝑏) = {𝑥 ∈ ℝ: 𝑎 < 𝑥 < 𝑏} ⇒ 𝑎, 𝑏 ∉ (𝑎, 𝑏)
 The half-open (or half-closed) intervals:
 [𝑎, 𝑏) = {𝑥 ∈ ℝ: 𝑎 ≤ 𝑥 < 𝑏} ⇒ 𝑎 ∈ [𝑎, 𝑏) and 𝑏 ∉ [𝑎, 𝑏)
 (𝑎, 𝑏] = {𝑥 ∈ ℝ: 𝑎 < 𝑥 ≤ 𝑏} ⇒ 𝑎 ∉ (𝑎, 𝑏] and 𝑏 ∈ (𝑎, 𝑏]
𝑎 and 𝑏 are the endpoints of these intervals.
The length of these intervals is 𝑏 − 𝑎.
 sup[𝑎, 𝑏] = sup(𝑎, 𝑏) = sup[𝑎, 𝑏) = sup(𝑎, 𝑏] = 𝑏
 inf[𝑎, 𝑏] = inf(𝑎, 𝑏) = inf[𝑎, 𝑏) = inf(𝑎, 𝑏] = 𝑎
If 𝑎 = 𝑏, then:
 [𝑎, 𝑎] = {𝑎}
 (𝑎, 𝑎) = [𝑎, 𝑎) = (𝑎, 𝑎] = ∅ (the empty set)
Unbounded Intervals 2.5.2: There are 5 types of unbounded intervals.
If 𝑎, 𝑏 ∈ ℝ:
 The infinite interval (−∞, ∞) = {𝑥 ∈ ℝ: −∞ < 𝑥 < ∞} = ℝ
 The infinite interval [𝑎, ∞) = {𝑥 ∈ ℝ: 𝑎 ≤ 𝑥 < ∞} ⇒ 𝑎 ∈ [𝑎, ∞)
 The infinite interval (𝑎, ∞) = {𝑥 ∈ ℝ: 𝑎 < 𝑥 < ∞} ⇒ 𝑎 ∉ (𝑎, ∞)
 The infinite interval (−∞, 𝑏] = {𝑥 ∈ ℝ: −∞ < 𝑥 ≤ 𝑏} ⇒ 𝑏 ∈ (−∞, 𝑏]
 The infinite interval (−∞, 𝑏) = {𝑥 ∈ ℝ: −∞ < 𝑥 < 𝑏} ⇒ 𝑏 ∉ (−∞, 𝑏)
The length of these intervals is infinite ∞.
 sup (−∞, ∞) = sup [𝑎, ∞) = sup (𝑎, ∞) = ∞
⇒ sup (−∞, ∞) , sup [𝑎, ∞) and sup (𝑎, ∞) does not exist.
 sup (−∞, 𝑏] = sup (−∞, 𝑏) = 𝑏
 inf (−∞, ∞) = inf (−∞, 𝑏] = inf (−∞, 𝑏) = −∞
⇒ inf (−∞, ∞) , inf (−∞, 𝑏]and inf (−∞, 𝑏) does not exist
 inf [𝑎, ∞) = inf (𝑎, ∞) = 𝑎
Characterization Theorem 2.5.3: If 𝑆 is a of real numbers that contains at least two points and
has the property:
if 𝑥, 𝑦 ∈ 𝑆 and 𝑥 < 𝑦, then [𝑥, 𝑦] ⊆ 𝑆,
then 𝑆 is an interval.
 23

Definition 2.5.4: A sequence {𝐼𝑛 : 𝑛 ∈ ℕ} of intervals is called nested if

𝐼1 ⊆ 𝐼2 ⊆ 𝐼3 ⊆ ⋯ or 𝐼1 ⊇ 𝐼2 ⊇ 𝐼3 ⊇ ⋯
Equivalently, {𝐼𝑛 : 𝑛 ∈ ℕ} is nested if 𝐼𝑛 ⊆ 𝐼𝑛+1 , ∀𝑛 ∈ ℕ or 𝐼𝑛 ⊇ 𝐼𝑛+1 , ∀𝑛 ∈ ℕ.

Theorem 2.5.5: Let {𝐼𝑛 : 𝑛 ∈ ℕ} be a sequence of nested intervals.

(1) If 𝐼1 ⊆ 𝐼2 ⊆ 𝐼3 ⊆ ⋯, then ∩ 𝐼𝑛 = 𝐼1 and ∪ 𝐼𝑛 = 𝐼∞ .
 To find ∪ 𝐼𝑛 = 𝐼∞ : Let 𝐼𝑛 be with end points 𝑎𝑛 , 𝑏𝑛 with 𝑎𝑛 ≤ 𝑏𝑛 , ∀𝑛 ∈ ℕ and let
𝑎 = lim 𝑎𝑛 , 𝑏 = lim 𝑏𝑛 = 𝑏.
𝑛→∞ 𝑛→∞
 If 𝑎 ∈ 𝐼𝑚 and 𝑏 ∈ 𝐼𝑛 for some 𝑚, 𝑛 ∈ ℕ ⇒ ∪ 𝐼𝑛 = 𝐼∞ = [𝑎, 𝑏]
 If 𝑎 ∉ 𝐼𝑚 , ∀𝑚 ∈ ℕ and 𝑏 ∈ 𝐼𝑛 for some 𝑛 ∈ ℕ ⇒ ∪ 𝐼𝑛 = 𝐼∞ = (𝑎, 𝑏]
 If 𝑎 ∈ 𝐼𝑚 for some 𝑚 ∈ ℕ and 𝑏 ∉ 𝐼𝑛 , ∀𝑛 ∈ ℕ ⇒ ∪ 𝐼𝑛 = 𝐼∞ = [𝑎, 𝑏)
 If 𝑎, 𝑏 ∉ 𝐼𝑛 , , ∀𝑛 ∈ ℕ ⇒ ∪ 𝐼𝑛 = 𝐼∞ = (𝑎, 𝑏)
(2)

(2) If 𝐼1 ⊇ 𝐼2 ⊇ 𝐼3 ⊇ ⋯, then ∩ 𝐼𝑛 = 𝐼∞ and ∪ 𝐼𝑛 = 𝐼1 .

To find ∩ 𝐼𝑛 = 𝐼∞ : Let 𝐼𝑛 be with end points 𝑎𝑛 , 𝑏𝑛 with

𝑎𝑛 ≤ 𝑏𝑛 , ∀𝑛 ∈ ℕ and let 𝑎 = lim 𝑎𝑛 , 𝑏 = lim 𝑏𝑛 = 𝑏.
𝑛→∞ 𝑛→∞
 If 𝑎, 𝑏 ∈ 𝐼𝑛 , ∀𝑛 ∈ ℕ ⇒ ∩ 𝐼𝑛 = 𝐼∞ = [𝑎, 𝑏]
 If 𝑎 ∉ 𝐼𝑚 for some 𝑚 ∈ ℕ and 𝑏 ∈ 𝐼𝑛 , ∀𝑛 ∈ ℕ ⇒ ∩ 𝐼𝑛 = 𝐼∞ = (𝑎, 𝑏]
 If 𝑎 ∈ 𝐼𝑛 , ∀𝑛 ∈ ℕ and 𝑏 ∉ 𝐼𝑚 for some 𝑚 ∈ ℕ ⇒ ∩ 𝐼𝑛 = 𝐼∞ = [𝑎, 𝑏)
 If 𝑎 ∉ 𝐼𝑚 and 𝑏 ∉ 𝐼𝑛 for some 𝑚, 𝑛 ∈ ℕ ⇒ ∩ 𝐼𝑛 = 𝐼∞ = (𝑎, 𝑏)

1 1
Example 2.5.6: Let 𝐼𝑛 = [−1 + 𝑛 , 1 − 𝑛], 𝑛 ∈ ℕ. Find ∪ 𝐼𝑛 and ∩ 𝐼𝑛 if they exist.
1 1 2 2 3 3
Solution: 𝐼1 = [0,0] = {0}, 𝐼2 = [− 2 , 2] , 𝐼3 = [− 3 , 3] , 𝐼4 = [− 4 , 4] , …

⇒ 𝐼1 ⊆ 𝐼2 ⊆ 𝐼3 ⊆ ⋯ ⇒ {𝐼𝑛 : 𝑛 ∈ ℕ} is a sequence of nested intervals

 ∩ 𝐼𝑛 = 𝐼1 = {0}
 ∪ 𝐼𝑛 = 𝐼∞ :
1
 𝑎 = lim (−1 + 𝑛) = −1 ∉ 𝐼𝑛 , ∀𝑛 ∈ ℕ ⇒ 𝑎 = −1 (‫)يستثنى‬
𝑛→∞
1
 𝑏 = lim (1 − 𝑛) = 1 ∉ 𝐼𝑛 , ∀𝑛 ∈ ℕ ⇒ 𝑏 = 1 (‫)يستثنى‬
𝑛→∞

⇒ ∪ 𝐼𝑛 = (𝑎, 𝑏) = (−1,1).
 24

1
Example 2.5.7: Let 𝐼𝑛 = (0, 𝑛), 𝑛 ∈ ℕ. Find ∪ 𝐼𝑛 and ∩ 𝐼𝑛 if they exist.
1 1 1
Solution: 𝐼1 = (0,1), 𝐼2 = (0, 2) , 𝐼3 = (0, 3) , 𝐼4 = (0, 4) , …

⇒ 𝐼1 ⊇ 𝐼2 ⊇ 𝐼3 ⊇ ⋯ ⇒ {𝐼𝑛 : 𝑛 ∈ ℕ} is a sequence of nested intervals

 ∪ 𝐼𝑛 = 𝐼1 = (0,1)
 ∩ 𝐼𝑛 = 𝐼∞ :
 𝑎 = lim 0 = 0 ∉ 𝐼1 ⇒ 𝑎 = 0 (‫)يستثنى‬
𝑛→∞
1
 𝑏 = lim = 0 ∉ 𝐼1 , ∀𝑛 ∈ ℕ ⇒ 𝑏 = 0 (‫)يستثنى‬
𝑛→∞ 𝑛

⇒ ∩ 𝐼𝑛 = (𝑎, 𝑏) = (0,0) = ∅ (the empty set)

1 1
Example 2.5.8: Let 𝐼𝑛 = (− 𝑛 , 𝑛], 𝑛 ∈ ℕ. Find ∪ 𝐼𝑛 and ∩ 𝐼𝑛 if they exist.
1 1 1 1 1 1
Solution: 𝐼1 = (−1,1], 𝐼2 = (− 2 , 2] , 𝐼3 = (− 3 , 3] , 𝐼4 = (− 4 , 4] , …

⇒ 𝐼1 ⊇ 𝐼2 ⊇ 𝐼3 ⊇ ⋯ ⇒ {𝐼𝑛 : 𝑛 ∈ ℕ} is a sequence of nested intervals

 ∪ 𝐼𝑛 = 𝐼1 = (−1,1]
 ∩ 𝐼𝑛 = 𝐼∞ :
1
 𝑎 = lim − 𝑛 = 0 ∈ 𝐼𝑛 , ∀𝑛 ∈ ℕ ⇒ 𝑎 = 0 (‫)ال يستثنى‬
𝑛→∞
1
 𝑏 = lim = 0 ∈ 𝐼𝑛 , ∀𝑛 ∈ ℕ ⇒ 𝑏 = 0 (‫)ال يستثنى‬
𝑛→∞ 𝑛

⇒ ∩ 𝐼𝑛 = 𝐼∞ = [𝑎, 𝑏] = [0,0] = {0}

1
Example 2.5.9: Let 𝐼𝑛 = [0,1 − 𝑛], 𝑛 ∈ ℕ. Find ∪ 𝐼𝑛 and ∩ 𝐼𝑛 if they exist.
1 2 3
Solution: 𝐼1 = [0,0] = {0}, 𝐼2 = [0, ] , 𝐼3 = [0, ] , 𝐼4 = [0, ] , …
2 3 4

⇒ 𝐼1 ⊆ 𝐼2 ⊆ 𝐼3 ⊆ ⋯ ⇒ {𝐼𝑛 : 𝑛 ∈ ℕ} is a sequence of nested intervals

 ∩ 𝐼𝑛 = 𝐼1 = {0}
 ∪ 𝐼𝑛 = 𝐼∞ :
 𝑎 = lim 0 = 0 ∈ 𝐼1 ⇒ 𝑎 = 0 (‫)ال يستثنى‬
𝑛→∞
1
 𝑏 = lim 1 − 𝑛 = 1 ∉ 𝐼𝑛 , ∀𝑛 ∈ ℕ ⇒ 𝑏 = 1 (‫)يستثنى‬
𝑛→∞

⇒ ∪ 𝐼𝑛 = [0,1)
 25

Example 2.5.10: Let 𝐼𝑛 = [𝑛, ∞), 𝑛 ∈ ℕ. Find ∪ 𝐼𝑛 and ∩ 𝐼𝑛 if they exist.

Solution: 𝐼1 = [1, ∞), 𝐼2 = [2, ∞), 𝐼3 = [3, ∞), 𝐼4 = [4, ∞), …
⇒ 𝐼1 ⊇ 𝐼2 ⊇ 𝐼3 ⊇ ⋯ ⇒ {𝐼𝑛 : 𝑛 ∈ ℕ} is a sequence of nested intervals
 ∪ 𝐼𝑛 = 𝐼1 = [1, ∞)
 ∩ 𝐼𝑛 = 𝐼∞ = (∞, ∞) = ∅ (the empty set)

1 1
Example 2.5.11: Let 𝐼𝑛 = (− 𝑛 , 1 + 𝑛), 𝑛 ∈ ℕ. Find ∪ 𝐼𝑛 and ∪ 𝐼𝑛 if they exist.
1 3 1 4 1 5
Solution: 𝐼1 = (−1,2), 𝐼2 = (− 2 , 2) , 𝐼3 = (− 3 , 3) , 𝐼4 = (− 4 , 4) , …

⇒ 𝐼1 ⊇ 𝐼2 ⊇ 𝐼3 ⊇ ⋯ ⇒ {𝐼𝑛 : 𝑛 ∈ ℕ} is a sequence of nested intervals

 ∪ 𝐼𝑛 = 𝐼1 = (−1,2)
 ∩ 𝐼𝑛 = 𝐼∞ = [0,1]

Exercises 2.5.13 (Pages 52): 7,8,9

 26

Chapter 3: Sequences and Series

Section 3.1: Sequences and their Limits

Definition 3.1.1: A sequence of real numbers (or a sequence in ℝ) is a function 𝑋: ℕ → ℝ and this
function is denote by the notations 𝑋, (𝑥𝑛 ), or (𝑥𝑛 : 𝑛 ∈ ℕ).
It is convenient to list the terms of a sequence as 𝑋 = (𝑥𝑛 : 𝑛 ∈ ℕ) = (𝑥1 , 𝑥2 , 𝑥3 , … )

Example 3.1.2: Plot the graph of the sequence 𝑋 = (2.5, −1,0, 4, 3, −1, 1, 1, 0.5, … ).
Solution: 𝑥1 = 2.5, 𝑥2 = −1, 𝑥3 = 0, 𝑥4 = 4, 𝑥5 = 3, 𝑥6 = −1, 𝑥7 = 1, 𝑥8 = 1, 𝑥9 = 0.5 , …

y
4.5
4
3.5
3
2.5
2
1.5
1
0.5 x

-0.5 1 2 3 4 5 6 7 8 9
-1
-1.5

Remark 3.1.3:
(1) Sequences may be denoted by other letters, such as 𝑌 = (𝑦𝑘 ), 𝑍 = (𝑧𝑗 ) and so on.
(2) Sequences are often defined by giving a formula for the 𝑛𝑡ℎ term 𝑥𝑛 . For example:
1 1 1 1 1
𝑋 = ( , , , , … ) = ( : 𝑛 ∈ ℕ)
2 4 6 8 2𝑛
1
or more simply 𝑋 = (2𝑛).
𝑛 1 2 3
(3) For example: If (𝑥𝑛 ) = (𝑛+1) ⇒ (𝑥𝑛 ) = (2 , 3 , 4 , … )

Definition 3.1.4: If 𝑏 ∈ ℝ, the sequence 𝐵 = (𝑏, 𝑏, 𝑏, … ), all of whose terms equal 𝑏, is called the
constant sequence 𝑏.
 The constant sequence 1 is the sequence (1,1,1, … )
 The constant sequence 0 is the sequence (0,0,0, … ).
 27

Sequences that are defied inductively (or recursively) 3.1.5:

A sequence (𝑥𝑛 ) is defined inductively (or recursively) if the first terms 𝑥1 , 𝑥2 , … , 𝑥𝑘 are
specified and a formula is given for obtaining 𝑥𝑛+1 (𝑛 ≥ 𝑘) in terms of the preceding terms.

Example 3.1.6: Let (𝑦𝑛 ) be the sequence defined inductively as:

𝑦𝑛+1 +𝑦𝑛
𝑦1 = 2, 𝑦2 = 1 and 𝑦𝑛+2 = ,𝑛 ≥ 1
2

Find the first five terms of this sequence.

Solution:
𝑦2 +𝑦1 1+2
𝑦𝑛+2| = = = 1.5 ⇒ 𝑦3 = 1.5
2 2
𝑛=1
𝑦3 +𝑦2 1.5+1
𝑦𝑛+2 | = = = 1.25 ⇒ 𝑦4 = 1.25
2 2
𝑛=2
𝑦4 +𝑦3 1.25+1.5
𝑦𝑛+2 | == = = 1.375 ⇒ 𝑦4 = 1.375
2 2
𝑛=3
⇒ The first 5 terms are: 2, 1, 1.5, 1.25, 1.375
Definition 3.1.7 (Fibonacci sequence (‫))متتالية فيبوناشي‬
The Fibonacci sequence (𝑓𝑛 ) is given inductively by:
𝑓1 = 1, 𝑓2 = 1 and 𝑓𝑛+1 = 𝑓𝑛 + 𝑓𝑛−1 , 𝑛 ≥ 2.

⇒ 𝑓𝑛+1 | = 𝑓2 + 𝑓1 = 1 + 1 = 2 ⇒ 𝑓3 = 2
𝑛=2

⇒ 𝑓𝑛+1 | = 𝑓3 + 𝑓2 = 2 + 1 = 3 ⇒ 𝑓4 = 3
𝑛=3

⇒ 𝑓5 = 𝑓4 + 𝑓3 = 3 + 2 = 5, 𝑓6 = 𝑓5 + 𝑓4 = 5 + 3 = 8, …
⇒ (𝑓𝑛 ) = (1, 1, 2, 3, 5, 8, 13, 21, 34, 55, … )
Question 3.1.8: What is 𝑓13 in the Fibonacci sequence (𝑓𝑛 )?
① ②

Definition 3.1.9: A sequence 𝑋 = (𝑥𝑛 ) in ℝ is said to converge to 𝑥 ∈ ℝ if ⏞

∀𝜀 > 0, ⏞
∃𝐾𝜀 ∈ ℕ such
③

that ⏞
|𝑥𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾𝜀 .

Remark 3.1.10:
 If (𝑥𝑛 ) converges to 𝑥 we say that 𝑥 is a limit of (𝑥𝑛 ) and we write lim 𝑥𝑛 = 𝑥
𝑛→∞

 If a sequence has no limit, that is lim 𝑥𝑛 does not exist, we say that the sequence (𝑥𝑛 )
𝑛→∞

is divergent.
 28

Remark 3.1.11:
(1) The notation 𝐾𝜀 is used to emphasize that the choice of 𝐾 depends on the value of 𝜀.
However, it is often convenient to write 𝐾 instead of 𝐾𝜀 .

(2) In most cases, a ‘‘small’’ value of 𝜀 will usually require a ‘‘large’’ value of 𝐾 to guarantee
that |𝑥𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾.

Recall that the 𝛿-neighborhood of 𝑎 is 𝑉𝛿 (𝑎) = (𝑎 − 𝛿, 𝑎 + 𝛿)

Theorem 3.1.12: Let (𝑥𝑛 ) be a sequence of real numbers, and let 𝑥 ∈ ℝ. The following statements
are equivalent.
(a) (𝑥𝑛 ) converges to 𝑥.
(b) ∀𝜀 > 0, ∃𝐾 ∈ ℕ such that |𝑥𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾.
(c) ∀𝜀 > 0, ∃𝐾 ∈ ℕ such that 𝑥 − 𝜀 < 𝑥𝑛 < 𝑥 + 𝜀, ∀𝑛 ≥ 𝐾.
(d) ∀𝜀, ∃𝐾 ∈ ℕ such that 𝑥𝑛 ∈ (𝑥 − 𝜀, 𝑥 + 𝜀), ∀𝑛 ≥ 𝐾.
(e) ∀𝜀, ∃𝐾 ∈ ℕ such that 𝑥𝑛 ∈ 𝑉𝜀 (𝑥), ∀𝑛 ≥ 𝐾.
Proof:
[(𝑥𝑛 ) converges to 𝑥] iff [∀𝜀 > 0, ∃𝐾 ∈ ℕ such that |𝑥𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾]
iff [∀𝜀 > 0, ∃𝐾 ∈ ℕ such that −𝜀 < 𝑥𝑛 − 𝑥 < 𝜀, ∀𝑛 ≥ 𝐾]
iff [∀𝜀 > 0, ∃𝐾 ∈ ℕ such that 𝑥 − 𝜀 < 𝑥𝑛 < 𝑥 + 𝜀, ∀𝑛 ≥ 𝐾]
iff [∀𝜀 > 0, ∃𝐾 ∈ ℕ such that 𝑥𝑛 ∈ (𝑥 − 𝜀, 𝑥 + 𝜀), ∀𝑛 ≥ 𝐾]
iff [∀𝜀 > 0, ∃𝐾 ∈ ℕ such that 𝑥𝑛 ∈ 𝑉𝜀 (𝑥), ∀𝑛 ≥ 𝐾].

 Recall that if 0 ≤ 𝑎 < 𝜀, ∀𝜀 > 0, then 𝑎 = 0

 Recall that if |𝑎| < 𝜀, ∀𝜀 > 0, then 𝑎 = 0

Theorem Uniqueness of Limit 3.1.13: A sequence in ℝ can have at most one limit.

Proof: Suppose that 𝑥 ′ and 𝑥 ′′ are both limits of (𝑥𝑛 ). Want to show that 𝑥 ′ = 𝑥 ′′ .
Let 𝜀 > 0 be given.
𝜀
Since lim 𝑥𝑛 = 𝑥 ′ ⇒ ∃𝐾 ′ ∈ ℕ such that |𝑥𝑛 − 𝑥 ′ | < 2 , ∀𝑛 ≥ 𝐾 ′ ……..①
𝑛→∞
𝜀
Since lim 𝑥𝑛 = 𝑥 ′′ ⇒ ∃𝐾 ′′ ∈ ℕ such that |𝑥𝑛 − 𝑥 ′′ | < , ∀𝑛 ≥ 𝐾 ′′ ……..②
2

Let 𝐾 = max(𝐾 ′ , 𝐾 ′′ ). If 𝑛 ≥ 𝐾 ⇒ 𝑛 ≥ 𝐾 ′ and 𝑛 ≥ 𝐾 ′′

𝜀 𝜀
⇒ |𝑥𝑛 − 𝑥 ′ | < 2 and |𝑥𝑛 − 𝑥 ′′ | < 2 , ∀𝑛 ≥ 𝐾 by ①&②
 29

Now, |𝑥 ′ − 𝑥 ′′ | = |(𝑥𝑛 − 𝑥 ′′ ) − (𝑥𝑛 − 𝑥 ′ )|

≤ |𝑥𝑛 − 𝑥 ′′ | + |(𝑥𝑛 − 𝑥 ′ | by the triangle inequality
𝜀 𝜀
< + =𝜀
2 2
⇒ |𝑥 ′ − 𝑥 ′′ | < 𝜀, ∀𝜀 > 0 ⇒ 𝑥 ′ − 𝑥 ′′ = 0 ⇒ 𝑥 ′ = 𝑥 ′′ .
Example 3.1.14: Let (𝑥𝑛 ) be the sequence given in the figure.
Find the smallest value of 𝐾 that satisfy the definition of limits in the following cases:
(1) 𝜀 = 1.5 (2) 𝜀 = 1 (3) 𝜀 = 0.5
y
5
4.5
4
3.5
3
2.5
2
1.5
1
0.5
x

-2 -1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
-0.5

Solution: First observe that lim 𝑥𝑛 = 1.5 ⇒ Want to find the smallest 𝐾 ∈ ℕ such that
𝑛→∞

𝑥𝑛 ∈ 𝑉𝜀 (1.5), ∀𝑛 ≥ 𝐾 ⇔ 𝑥𝑛 ∈ (1.5 − 𝜀, 1.5 + 𝜀), ∀𝑛 ≥ 𝐾

(1) 𝜀 = 1.5 ⇒ 𝑥𝑛 ∈ (1.5 − 1.5,1.5 + 1.5), ∀𝑛 ≥ 𝐾 ⇒ 𝑥𝑛 ∈ (0,3), ∀𝑛 ≥ 𝐾 ⇒ 𝐾 = 5
(2) 𝜀 = 1 ⇒ 𝑥𝑛 ∈ (1.5 − 1,1.5 + 1), ∀𝑛 ≥ 𝐾 ⇒ 𝑥𝑛 ∈ (0.5,2.5), ∀𝑛 ≥ 𝐾 ⇒⇒ 𝐾 = 9
(3) 𝜀 = 0.5 ⇒ 𝑥𝑛 ∈ (1.5 − 0.5,1.5 + 0.5), ∀𝑛 ≥ 𝐾 ⇒ 𝑥𝑛 ∈ (1,2), ∀𝑛 ≥ 𝐾 ⇒ ⇒ 𝐾 = 15
3𝑛+5
Example 3.1.15: Using definition, show that lim =3
𝑛→∞ 𝑛+1
3𝑛+5
Solution: Let 𝜀 > 0 be given. Take 𝑥𝑛 = and 𝑥 = 3.
𝑛+1
?
⏞ ∈ ℕ such that |𝑥𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾.
Want to find 𝐾
3𝑛+5 3𝑛+5 3(𝑛+1) 2 2 2 2
|𝑥𝑛 − 𝑥| = | − 3| = | 𝑛+1 − | = |𝑛+1| = 𝑛+1 ≤ 𝑛 ⇒ |𝑥𝑛 − 𝑥| ≤ 𝑛…….①
𝑛+1 𝑛+1
2 2 2
Let < 𝜀 ⇔ 𝜀 < 𝑛 ⇔ 𝑛 > 𝜀 ……….②
𝑛
2
By the Archimedean Property ∃𝐾 ∈ ℕ such that 𝐾 > 𝜀 .
2 2
If 𝑛 ≥ 𝐾 ⇒ 𝑛 > 𝜀 . So, by ② we have < 𝜀, ∀𝑛 ≥ 𝐾.
𝑛

Now, ① ⇒ |𝑥𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾

3𝑛+5
⇒ lim =3
𝑛→∞ 𝑛+1
 30

3𝑛3 −𝑛2 +𝑛+5

Example 3.1.16: Using definition, show that lim =0
𝑛→∞ 𝑛5 −𝑛+3

3𝑛3 −𝑛2 +𝑛+5

Solution: Let 𝜀 > 0 be given. Take 𝑥𝑛 = and 𝑥 = 0.
𝑛5 −𝑛+3
?
⏞ ∈ ℕ such that |𝑥𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾.
Want to find 𝐾
3𝑛3 −𝑛2 +𝑛+5 3𝑛3 −𝑛2 +𝑛+5 3𝑛3 +𝑛+5 3𝑛3 +𝑛3 +5𝑛3
|𝑥𝑛 − 𝑥| = | − 0| = ≤ ≤
𝑛5 −𝑛+3 𝑛5 −𝑛+3 𝑛5 −𝑛+3 𝑛5 −𝑛+3
9𝑛3 9𝑛3 9𝑛3 9𝑛3 18 18
≤ 𝑛5 −𝑛+3 ≤ 𝑛5 −𝑛 ≤ 1 = 1 5 = 𝑛2 ≤
𝑛5 − 𝑛5 𝑛 𝑛
2 2

18
⇒ |𝑥𝑛 − 𝑥| ≤ ……①
𝑛
18 18 18
Let <𝜀 ⇔ <𝑛 ⇔𝑛> ……….②
𝑛 𝜀 𝜀
18
By the Archimedean Property ∃𝐾 ∈ ℕ such that 𝐾 > .
𝜀
18 18
If 𝑛 ≥ 𝐾 ⇒ 𝑛 > . So, by ② we have < 𝜀, ∀𝑛 ≥ 𝐾.
𝜀 𝑛

Now, ① ⇒ |𝑥𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾

3𝑛3 −𝑛2 +𝑛+5
⇒ lim =0
𝑛→∞ 𝑛5 −𝑛+3

Example 3.1.17: Using definition, show that lim (√𝑛 + 1 − √𝑛) = 0

𝑛→∞

Solution:
Let 𝜀 > 0 be given. Take 𝑥𝑛 = √𝑛 + 1 − √𝑛 and 𝑥 = 0.
?
⏞ ∈ ℕ such that |𝑥𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾.
Want to find 𝐾
√𝑛+1+√𝑛
|𝑥𝑛 − 𝑥| = |√𝑛 + 1 − √𝑛 − 0| = √𝑛 + 1 − √𝑛 = (√𝑛 + 1 − √𝑛) ×
√𝑛+1+√𝑛
1 1
= < …………①
√𝑛+1+√𝑛 √𝑛
1 1 1
Let < 𝜀 ⇔ √𝑛 > 𝜀 ⇔ 𝑛 > 𝜀2…….②
√𝑛
1
By the Archimedean Property ∃𝐾 ∈ ℕ such that 𝐾 > 𝜀2 .
1 1
If 𝑛 ≥ 𝐾 ⇒ 𝑛 > 𝜀2 and by ② we have < 𝜀, ∀𝑛 ≥ 𝐾
√𝑛

Now, ① ⇒ |𝑥𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾

⇒ lim (√𝑛 + 1 − √𝑛) = 0
𝑛→∞
 31

Example 3.1.18: Using definition show that if 0 < 𝑏 < 1 lim 𝑏 𝑛 = 0

𝑛→∞
𝑛
Solution: Let 𝜀 > 0 be given. Take 𝑥𝑛 = 𝑏 and 𝑥 = 0.
?
⏞ ∈ ℕ such that |𝑥𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾.
Want to find 𝐾
ln(ε)
|𝑥𝑛 − 𝑥| < 𝜀 ⇔ 𝑏 𝑛 < 𝜀 ⇔ ln(𝑏 𝑛 ) < ln(𝜀) ⇔ 𝑛 ln(𝑏)
⏟ < ln(ε) ⇔ 𝑛 > ln(𝑏) ………….①
‫سالب‬

ln(ε)
By the Archimedean Property take 𝐾 ∈ ℕ such that 𝐾 > ln(𝑏).
ln(ε)
If 𝑛 ≥ 𝐾 ⇒ 𝑛 > ln(𝑏) , ∀𝑛 ≥ 𝐾 and by ① we have |𝑥𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾
1
Example 3.1.19: Using definition, show that lim =0
𝑛→∞ 𝑛
?
1
⏞ ∈ ℕ such that
Solution: Let 𝜀 > 0 be given. Take 𝑥𝑛 = 𝑛 and 𝑥 = 0. Want to find 𝐾
|𝑥𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾.
1 1 1
|𝑥𝑛 − 𝑥| < 𝜀 ⇔ | − 0| < 𝜀 ⇔ < 𝜀 ⇔ 𝑛 > 𝜀 ……….①
𝑛 𝑛
1 1
By the Archimedean Property take 𝐾 ∈ ℕ such that 𝐾 > 𝜀 : If 𝑛 ≥ 𝐾 ⇒ 𝑛 > 𝜀 , ∀𝑛 ≥ 𝐾
1
⇒ ① holds ⇒ |𝑥𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾 ⇒ lim =0
𝑛→∞ 𝑛

Recall that: If |𝑎| < 𝜀, ∀𝜀 > 0, then 𝑎 = 0

Example 3.1.20: Using definition, show that the sequence ((−1)𝑛 ) = (−1,1, −1,1, … ) diverges.
Solution: The proof by contradiction.
Assume that the sequence is convergent to a real number 𝑥 ⇒ lim 𝑥𝑛 = 𝑥.
𝑛→∞

Observe that 𝑥1 = −1, 𝑥3 = −1, 𝑥5 = −1, …. And 𝑥2 = 1, 𝑥4 = 1, 𝑥6 = 1, …

−1 , 𝑛 is odd
⇒ 𝑥𝑛 = { }
1 , 𝑛 is even
Take 𝜀 > 0 be given ⇒ ∃𝐾 ∈ ℕ such that
|−1 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾 , 𝑛 is odd |1 + 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾 , 𝑛 is odd
|𝑥𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾 ⇔ { }={ }
|1 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾 , 𝑛 is even |1 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾 , 𝑛 is even
⇒ |1 + 𝑥| < 𝜀, ∀𝜀 > 0 and |1 − 𝑥| < 𝜀, ∀𝜀 > 0 ⇒ 1 + 𝑥 = 0 and 1 − 𝑥 = 0
⇒ 𝑥 = −1 and 𝑥 = 1 ⇒ lim 𝑥𝑛 = −1 and lim 𝑥𝑛 = 1
𝑛→∞ 𝑛→∞

⇒ The sequence has two different limits which is a contradiction (since the limit is unique)
⇒ Our assumption is false ⇒ The sequence diverges.
 32

Example 3.1.21: Prove that lim 𝑥𝑛 = 0 if and only if lim |𝑥𝑛 | = 0.

𝑛→∞ 𝑛→∞

Solution:
lim 𝑥𝑛 = 0 ⇔ ∀𝜀 > 0, ∃𝐾 ∈ ℕ such that |𝑥𝑛 − 0| < 𝜀, ∀𝑛 ≥ 𝐾
𝑛→∞

⇔ ∀𝜀 > 0, ∃𝐾 ∈ ℕ such that |𝑥𝑛 | < 𝜀, ∀𝑛 ≥ 𝐾

⇔ ∀𝜀 > 0, ∃𝐾 ∈ ℕ such that | |𝑥𝑛 | | < 𝜀, ∀𝑛 ≥ 𝐾
⇔ ∀𝜀 > 0, ∃𝐾 ∈ ℕ such that | |𝑥𝑛 | − 0 | < 𝜀, ∀𝑛 ≥ 𝐾
⇔ lim |𝑥𝑛 | = 0
𝑛→∞

Example 3.1.22: Let 𝑥 ≠ 0. Show that lim |𝑥𝑛 | = |𝑥| ⇏ lim 𝑥𝑛 = 𝑥.

𝑛→∞ 𝑛→∞

Solution: Let 𝑥𝑛 = (−1)𝑛

⇒ (|𝑥𝑛 |) = (1,1,1, … ) is the constant sequence 1 ⇒ lim |𝑥𝑛 | = 1 = |1| ⇒ lim |𝑥𝑛 | = |1|
𝑛→∞ 𝑛→∞

Also, lim 𝑥𝑛 ≠ 1 because lim 𝑥𝑛 does not exist (by Example 3.1.20 (𝑥𝑛 ) = ((−1)𝑛 ) diverges)
𝑛→∞ 𝑛→∞

The Tail of a Sequence

Definition 3.1.23: If 𝑋 = (𝑥1 , 𝑥2 , … , 𝑥𝑛 , … ) is a sequence of real numbers and if 𝑚 is a given

natural number, then the 𝒎-tail of 𝑋 is the sequence:
𝑋𝑚 = (𝑥𝑚+𝑛 : 𝑛 ∈ ℕ) = (𝑥𝑚+1 , 𝑥𝑚+2 , … ) = (𝑥𝑚+𝑛 )

The terms 𝑥1 , 𝑥2 , … , 𝑥𝑚 were dropped from the sequence 𝑋 to get the 𝑚-tail 𝑋𝑚 .

The 1𝑠𝑡 term of 𝑋𝑚 is the 𝑥𝑚+1

The 2𝑛𝑑 term of 𝑋𝑚 is the 𝑥𝑚+2

Example 3.1.24: The 3-tail of the sequence 𝑋 = (2, 4, 6, 8, … , 2𝑛, … ) is the sequence:
𝑋3 = (𝑥4 , 𝑥5 , … ) = (8, 10, 12, … )

Lemma 3.1.25: Let 𝑋 = (𝑥𝑛 : 𝑛 ∈ ℕ) be a sequence of real numbers and let 𝑚 ∈ ℕ. Then the 𝑚-
tail 𝑋𝑚 = (𝑥𝑚+𝑛 : 𝑛 ∈ ℕ) converges to 𝑥 if and only if 𝑋 converges to 𝑥.
In this case, lim 𝑥𝑛 = lim 𝑥𝑚+𝑛 = 𝑥.
𝑛→∞ 𝑛→∞
 33

Definition 3.1.26: A sequence (𝑥𝑛 ) is called ultimately has a certain property if some tail of (𝑥𝑛 )
has this property.

Example 3.1.27:
(1) The sequence 𝑋 = (3, 4, 5, 5, … , 5, … ) is ultimately constant since the tail
𝑋2 = (5, 5, 5, … . ) is constant.
(2) The sequence 𝑌 = (3, 5, 3, 5, 3, 5, … ) is not ultimately constant since it doesn’t has a
constant tail.
Theorem 3.1.28 (Comparison Theorem of sequences):
Let (𝑥𝑛 ) be a sequence of real numbers and let 𝑥 ∈ ℝ. If (𝑎𝑛 ) is a sequence of positive real numbers
with lim 𝑎𝑛 = 0 and if for some constant 𝐶 > 0 and some 𝐾 ∈ ℕ we have
𝑛→∞

|𝑥𝑛 − 𝑥| ≤ 𝐶𝑎𝑛 , ∀𝑛 ≥ 𝐾, then the sequence (𝑥𝑛 ) with lim 𝑥𝑛 = 𝑥.

𝑛→∞
?
⏞ ∈ ℕ such that |𝑥𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾.
Proof: Let 𝜀 > 0 be given. Want to find 𝐾
𝜀
In the definition of lim 𝑎𝑛 = 0: Take > 0 ⇒ ∃𝐾 ∈ ℕ such that
𝑛→∞ 𝐶
𝜀
|𝑎𝑛 − 0| < , ∀𝑛 ≥ 𝐾
𝜀
𝐶
⇒ 𝑎𝑛 < 𝐶 , ∀𝑛 ≥ 𝐾 … … . . ①
𝜀
Now, |𝑥𝑛 − 𝑥| ≤ 𝐶𝑎𝑛 < 𝐶 (𝐶) = 𝜀, ∀𝑛 ≥ 𝐾 by ①
⇒ |𝑥𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾 ⇒ The sequence (𝑥𝑛 ) converges with lim 𝑥𝑛 = 𝑥.

1
Example 3.1.29: Show that if 𝑎 > 0, then lim =0
𝑛→∞ 1+𝑛𝑎

Solution:
1 1 1 1 1
| − 0| = < = ( ) , ∀𝑛 ≥ 1 … … … ①
1 + 𝑎𝑛 1 + 𝑛𝑎 𝑛𝑎 𝑎 𝑛
1
But lim = 0……….②
𝑛→∞ 𝑛
1
①&②& Comparison Theorem 26 ⇒ lim = 0.
𝑛→∞ 1+𝑛𝑎

Exercises 3.1. (pages 62): 4−13,18

 34

Section 3.2: Limit Theorems

Definition 3.2.1: A sequence (𝑥𝑛 ) of real numbers is said to be bounded if ∃𝑀 > 0 such that
|𝑥𝑛 | ≤ 𝑀, ∀𝑛 ∈ ℕ.

Thus, the sequence (𝑥𝑛 ) is bounded if and only if the set {𝑥𝑛 : 𝑛 ∈ ℕ} is a bounded subset of ℝ
⇒ (𝑥𝑛 ) is bounded if and only if ∃𝑀 > 0 such that −𝑀 ≤ 𝑥𝑛 ≤ 𝑀, ∀𝑛 ∈ ℕ.

Theorem 3.2.2: If a sequence (𝑥𝑛 ) is convergent, then (𝑥𝑛 ) is bounded.

Proof. Suppose that (𝑥𝑛 ) converges and let lim 𝑥𝑛 = 𝑥.
In the definition of lim 𝑥𝑛 = 𝑥 take 𝜀 = 1 ⇒ ∃𝐾 ∈ ℕ such that |𝑥𝑛 − 𝑥| < 1, ∀𝑛 ≥ 𝐾……….①
𝑛→∞
⇒ |𝑥𝑛 | = |𝑥𝑛 − 𝑥 + 𝑥| ≤ |𝑥𝑛 − 𝑥| + |𝑥| ⏟
⏟ < 1 + |𝑥| , ∀𝑛 ≥𝐾
by the triangle inequality by ①
⇒ |𝑥𝑛 | ≤ 1 + |𝑥|, ∀𝑛 ≥ 𝐾…………..②
Let 𝑀 = sup{ |𝑥1 |, |𝑥2 |, … , |𝑥𝐾−1 |, 1 + |𝑥|}.
⇒ |𝑥𝑛 | ≤ 𝑀, ∀𝑛 ∈ ℕ ⇒ (𝑥𝑛 ) is bounded

Remark 3.2.3:
(1) The convers of Theorem 3.2.2 is not true. Consider the following counterexample:
Let 𝑥𝑛 = (−1)𝑛 ⇒ |𝑥𝑛 | ≤ 2, ∀𝑛 ∈ ℕ ⇒ (𝑥𝑛 ) is bounded. But (𝑥𝑛 ) = ((−1)𝑛 ) diverges
(2) The contraposition of Theorem 3.2.2 is the following:
If a sequence (𝑥𝑛 ) is unbounded, then (𝑥𝑛 ) is divergent.
For example:
 The sequence (𝑛) = (1,2,3,4, … ) diverges (since it is unbounded)
 The sequence ((−1)𝑛 𝑛) = (−1,2, −3,4, −5,6, … ) diverges (since it is unbounded).

Theorem 3.2.4:
(1) Let (𝑥𝑛 ) be a sequence that converges to 𝑥, and let 𝑐 ∈ ℝ. Then (𝑐𝑥𝑛 ) converges to 𝑐𝑥, that
is lim (𝑐𝑥𝑛 ) = 𝑐 lim 𝑥𝑛
𝑛→∞ 𝑛→∞
(2) Let (𝑥𝑛 ) and (𝑦𝑛 ) be sequences that converge to 𝑥 and 𝑦, respectively, and let 𝑐 ∈ ℝ. Then
a) (𝑥𝑛 + 𝑦𝑛 ) converges to 𝑥 + 𝑦, that is lim (𝑥𝑛 + 𝑦𝑛 ) = lim 𝑥𝑛 + lim 𝑦𝑛
𝑛→∞ 𝑛→∞ 𝑛→∞
b) (𝑥𝑛 − 𝑦𝑛 ) converges to 𝑥 − 𝑦, that is lim (𝑥𝑛 − 𝑦𝑛 ) = lim 𝑥𝑛 − lim 𝑦𝑛
𝑛→∞ 𝑛→∞ 𝑛→∞
c) (𝑥𝑛 𝑦𝑛 ) converges to 𝑥𝑦, , that is lim 𝑥𝑛 𝑦𝑛 = ( lim 𝑥𝑛 ) ( lim 𝑦𝑛 )
𝑛→∞ 𝑛→∞ 𝑛→∞
(3) If (𝑥𝑛 ) converges to 𝑥 and (𝑦𝑛 ) is a sequence of nonzero real numbers that converges to 𝑦 and
𝑥 𝑥
if 𝑦 ≠ 0, then the quotient sequence (𝑦𝑛) converges to 𝑦, that is
𝑛
𝑥𝑛 𝑛→∞lim 𝑥𝑛
lim =
𝑛→∞ 𝑦𝑛 lim 𝑦𝑛
𝑛→∞
 35

Proof:
(1) Want to show that lim(𝑐𝑥𝑛 ) = 𝑐𝑥 that is ∀𝜀 > 0,∃𝐾 ∈ ℕ such that
|𝑐𝑥𝑛 − 𝑐𝑥| < 𝜀, ∀𝑛 ≥ 𝐾.
Let 𝜀 > 0 be given
Since lim 𝑥𝑛 = 𝑥 ⇒ If 𝜀1 > 0, then ∃𝐾 ∈ ℕ such that |𝑥𝑛 − 𝑥| < 𝜀1 , ∀𝑛 ≥ 𝐾……①
𝑛→∞
𝜀 𝜀
Take 𝜀1 = |𝑐|+1 so ① ⇒ |𝑥𝑛 − 𝑥| < |𝑐|+1 , ∀𝑛 ≥ 𝐾……②
𝜀 |𝑐|
So, |𝑐𝑥𝑛 − 𝑐𝑥| = |𝑐||𝑥𝑛 − 𝑥| < |𝑐| |𝑐|+1 = |𝑐|+1 𝜀 ≤ 𝜀, ∀𝑛 ≥ 𝐾
⇒ |𝑐𝑥𝑛 − 𝑐𝑥| < 𝜀, ∀𝑛 ≥ 𝐾
∴ (𝑐𝑥𝑛 ) converges to 𝑐𝑥 and lim (𝑐𝑥𝑛 ) = 𝑐 lim 𝑥𝑛
𝑛→∞ 𝑛→∞
(2)
(a) Want to show that lim (𝑥𝑛 + 𝑦𝑛 ) = 𝑥 + 𝑦 that is ∀𝜀 > 0,∃𝐾 ∈ ℕ such that
𝑛→∞
|(𝑥𝑛 + 𝑦𝑛 ) − (𝑥 + 𝑦)| < 𝜀, ∀𝑛 ≥ 𝐾.
Let 𝜀 > 0 be given
𝜀 𝜀
In the definition of lim 𝑥𝑛 = 𝑥 take 2 > 0: ∃𝐾1 ∈ ℕ such that |𝑥𝑛 − 𝑥| < 2 , ∀𝑛 ≥ 𝐾1 ……③
𝑛→∞
𝜀 𝜀
In the definition of lim 𝑦𝑛 = 𝑦 take 2 > 0: ∃𝐾2 ∈ ℕ such that |𝑦𝑛 − 𝑦| < 2 , ∀𝑛 ≥ 𝐾2...…④
𝑛→∞
Let 𝐾 = max( 𝐾1 , 𝐾2 ) ⇒ 𝐾 ∈ ℕ
If 𝑛 ≥ 𝐾 ⇒ 𝑛 ≥ 𝐾1 and 𝑛 ≥ 𝐾2
𝜀
From ③: |𝑥𝑛 − 𝑥| < 2 , ∀𝑛 ≥ 𝐾 ………⑤
𝜀
From ④: |𝑦𝑛 − 𝑦| < 2 , ∀𝑛 ≥ 𝐾 ………⑥

𝜀 𝜀
Now, |(𝑥𝑛 + 𝑦𝑛 ) − (𝑥 + 𝑦)| ≤
⏟ |𝑥𝑛 − 𝑥| + |𝑦𝑛 − 𝑦| <
⏟ 2 + 2 = 𝜀, ∀𝑛 ≥ 𝐾
by the triangle inequality
by ⑤&⑥
⇒ |(𝑥𝑛 + 𝑦𝑛 ) − (𝑥 + 𝑦)| < 𝜀, ∀𝑛 ≥ 𝐾
∴ (𝑥𝑛 + 𝑦𝑛 ) converges to 𝑥 + 𝑦, that is lim (𝑥𝑛 + 𝑦𝑛 ) = lim 𝑥𝑛 + lim 𝑦𝑛
𝑛→∞ 𝑛→∞ 𝑛→∞
(b) lim (𝑥𝑛 − 𝑦𝑛 ) = lim (𝑥𝑛 + (−𝑦𝑛 )) = (−𝑦𝑛 ) = lim 𝑥𝑛 − lim 𝑦𝑛
𝑛→∞ 𝑛→∞ ⏟ lim 𝑥𝑛 + lim
𝑛→∞ ⏟ 𝑛→∞ 𝑛→∞ 𝑛→∞
by part (a) by part (1)

(c) Want to show that lim 𝑥𝑛 𝑦𝑛 = 𝑥𝑦 that is ∀𝜀 > 0,∃𝐾 ∈ ℕ such that
𝑛→∞
|𝑥𝑛 𝑦𝑛 − 𝑥𝑦| < 𝜀, ∀𝑛 ≥ 𝐾.
|𝑥𝑛 𝑦𝑛 − 𝑥𝑦| = |𝑥𝑛 𝑦𝑛 − 𝑥𝑦𝑛 + 𝑥𝑦𝑛 − 𝑥𝑦| = |(𝑥𝑛 − 𝑥)𝑦𝑛 + 𝑥(𝑦𝑛 − 𝑦)|
≤ |(𝑥𝑛 − 𝑥)𝑦𝑛 | + |𝑥(𝑦𝑛 − 𝑦)|
= |𝑥𝑛 − 𝑥||𝑦𝑛 | + |𝑥||𝑦𝑛 − 𝑦|………⑦
Let 𝜀 > 0 be given
 (𝑦𝑛 ) converges ⇒ (𝑦𝑛 ) bounded (see Theorem 3.2.2) ⇒ ∃𝑀 > 0 such that
|𝑦𝑛 | ≤ 𝑀, ∀𝑛 ∈ ℕ……….⑧
𝜀
 In the definition of lim 𝑥𝑛 = 𝑥 take 2𝑀 > 0: ∃𝐾1 ∈ ℕ such that
𝑛→∞
𝜀
|𝑥𝑛 − 𝑥| < , ∀𝑛 ≥ 𝐾1 ……⑨
2𝑀
𝜀
 In the definition of lim 𝑦𝑛 = 𝑦 take 2(|𝑥|+1) > 0: ∃𝐾2 ∈ ℕ such that
𝑛→∞
𝜀
|𝑦𝑛 − 𝑦| < , ∀𝑛 ≥ 𝐾2 ...… ⑩
2(|𝑥|+1)
Let 𝐾 = max( 𝐾1 , 𝐾2 ) ⇒ 𝐾 ∈ ℕ
 36

If 𝑛 ≥ 𝐾 ⇒ 𝑛 ≥ 𝐾1 and 𝑛 ≥ 𝐾2, so
𝜀
⑨ ⇒ |𝑥𝑛 − 𝑥| < 2𝑀 , ∀𝑛 ≥ 𝐾……⑪
𝜀
⑩ ⇒ |𝑦𝑛 − 𝑦| < 2(|𝑥|+1) , ∀𝑛 ≥ 𝐾...… ⑫
Now, |𝑥𝑛 𝑦𝑛 − 𝑥𝑦| = |𝑥𝑛 𝑦𝑛 − 𝑥𝑦𝑛 + 𝑥𝑦𝑛 − 𝑥𝑦|
= |(𝑥𝑛 − 𝑥)𝑦𝑛 + 𝑥(𝑦𝑛 − 𝑦)|
≤ |(𝑥𝑛 − 𝑥)𝑦𝑛 | + |𝑥(𝑦𝑛 − 𝑦)| (by the triangle inequality)
= |𝑥𝑛 − 𝑥||𝑦𝑛 | + |𝑥||𝑦𝑛 − 𝑦|
𝜀 𝜀
< 2𝑀 𝑀 + |𝑥| 2(|𝑥|+1) (by ⑧, ⑨&⑩)
𝜀 |𝑥| 𝜀
= 2 + (|𝑥|+1) 2
𝜀 𝜀
≤2+2
= 𝜀, ∀𝑛 ≥ 𝐾
⇒ |𝑥𝑛 𝑦𝑛 − 𝑥𝑦| < 𝜀, ∀𝑛 ≥ 𝐾
∴ (𝑥𝑛 𝑦𝑛 ) converges to 𝑥𝑦, that is lim (𝑥𝑛 𝑦𝑛 ) = ( lim 𝑥𝑛 ) ( lim 𝑦𝑛 )
𝑛→∞ 𝑛→∞ 𝑛→∞

(3) Is an exercise

Theorem 3.2.5: Let (𝑥𝑛 ) is a convergent sequence.

(1) If 𝑥𝑛 ≥ 0, ∀𝑛 ∈ ℕ (or 𝑥𝑛 > 0, ∀𝑛 ∈ ℕ), then lim 𝑥𝑛 ≥ 0.
𝑛→∞
(2) If 𝑥𝑛 ≤ 0, ∀𝑛 ∈ ℕ (or 𝑥𝑛 < 0, ∀𝑛 ∈ ℕ), then lim 𝑥𝑛 ≤ 0.
𝑛→∞

Proof: We prove part (1). The proof of part (2) is an exercise.

Since (𝑥𝑛 ) converges, let lim 𝑥𝑛 = 𝑥.
𝑛→∞
Definition ⇒ If 𝜀 > 0, then ∃𝐾 ∈ ℕ such that |𝑥𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾……①
Want to show that 𝑥 ≥ 0: We use the proof by contradiction:
Assume that 𝑥 < 0 ⇒ −𝑥 > 0
In ①: Take 𝜀 = −𝑥 ⇒ ∃𝐾 ∈ ℕ such that |𝑥𝑛 − 𝑥| < −𝑥, ∀𝑛 ≥ 𝐾
⇒ −(−𝑥) < 𝑥 ⏟𝑛 − 𝑥 < −𝑥 , ∀𝑛 ≥ 𝐾 ⇒ 𝑥𝑛 − 𝑥 < −𝑥, ∀𝑛 ≥ 𝐾
‫ركز على هذا الجزء‬
⇒ 𝑥𝑛 < 0, ∀𝑛 ≥ 𝐾 ⇒ we have a contradiction (since 𝑥𝑛 ≥ 0, ∀𝑛 ∈ ℕ)
Our assumption is false ⇒ 𝑥 ≥ 0 ⇒ lim 𝑥𝑛 ≥ 0
𝑛→∞

Question 3.2.6: Let (𝑥𝑛 ) be convergent. Is 𝑥𝑛 > 0, ∀𝑛 ∈ ℕ ⇒ lim 𝑥𝑛 > 0?

𝑛→∞
Answer: The answer is No. For Example:
1 1
Take 𝑥𝑛 = 𝑛 > 0, ∀𝑛 ∈ ℕ and lim 𝑥𝑛 = lim =0≯0
𝑛→∞ 𝑛→∞ 𝑛
That is 𝑥𝑛 > 0, ∀𝑛 ∈ ℕ ⇏ lim 𝑥𝑛 > 0.
𝑛→∞
 37

Theorem 3.2.7:
(1) If (𝑥𝑛 ) and (𝑦𝑛 ) are convergent sequences and if 𝑥𝑛 ≤ 𝑦𝑛 , ∀𝑛 ∈ ℕ (or 𝑥𝑛 < 𝑦𝑛 , ∀𝑛 ∈ ℕ), then
lim 𝑥𝑛 ≤ lim 𝑦𝑛 .
𝑛→∞ 𝑛→∞

(2) If (𝑥𝑛 ) is a convergent sequence and 𝑎 ≤ 𝑥𝑛 ≤ 𝑏, ∀𝑛 ∈ ℕ (or 𝑎 < 𝑥𝑛 < 𝑏, ∀𝑛 ∈ ℕ), then 𝑎 ≤
lim 𝑥𝑛 ≤ 𝑏.
Proof:
(1) Let 𝑧𝑛 = 𝑦𝑛 − 𝑥𝑛 , 𝑛 ∈ ℕ ⇒ 𝑧𝑛 ≥ 0, ∀𝑛 ∈ ℕ (since 𝑥𝑛 ≤ 𝑦𝑛 , ∀𝑛 ∈ ℕ)
⇒ lim 𝑧𝑛 ≥ 0 ⇒ lim (𝑦𝑛 − 𝑥𝑛 ) ≥ 0 ⇒ lim 𝑦𝑛 − lim 𝑥𝑛 ≥ 0 ⇒ lim 𝑦𝑛 ≥ lim 𝑥𝑛
𝑛→∞ 𝑛→∞ 𝑛→∞ 𝑛→∞ 𝑛→∞ 𝑛→∞

(2) 𝑎 ≤ 𝑥𝑛 ≤ 𝑏, ∀𝑛 ∈ ℕ ⇒ lim 𝑎 ≤ lim 𝑥𝑛 ≤ lim 𝑏, ∀𝑛 ∈ ℕ (by part (1))

𝑛→∞ 𝑛→∞ 𝑛→∞

⇒ 𝑎 ≤ lim 𝑥𝑛 ≤ 𝑏
𝑛→∞

Squeeze Theorem 3.2.8: Let (𝑥𝑛 ), (𝑦𝑛 ), and (𝑧𝑛 ) be sequences of real numbers such that
𝑥𝑛 ≤ 𝑦𝑛 ≤ 𝑧𝑛 , ∀𝑛 ∈ ℕ. If (𝑥𝑛 ) and (𝑧𝑛 ) are convergent with lim 𝑥𝑛 = lim 𝑧𝑛 = 𝑤. Then (𝑦𝑛 ) is
𝑛→∞ 𝑛→∞
convergent and lim 𝑦𝑛 = 𝑤.
𝑛→∞

Proof: Let lim 𝑥𝑛 = lim 𝑧𝑛 = 𝑤

𝑛→∞ 𝑛→∞

Want to show that if 𝜀 > 0,∃𝐾 ∈ ℕ such that |𝑦𝑛 − 𝑤| < 𝜀, ∀𝑛 ≥ 𝐾

Let 𝜀 > 0 be given
By the definition of lim 𝑥𝑛 = 𝑤 ⇒ ∃𝐾1 ∈ ℕ such that |𝑥𝑛 − 𝑤| < 𝜀, ∀𝑛 ≥ 𝐾1 ……①
𝑛→∞

By the definition of lim 𝑧𝑛 = 𝑤 ⇒ ∃𝐾2 ∈ ℕ such that |𝑧𝑛 − 𝑤| < 𝜀, ∀𝑛 ≥ 𝐾2 ……②

𝑛→∞

Let 𝐾 = max(𝐾1 , 𝐾2 ) ⇒ 𝐾 ∈ ℕ.

If 𝑛 ≥ 𝐾 ⇒ 𝑛 ≥ 𝐾1 and 𝑛 ≥ 𝐾2

①&②: ⇒ |𝑥𝑛 − 𝑤| < 𝜀 and |𝑧𝑛 − 𝑤| < 𝜀, ∀𝑛 ≥ 𝐾

⇒ −𝜀 < 𝑥𝑛 − 𝑤 < 𝜀 and −𝜀 < 𝑧𝑛 − 𝑤 < 𝜀, ∀𝑛 ≥ 𝐾
⇒ 𝑤 − 𝜀 < 𝑥𝑛 < 𝑤 + 𝜀 and 𝑤 − 𝜀 < 𝑧𝑛 < 𝑤 + 𝜀, ∀𝑛 ≥ 𝐾
⇒ 𝑤 − 𝜀 < 𝑥𝑛 ≤ 𝑦𝑛 ≤ 𝑧𝑛 < 𝑤 + 𝜀, ∀𝑛 ≥ 𝐾
⇒ 𝑤 − 𝜀 < 𝑦𝑛 < 𝑤 + 𝜀, ∀𝑛 ≥ 𝐾
⇒ −𝜀 < 𝑦𝑛 − 𝑤 < 𝜀, ∀𝑛 ≥ 𝐾 ⇒ |𝑦𝑛 − 𝑤| < 𝜀, ∀𝑛 ≥ 𝐾
⇒ (𝑦𝑛 ) is convergent and lim 𝑦𝑛 = 𝑤
𝑛→∞
 38

Remark 3.2.9:
(1) A sequence (𝑥𝑛 ) converges to 𝑥 iff any tail (𝑥𝑚+𝑛 ) of (𝑥𝑛 ) also converges to 𝑥, that is:
lim 𝑥𝑛 = 𝑥 iff lim 𝑥𝑚+𝑛 = 𝑥, ∀𝑚 ∈ ℕ.
𝑛→∞ 𝑛→∞
(2) Theorems 3.2.5,3.3.7 and 3.2.8 also true for the tails of sequences, that is:
Theorem 3.2.10: Let (𝑥𝑛 ) is a convergent sequence.
(1) If 𝑥𝑛 ≥ 0, ∀𝑛 ≥ 𝐾 (or 𝑥𝑛 > 0, ∀𝑛 ≥ 𝐾), then lim 𝑥𝑛 ≥ 0.
𝑛→∞
(2) If 𝑥𝑛 ≤ 0, ∀𝑛 ≥ 𝐾 (or 𝑥𝑛 < 0, ∀𝑛 ≥ 𝐾), then lim 𝑥𝑛 ≤ 0.
𝑛→∞
Theorem 3.2.11:
(1) If (𝑥𝑛 ) and (𝑦𝑛 ) are convergent sequences and if 𝑥𝑛 ≤ 𝑦𝑛 , ∀𝑛 ≥ 𝐾 (or 𝑥𝑛 ≤ 𝑦𝑛 , ∀𝑛 ≥ 𝐾),
then lim 𝑥𝑛 ≤ lim 𝑦𝑛 .
𝑛→∞ 𝑛→∞
(2) If (𝑥𝑛 ) is a convergent sequence and 𝑎 ≤ 𝑥𝑛 ≤ 𝑏, ∀𝑛 ≥ 𝐾 (or 𝑎 < 𝑥𝑛 < 𝑏, ∀𝑛 ≥ 𝐾),
then 𝑎 ≤ lim 𝑥𝑛 ≤ 𝑏.
Squeeze Theorem 3.2.12: Let (𝑥𝑛 ), (𝑦𝑛 ), and (𝑧𝑛 ) be sequences of real numbers such that
𝑥𝑛 ≤ 𝑦𝑛 ≤ 𝑧𝑛 , ∀𝑛 ≥ 𝐾. If (𝑥𝑛 ) and (𝑧𝑛 ) are convergent with lim 𝑥𝑛 = lim 𝑧𝑛 = 𝑤. Then
𝑛→∞ 𝑛→∞
(𝑦𝑛 ) is convergent and lim 𝑦𝑛 = 𝑤.
𝑛→∞

Ratio Test for Convergence of sequences 3.2.13:

𝑥𝑛+1
Let (𝑥𝑛 ) be a sequence of positive real numbers such that lim =𝐿
𝑛→∞ 𝑥𝑛
(1) If 0 ≤ 𝐿 < 1, then (𝑥𝑛 ) converges and lim 𝑥𝑛 = 0.
𝑛→∞
(2) If 𝐿 > 1, then (𝑥𝑛 ) is unbounded and so diverges.
(3) If 𝐿 = 1, then the test fails.
Proof:
(1) Let 0 ≤ 𝐿 < 1 ⇒ ∃𝑟 ∈ (𝐿, 1) that is 𝐿 < 𝑟 < 1 ⇒ 𝑟 − 𝐿 > 0
𝑥
Take 𝜀 = 𝑟 − 𝐿 in the definition of lim 𝑥𝑛+1 = 𝐿 ⇒ ∃𝐾 ∈ ℕ such that
𝑛→∞ 𝑛
𝑥𝑛+1
| − 𝐿| < 𝑟 − 𝐿, ∀𝑛 ≥ 𝐾
𝑥𝑛
𝑥
⇒ −(𝑟 − 𝐿) < 𝑥𝑛+1 − 𝐿 < 𝑟 − 𝐿, ∀𝑛 ≥ 𝐾
𝑛
𝑥𝑛+1 𝑥𝑛+1
⇒ − 𝐿 < 𝑟 − 𝐿, ∀𝑛 ≥ 𝐾 ⇒ < 𝑟, ∀𝑛 ≥ 𝐾
𝑥𝑛 𝑥𝑛
⇒ 𝑥𝑛+1 < 𝑟𝑥𝑛 , ∀𝑛 ≥ 𝐾……①
For 𝑛 = 𝐾 ∶ ①⇒ 𝑥𝐾+1 < 𝑟𝑥𝐾
For 𝑛 = 𝐾 + 1 ∶ ① ⇒ 𝑥𝐾+2 < 𝑟𝑥𝐾+1 ⇒ 𝑥𝐾+2 < 𝑟 2 𝑥𝐾
For 𝑛 = 𝐾 + 2 ∶ ① ⇒ 𝑥𝐾+3 < 𝑟𝑥𝐾+2 ⇒ 𝑥𝐾+3 < 𝑟 3 𝑥𝐾
For 𝑛 = 𝐾 + 3 ∶ ① ⇒ 𝑥𝐾+4 < 𝑟𝑥𝐾+3 ⇒ 𝑥𝐾+4 < 𝑟 4 𝑥𝐾
⋮
For general 𝑛 ≥ 𝐾 ∶ ① ⇒ 𝑥𝑛 < 𝑟 𝑛−𝐾 𝑥𝐾 = 𝐶𝑟 𝑛 , where 𝐶 = 𝑟 −𝐾 𝑥𝐾
But 𝑥𝑛 > 0, ∀𝑛 ∈ ℕ ⇒ 0 < 𝑥𝑛 < 𝐶𝑟 𝑛 , ∀𝑛 ≥ 𝐾…………②

Since 0 < 𝑟 < 1 ⇒ lim 𝐶𝑟 𝑛 = 0 ……….③

𝑛→∞
② ⇒ lim 0 ≤ lim 𝑥𝑛 ≤ lim 𝐶𝑟 𝑛 ⇒ 0 ≤ lim 𝑥𝑛 ≤ ⏟ 0 ⇒ lim 𝑥𝑛 = 0.
𝑛→∞ 𝑛→∞ 𝑛→∞ 𝑛→∞ 𝑛→∞
by ③
(2) Let 𝐿 > 1 ⇒ ∃𝑟 ∈ (1, 𝐿) that is 1 < 𝑟 < 𝐿 ⇒ 𝐿 − 𝑟 > 0
 39

𝑥
Take 𝜀 = 𝐿 − 𝑟 in the definition of lim 𝑥𝑛+1 = 𝐿 ⇒ ∃𝐾 ∈ ℕ such that
𝑛→∞ 𝑛
𝑥𝑛+1
| − 𝐿| < 𝐿 − 𝑟, ∀𝑛 ≥ 𝐾
𝑥𝑛
𝑥
⇒ −(𝐿 − 𝑟) < 𝑥𝑛+1 − 𝐿 < 𝐿 − 𝑟, ∀𝑛 ≥ 𝐾
𝑛
𝑥𝑛+1 𝑥𝑛+1
⇒ −𝐿 + 𝑟 < − 𝐿, ∀𝑛 ≥ 𝐾 ⇒ 𝑟 < , ∀𝑛 ≥ 𝐾
𝑥𝑛 𝑥𝑛
⇒ 𝑥𝑛+1 > 𝑟𝑥𝑛 , ∀𝑛 ≥ 𝐾…..④
For 𝑛 = 𝐾 ∶ ④ ⇒ 𝑥𝐾+1 > 𝑟𝑥𝐾
For 𝑛 = 𝐾 + 1 ∶ ④ ⇒ 𝑥𝐾+2 > 𝑟𝑥𝐾+1 ⇒ 𝑥𝐾+2 > 𝑟 2 𝑥𝐾
For 𝑛 = 𝐾 + 2 ∶ ④ ⇒ 𝑥𝐾+3 > 𝑟𝑥𝐾+2 ⇒ 𝑥𝐾+3 > 𝑟 3 𝑥𝐾
For 𝑛 = 𝐾 + 3 ∶ ④ ⇒ 𝑥𝐾+4 < 𝑟𝑥𝐾+3 ⇒ 𝑥𝐾+4 > 𝑟 4 𝑥𝐾
⋮
For general 𝑛 ≥ 𝐾 ∶ ④ ⇒ 𝑥𝑛 > 𝑟 𝑛−𝐾 𝑥𝐾 = 𝐶𝑟 𝑛 ,……..⑤
where 𝐶 = 𝑟 −𝐾 𝑥𝐾
Since 𝑟 > 1 and 𝐶 > 0⇒ lim 𝐶𝑟 𝑛 = ∞ ……….⑥
𝑛→∞
⑤&⑥⇒ lim 𝑥𝑛 = ∞ ⇒ (𝑥𝑛 ) diverges.
𝑛→∞
(3) To show that the test fails if 𝐿 = 1:
1
1 𝑥𝑛+1 𝑛
 Take 𝑥𝑛 = 𝑛 ⇒ (𝑥𝑛 ) converges and 𝐿 = lim = lim 𝑛+1
1 = lim =1
𝑥𝑛 𝑛+1
𝑛
𝑥𝑛+1 𝑛+1
 Take 𝑥𝑛 = 𝑛 ⇒ (𝑥𝑛 ) diverges and 𝐿 = lim = lim =1
𝑥𝑛 𝑛

Root Test for Convergence of sequences 3.2.14:

Let (𝑥𝑛 ) be a sequence of positive real numbers such that lim 𝑛√𝑥𝑛 = 𝐿
𝑛→∞
(1) If 0 ≤ 𝐿 < 1, then (𝑥𝑛 ) converges and lim 𝑥𝑛 = 0.
𝑛→∞
(2) If 𝐿 > 1, then (𝑥𝑛 ) is unbounded and so diverges
(3) If 𝐿 = 1, then the test fails.

Proof:
(1) Let 0 ≤ 𝐿 < 1 ⇒ ∃𝑟 ∈ (𝐿, 1) that is 𝐿 < 𝑟 < 1 ⇒ 𝑟 − 𝐿 > 0
Take 𝜀 = 𝑟 − 𝐿 in the definition of lim 𝑛√𝑥𝑛 = 𝐿 ⇒ ∃𝐾 ∈ ℕ such that
𝑛→∞
𝑛
| √𝑥𝑛 − 𝐿| < 𝑟 − 𝐿, ∀𝑛 ≥ 𝐾
𝑛
⇒ −(𝑟 − 𝐿) < √𝑥𝑛 − 𝐿 < 𝑟 − 𝐿, ∀𝑛 ≥ 𝐾
⇒ 𝑛√𝑥𝑛 − 𝐿 < 𝑟 − 𝐿, ∀𝑛 ≥ 𝐾 ⇒ 𝑛√𝑥𝑛 < 𝑟, ∀𝑛 ≥ 𝐾
⇒ 𝑥𝑛 < 𝑟 𝑛 , ∀𝑛 ≥ 𝐾……①
But 𝑥𝑛 > 0, ∀𝑛 ∈ ℕ ⇒ 0 < 𝑥𝑛 < 𝑟 𝑛 , ∀𝑛 ≥ 𝐾…………②
Since 0 < 𝑟 < 1 we have lim 𝑟 𝑛 = 0
𝑛→∞
①&② ⇒ lim 0 ≤ lim 𝑥𝑛 ≤ lim 𝑟 𝑛 (by the squeeze theorem) ⇒ 0 ≤ lim 𝑥𝑛 ≤ 0
𝑛→∞ 𝑛→∞ 𝑛→∞ 𝑛→∞
⇒ lim 𝑥𝑛 = 0.
𝑛→∞
 40

(2) Let 𝐿 > 1 ⇒ ∃𝑟 ∈ (1, 𝐿) that is 1 < 𝑟 < 𝐿 ⇒ 𝐿 − 𝑟 > 0

Take 𝜀 = 𝐿 − 𝑟 in the definition of lim 𝑛√𝑥𝑛 = 𝐿 ⇒ ∃𝐾 ∈ ℕ such that

𝑛→∞
𝑛
| √𝑥𝑛 − 𝐿| < 𝐿 − 𝑟, ∀𝑛 ≥ 𝐾
𝑛
⇒ −(𝐿 − 𝑟) < √𝑥𝑛 − 𝐿 < 𝐿 − 𝑟, ∀𝑛 ≥ 𝐾
⇒ −𝐿 + 𝑟 < 𝑛√𝑥𝑛 − 𝐿, ∀𝑛 ≥ 𝐾
⇒ 𝑟 < 𝑛√𝑥𝑛 , ∀𝑛 ≥ 𝐾 ⇒ 𝑥𝑛 > 𝑟 𝑛 , ∀𝑛 ≥ 𝐾…..③
Since 𝑟 > 1 we have lim 𝑟 𝑛 = ∞ ……….④
𝑛→∞
③&④⇒ lim 𝑥𝑛 = ∞ ⇒ (𝑥𝑛 ) diverges.
(3) To show that the test fails if 𝐿 = 1:
1 𝑛
 Take 𝑥𝑛 = (1 + 𝑛)
1
⇒ (𝑥𝑛 ) converges to 𝑒 and 𝐿 = lim 𝑛√𝑥𝑛 = lim (1 + 𝑛) = 1
𝑛→∞ 𝑛→∞
 Take 𝑥𝑛 = 𝑛 ⇒ (𝑥𝑛 ) diverges and
1
1 ln(𝑛𝑛 ) ln(𝑛) ln(𝑛)
𝑛 lim
𝐿 = lim √𝑥𝑛 = lim 𝑛𝑛 = lim 𝑒 = lim 𝑒 𝑛 = 𝑒 𝑛→∞ 𝑛
𝑛→∞ 𝑛→∞ 𝑛→∞ 𝑛→∞
1/𝑛 1
lim lim
= 𝑒 𝑛→∞ 1 = 𝑒 𝑛→∞𝑛 = 𝑒0 = 1

Solving Problems:
sin 𝑛
(Q 3.2.15): Show that the sequence ( ) converges and find its limit.
𝑛
1 sin 𝑛 1
Solution: −1 ≤ sin 𝑛 ≤ 1, ∀𝑛 ∈ ℕ ⇒ −
⏟𝑛 ≤ ⏟ ≤⏟ , ∀𝑛 ∈ ℕ
𝑛 𝑛
𝑥𝑛 𝑦𝑛 𝑧𝑛
1 1
Since lim (− 𝑛) = lim =0
𝑛→∞ 𝑛→∞ 𝑛
sin 𝑛 sin 𝑛
⇒( ) converges and lim = 0 (by the squeeze theorem)
𝑛 𝑛→∞ 𝑛

(−1)𝑛
(Q 3.2.16) Show that the sequence ( ) converges and find its limit.
𝑛
1 (−1)𝑛 1
Solution: −1 ≤ (−1)𝑛 ≤ 1, ∀𝑛 ∈ ℕ ⇒ −
⏟𝑛 ≤ ⏟ ≤⏟ , ∀𝑛 ∈ ℕ
𝑛 𝑛
𝑥𝑛 𝑦𝑛 𝑧𝑛

1 1 (−1)𝑛 (−1)𝑛
Since lim (− 𝑛) = lim =0⇒( ) converges and lim = 0 (by the squeeze theorem)
𝑛→∞ 𝑛→∞ 𝑛 𝑛 𝑛→∞ 𝑛

Note
0 ±∞
(1) Some of the indeterminate Forms are: ∞ − ∞, 0 , ±∞ , 0∞, ∞0 , 00 , 1∞
𝑓(𝑥) 0 ±∞ 𝑓(𝑥) 𝑓 ′ (𝑥)
(2) L’Hospital’s Rule: If lim = 0 or ⇒ lim = lim
𝑛→∞ 𝑔(𝑥) ±∞ 𝑛→∞ 𝑔(𝑥) 𝑛→∞ 𝑔′ (𝑥)
 41

2
(Q 3.2.17) Show that the sequence ((𝑛!)1/𝑛 ) converges and find its limit.
Solution: 𝑛! = 1
⏟× 2 × ⋯ × (𝑛 − 1) × 𝑛
𝑛−times
𝑛 × 𝑛 × ⋯ × 𝑛 × 𝑛 = 𝑛𝑛 ⇒ 1 ≤ 𝑛! ≤ 𝑛𝑛
⇒ 1 ≤ 𝑛! ≤ ⏟
𝑛−times
2 2 2 2
⇒ 11/𝑛 ≤ (𝑛!)1/𝑛 ≤ (𝑛𝑛 )1/𝑛 ⇒ 1 ≤ (𝑛!)1/𝑛 ≤ 𝑛1/𝑛 ……….①
 lim 1 = 1………②
𝑛→∞
1 ln 𝑛 ln 𝑛 1/𝑛 1
1/𝑛 ln 𝑛𝑛 lim lim lim
 lim 𝑛 = lim 𝑒 = lim 𝑒 𝑛 = 𝑒 𝑛→∞ 𝑛 = 𝑒 𝑛→∞ 1 = 𝑒 𝑛→∞𝑛 = 𝑒 0 = 1…③
𝑛→∞ 𝑛→∞ 𝑛→∞
1
1/𝑛2
①&②&③: ⇒ ((𝑛!) ) converges and lim (𝑛!)𝑛2 = 1 (by the squeeze theorem)
𝑛→∞

(Q 3.2.18) Find each of the following limit if it exists:

(1) lim 𝑛1/ ln(𝑛) (2) lim 𝑛2 𝑎𝑛 , where 0 < 𝑎 < 1
𝑛→∞ 𝑛→∞
𝑛! 𝑎𝑛+1 +𝑏𝑛+1
(3) lim (4) lim , where 0 < 𝑎 < 𝑏
𝑛→∞ 𝑛𝑛 𝑛→∞ 𝑎𝑛 +𝑏𝑛
Solution:
1/ ln(𝑛) )
(1) lim 𝑛1/ ln(𝑛) = lim 𝑒 ln(𝑛 = lim 𝑒 (ln(𝑛))/ ln(𝑛) = 𝑒
𝑛→∞ 𝑛→∞ 𝑛→∞
1 1
(2) 0 < 𝑎 < 1 ⇒ 𝑎 > 1 ⇒ Let 𝑏 = 𝑎 ⇒ 𝑏 > 1 and lim 𝑏 𝑛 = ∞
𝑛→∞
𝑛2 2𝑛 2
lim 𝑛2 𝑎𝑛 = lim 𝑛 = lim 𝑛 = lim 𝑛 =0
𝑛→∞ 𝑛→∞ 𝑏 𝑛→∞ 𝑏 ln 2 𝑛→∞ 𝑏 (ln 2)2
𝑛! 1×2×⋯×(𝑛−1)×𝑛 1 2 𝑛−1 𝑛 1 1
(3) 0 < 𝑛𝑛 = = 𝑛 × 𝑛 × ⋯× ×𝑛 ≤ 𝑛 ×1 ×⋯×1×1 = 𝑛
𝑛×𝑛×⋯×𝑛×𝑛 𝑛
𝑛! 1
0 < 𝑛𝑛 ≤ 𝑛 , ∀𝑛 ∈ ℕ……….①
𝑛!
The squeeze theorem ⇒ lim =0
𝑛→∞ 𝑛𝑛
𝑎
(4) Let 𝑟 = 𝑏 ⇒ 0 < 𝑟 < 1 (since 0 < 𝑎 < 𝑏) ⇒ lim 𝑟 𝑛 = 0
𝑛→∞

𝑎𝑛+1 𝑏𝑛+1
𝑎𝑛+1 +𝑏𝑛+1 + 𝑛 𝑎𝑟 𝑛 +𝑏 lim (𝑎𝑟 𝑛 +𝑏) 𝑎×0+𝑏
𝑏𝑛 𝑏 𝑛→∞
lim = lim 𝑛 𝑛 = lim = = =𝑏
𝑛→∞ 𝑎𝑛 +𝑏𝑛 𝑛→∞ 𝑎𝑛 +𝑏𝑛 𝑛→∞ 𝑟 𝑛 +1 lim (𝑟 𝑛 +1)
𝑛→∞
0+1
𝑏 𝑏

𝑛! 𝑛𝑛
(Q 3.2.19) Find lim and lim if exist.
𝑛→∞ 𝑛𝑛 𝑛→∞ 𝑛!
Solution:
𝑛! 𝑛!
 lim 𝑛𝑛 : Let 𝑥𝑛 = 𝑛𝑛
𝑛→∞
𝑥𝑛+1 (𝑛+1)! 𝑛𝑛 (𝑛+1)(𝑛!) 𝑛𝑛 𝑛 𝑛 1
⇒ 𝐿 = lim = lim (𝑛+1)𝑛+1
× = lim × = lim (𝑛+1) = 𝑒
𝑛→∞ 𝑥𝑛 𝑛→∞ 𝑛! 𝑛→∞ (𝑛+1)(𝑛+1)𝑛 𝑛! 𝑛→∞
𝑛!
0 ≤ 𝐿 < 1 ⇒ The Ratio Test applies ⇒ lim =0
𝑛→∞ 𝑛𝑛
𝑛! 𝑛!
 lim : Let 𝑦𝑛 = 𝑛𝑛
𝑛→∞ 𝑛𝑛
𝑦𝑛+1
𝐿 = lim = 𝑒 ⇒ 𝐿 > 1 ⇒ The Ratio Test applies
𝑛→∞ 𝑦𝑛
𝑛𝑛 𝑛𝑛
⇒ ( 𝑛! ) is unbounded and so lim =∞
𝑛→∞ 𝑛!
 42

(Q 3.2.20) Let (𝑥𝑛 ) be a sequence of real numbers. Prove that if lim 𝑥𝑛 = 𝑥, then lim |𝑥𝑛 | = |𝑥|, that
𝑛→∞ 𝑛→∞
is

lim |𝑥𝑛 | = | lim 𝑥𝑛 |

𝑛→∞ 𝑛→∞

Proof: Let 𝜀 > 0 be given want to find 𝐾 ∈ ℕ such that ||𝑥𝑛 | − |𝑥|| < 𝜀, ∀𝑛 ≥ 𝐾

Observe that: from the triangle inequality ||𝑥𝑛 | − |𝑥|| ≤ |𝑥𝑛 − 𝑥| ………①

By the definition of lim 𝑥𝑛 = 𝑥: Take 𝜀 > 0 ⇒ ∃𝐾 ∈ ℕ such that |𝑥𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾……..②

①&②: ⇒ ||𝑥𝑛 | − |𝑥|| < 𝜀, ∀𝑛 ≥ 𝐾.

⇒ lim |𝑥𝑛 | = |𝑥|.

𝑛→∞

(Q 3.2.21) Prove that if 𝑎, 𝑏 ≥ 0, then: |√𝑎 − √𝑏| ≤ √|𝑎 − 𝑏|

Proof: Exercise.
(Q 3.2.22) Let (𝑥𝑛 ) be a sequence of positive real numbers. Prove that if lim 𝑥𝑛 = 𝑥, then
𝑛→∞

lim √𝑥𝑛 = √𝑥, that is lim √𝑥𝑛 = √ lim 𝑥𝑛 .

𝑛→∞ 𝑛→∞ 𝑛→∞

Proof: Let 𝜀 > 0 be given want to find 𝐾 ∈ ℕ such that |√𝑥𝑛 − √𝑥. | < 𝜀, ∀𝑛 ≥ 𝐾

Observe that |√𝑥𝑛 − √𝑥| ≤ √|𝑥𝑛 − 𝑥| ………①

By the definition of lim 𝑥𝑛 = 𝑥: Take 𝜀 2 > 0 ⇒ ∃𝐾 ∈ ℕ such that |𝑥𝑛 − 𝑥| < 𝜀 2 , ∀𝑛 ≥ 𝐾…….②

①&②: ⇒ |√𝑥𝑛 − √𝑥. | < √𝜀 2 = 𝜀, ∀𝑛 ≥ 𝐾.

⇒ lim √𝑥𝑛 = √𝑥.

𝑛→∞

Exercises 3.2.23 (pages 69+70): 2-5, 7,9,13,15,22

 43

Section 3.3: Monotone Sequences

Definition 3.3.1: A sequence (𝑥𝑛 ) of real numbers is said to be:

(1) increasing if 𝑥1 ≤ 𝑥2 ≤ 𝑥3 ≤ ⋯, that is 𝑥𝑛 ≤ 𝑥𝑛+1 , ∀𝑛 ∈ ℕ.
(2) decreasing if 𝑥1 ≥ 𝑥2 ≥ 𝑥3 ≥ ⋯, that is 𝑥𝑛 ≥ 𝑥𝑛+1 , ∀𝑛 ∈ ℕ.
(3) monotone if either (𝑥𝑛 ) is increasing or decreasing

Example 3.3.2:
(1) The following sequences are increasing (and so they are monotone):
(1,2,3, … ), (1,2,2,3,3,3, … ), (𝑎, 𝑎2 , 𝑎3 , … ), where 𝑎 > 1
(2) The following sequences are decreasing (and so they are monotone):
1 1 1 1 1
(1, 2 , 3 , … ) , (1, 2 , 22 , 23 , … ) , (𝑏, 𝑏 2 , 𝑏 3 , … ), where 0 < 𝑏 < 1
(3) The following sequences are not monotone:
1 1
(1, −1,1, −1, … ), (1,2, , 3,4, , … )
2 4
(4) The following sequences are not monotone but they are ultimately monotone:
1 1
(7,6,2,1,2,3, … ) , (−2,0,1, , , … )
⏟
⏟ 2 3
the tail 𝑋 increasing
3
the tail 𝑌2 decreasing

Recall that:
A sequence (𝑥𝑛 ) is called ultimately has a certain property if some tail of (𝑥𝑛 ) has this property.

Remark 3.3.3:
(1) A sequence (𝑥𝑛 ) is increasing iff (−𝑥𝑛 ) is decreasing.
(2) A sequence (𝑥𝑛 ) is decreasing iff (−𝑥𝑛 ) is increasing.
(3) Recall that if 𝑢∗ is a real number, then sup 𝑆 = 𝑢∗ iff the following two conditions hold:
𝐶1 : 𝑢∗ is an UB of 𝑆.
𝐶2′ : if 𝑢 < 𝑢∗ , then 𝑢 is not an UB of 𝑆.

Monotone Convergence Theorem (MCT) 3.3.4:

Let (𝑥𝑛 ) be a monotone sequence of real numbers. Then
(𝑥𝑛 ) converges iff (𝑥𝑛 ) is bounded.
Further
(1) If (𝑥𝑛 ) is a bounded increasing sequence, then lim 𝑥𝑛 = sup{𝑥𝑛 : 𝑛 ∈ ℕ}
𝑛→∞
(2) If (𝑥𝑛 ) is a bounded decreasing sequence, then lim 𝑥𝑛 = inf{𝑥𝑛 : 𝑛 ∈ ℕ}
𝑛→∞

Proof:
Suppose that (𝑥𝑛 ) converges ⇒ (𝑥𝑛 ) is bounded (by Theorem 3.2.2)
Suppose that (𝑥𝑛 ) is monotone and bounded.
Since (𝑥𝑛 ) is monotone ⇒ (𝑥𝑛 ) is increasing or decreasing:
Case 1: Let (𝑥𝑛 ) be increasing ⇒ 𝑥𝑛 ≤ 𝑥𝑛+1 , ∀𝑛 ∈ ℕ
 44

Since (𝑥𝑛 ) is bounded, then by the Completeness Theorem we have (𝑥𝑛 ) has a
supremum. So, let sup{𝑥𝑛 : 𝑛 ∈ ℕ} = 𝑥.
By using the squeeze theorem we will show that lim 𝑥𝑛 = 𝑥
𝑛→∞
Since sup{𝑥𝑛 : 𝑛 ∈ ℕ} = 𝑥
Condition 𝐶1 ⇒ 𝑥 is an UB of {𝑥𝑛 : 𝑛 ∈ ℕ} ⇒ 𝑥𝑛 ≤ 𝑥, ∀𝑛 ∈ ℕ
⇒ 𝑥𝑛 − 𝑥 ≤ 0, ∀𝑛 ∈ ℕ…………..①
1
Since 𝑥 − 𝑛 < 𝑥, ∀𝑛 ∈ ℕ and 𝑥 = sup{𝑥𝑛 : 𝑛 ∈ ℕ}. Then:
1
Condition 𝐶2′ ⇒ ∀𝑛 ∈ ℕ, 𝑥 − 𝑛 is not an UB of {𝑥𝑛 : 𝑛 ∈ ℕ}
1
⇒ ∃𝐾 ∈ ℕ such that 𝑥 − 𝑛 < 𝑥𝐾 ……….②
But (𝑥𝑛 ) is increasing ⇒ 𝑥𝐾 ≤ 𝑥𝑛 , ∀𝑛 ≥ 𝐾……..③
1 1
②&③: ⇒ 𝑥 − 𝑛 < 𝑥𝑛 , ∀𝑛 ≥ 𝐾 ⇒ − 𝑛 < 𝑥𝑛 − 𝑥, ∀𝑛 ≥ 𝐾………④
1
①&④: ⇒ − 𝑛 < 𝑥𝑛 − 𝑥 < 0, ∀𝑛 ≥ 𝐾…………⑤
1
The squeeze theorem ⑤ ⇒ lim (− 𝑛) ≤ lim (𝑥𝑛 − 𝑥) ≤ 0 ⇒ 0 ≤ lim (𝑥𝑛 − 𝑥) ≤ 0
𝑛→∞ 𝑛→∞ 𝑛→∞
⇒ lim(𝑥𝑛 − 𝑥) = 0 ⇒ lim 𝑥𝑛 = 𝑥 ⇒ (𝑥𝑛 ) converges and lim 𝑥𝑛 = sup{𝑥𝑛 : 𝑛 ∈ ℕ}
Case 2: Let (𝑥𝑛 ) be decreasing ⇒ (−𝑥𝑛 ) is increasing
Since (𝑥𝑛 ) is bounded ⇒ (−𝑥𝑛 ) is bounded
∴ (−𝑥𝑛 ) is increasing bounded sequence
Case 1 ⇒ (−𝑥𝑛 ) converges and lim (−𝑥𝑛 ) = sup{−𝑥𝑛 : 𝑛 ∈ ℕ}
𝑛→∞
⇒ (𝑥𝑛 ) converges and − ( lim 𝑥𝑛 ) = − inf{𝑥𝑛 : 𝑛 ∈ ℕ}
𝑛→∞
⇒ (𝑥𝑛 ) converges and lim 𝑥𝑛 = inf{𝑥𝑛 : 𝑛 ∈ ℕ}
𝑛→∞

Remark 3.3.5: The contraposition of the MCT 3.3.4 is: Let (𝑥𝑛 ) be a monotone sequence. Then
(𝑥𝑛 ) diverges iff (𝑥𝑛 ) is unbounded.

1
Example 3.3.6: Using the MCT 3.3.4, show that the sequence ( ) converges and find its limit.
√𝑛+1
1 1 1 1
Solution: ( )=( , , , … ) ⇒ The sequence is decreasing………①
√𝑛+1 √2 √3 √4
1 1
Also, 0 < ≤ , ∀𝑛 ∈ ℕ ⇒ The sequence is bounded………..②
√𝑛+1 √2
1 1 1 1 1 1
MCT ⇒ ( ) converges and lim = inf { : 𝑛 ∈ ℕ} = inf { , , ,…} = 0
√𝑛+1 𝑛→∞ √𝑛+1 √𝑛+1 √2 √3 √4

Proof by Induction of a proposition 𝑃(𝑛), 𝑛 ∈ ℕ:

Base Step: Show 𝑃(1) is true
Inductive Step: Show 𝑃(𝑘) ⇒ 𝑃(𝑘 + 1)
①: Suppose that 𝑃(𝑘) is true
②: Show that 𝑃(𝑘 + 1) is true
 45

Example 3.3.7: Let (𝑥𝑛 ) be the sequence defined

inductively by 𝑥1 = 1, 𝑥𝑛+1 = √2 + 𝑥𝑛 , 𝑛 ∈ ℕ
(1) Show that (𝑥𝑛 ) is convergent.
(2) Find lim 𝑥𝑛
𝑛→∞
(3) Find sup(𝑥𝑛 ) and inf(𝑥𝑛 )
Solution:
(2) We use the MCT, that is we want to show that (𝑥𝑛 ) is monotone and bounded:
First we make a guessing to see if the sequence is monotone and bounded:
𝑥1 = 1
𝑥2 = √2 + 𝑥1 = √2 + 1 = √3 ≅ 1.7 (𝑥𝑛 ) increasing that is 𝑥𝑛 ≤ 𝑥𝑛+1 , ∀𝑛 ∈ ℕ
⇒ &
𝑥3 = √2 + 𝑥2 ≅ √2 + 1.7 = √3.7 ≅ 1.9
(𝑥𝑛 ) bounded with 𝑥1 ≤ 𝑥𝑛 ≤ 2, ∀𝑛 ∈ ℕ
𝑥4 = √2 + 𝑥3 ≅ √2 + 1.9 = √3.9 ≅ 1.97}

Claim 1: (𝒙𝒏 ) is increasing: we show 𝑥𝑛 ≤ 𝑥𝑛+1 , ∀𝑛 ∈ ℕ (by induction)

Base Step: At 𝑛 = 1:
𝑥1 = 1 and 𝑥2 = √3 ⇒ 𝑥1 ≤ 𝑥2 is true
Inductive Step:
 Suppose that the inequality holds at 𝑛 = 𝑘, that is
𝑥𝑘 ≤ 𝑥𝑘+1 ……….①
 At 𝑛 = 𝑘 + 1:
𝑥𝑘+1 = √2 + 𝑥𝑘 ⏟ ≤ √2 + 𝑥𝑘+1 = 𝑥𝑘+2
by ①
𝑥𝑘+1 ≤ 𝑥𝑘+2 is true
∴ 𝑥𝑛 ≤ 𝑥𝑛+1 , ∀𝑛 ∈ ℕ ⇒ (𝑥𝑛 ) is increasing

Claim 2: (𝒙𝒏 ) is bounded:

 Since (𝑥𝑛 ) is increasing: 𝑥1 ≤ 𝑥2 ≤ 𝑥3 ≤ ⋯
⇒ 𝑥1 = 1 is a LB of (𝑥𝑛 ) ⇒ (𝑥𝑛 ) bounded below by 𝑥1 = 1
⇒ 1 ≤ 𝑥𝑛 , ∀𝑛 ∈ ℕ……②
 We show that 𝑥𝑛 ≤ 2, ∀𝑛 ∈ ℕ (by induction)
Base Step: At 𝑛 = 1:
𝑥1 = 1 ⇒ 𝑥1 ≤ 2 is true
Inductive Step: Suppose that the inequalities hold at 𝑛 = 𝑘, that is
𝑥𝑘 ≤ 2……….③
At 𝑛 = 𝑘 + 1:
𝑥𝑘+1 = √2 + 𝑥𝑘 ⏟≤ √2 + 2 = √4 = 2
by ③
⇒ 𝑥𝑘+1 ≤ 2 is true
∴ 𝑥𝑛 ≤ 2, ∀𝑛 ∈ ℕ ………….④
②&④: ⇒ 1 ≤ 𝑥𝑛 ≤ 2, ∀𝑛 ∈ ℕ ⇒ (𝑥𝑛 ) is bounded
∴ Claim 1& Claim 2 ⇒ (𝑥𝑛 ) is increasing and bounded
MCT ⇒ (𝑥𝑛 ) converges.

(3) To find lim 𝑥𝑛 : By part (1) (𝑥𝑛 ) converges so let lim 𝑥𝑛 = 𝑥

𝑛→∞ 𝑛→∞
 46

𝑥𝑛+1 = √2 + 𝑥𝑛 ⇒ lim 𝑥𝑛+1 = √2 + lim 𝑥𝑛 ⇒ 𝑥 = √2 + 𝑥

𝑛→∞ 𝑛→∞
2 2
⇒ 𝑥 = 2 + 𝑥 ⇒ 𝑥 − 𝑥 − 2 = 0 ⇒ (𝑥 − 2)(𝑥 + 1) = 0
⇒ 𝑥 = 2 or 𝑥 = −1……….⑤
Observe that: 1
⏟≤ 𝑥𝑛 ≤ 2, ∀𝑛 ∈ ℕ ⇒ 1 ≤ lim 𝑥𝑛 ≤ 2
𝑛→∞
the study of boundedness
⇒ 1 ≤ 𝑥 ≤ 2………….⑥
⑤&⑥: ⇒ 𝑥 = 2 ⇒ lim 𝑥𝑛 = 2
𝑛→∞

(4) Since (𝑥𝑛 ) is increasing that is 𝑥1 ≤ 𝑥2 ≤ 𝑥3 ≤ ⋯

⇒ sup(𝑥𝑛 ) = 𝑥∞ = lim 𝑥𝑛 = 2 (by the MCT) and inf(𝑥𝑛 ) = 𝑥1 = 1
𝑛→∞

Example 3.3.8: Let (𝑥𝑛 ) be the sequence defined

inductively by 𝑥1 ≥ 2, 𝑥𝑛+1 = 1 + √𝑥𝑛 − 1, 𝑛 ∈ ℕ
(1) Show that (𝑥𝑛 ) is convergent.
(2) Find lim 𝑥𝑛
𝑛→∞
(3) Find sup(𝑥𝑛 ) and inf(𝑥𝑛 )
Solution:
(1) We use the MCT, that is we want to show that (𝑥𝑛 ) is monotone and bounded:
First we make a guessing to see if the sequence is monotone and bounded:
𝑥1 ≥ 2 ⇒ take 𝑥1 = 10 (just for guessing)
𝑥2 = 1 + √𝑥1 − 1 = 1 + √10 − 1 = 4 (𝑥𝑛 ) decreasing that is 𝑥𝑛 ≤ 𝑥𝑛+1 , ∀𝑛 ∈ ℕ
⇒ &
𝑥3 = 1 + √𝑥2 − 1 = 1 + √4 − 1 = 1 + √3 ≅ 2.7
(𝑥𝑛 ) bounded with 2 ≤ 𝑥𝑛 ≤ 𝑥1 , ∀𝑛 ∈ ℕ
𝑥4 = 1 + √𝑥3 − 1 ≅ 1 + √1.7 ≅ 2.3 }

Claim 1: (𝒙𝒏 ) is decreasing: we show 𝑥𝑛 ≥ 𝑥𝑛+1 , ∀𝑛 ∈ ℕ (by induction)

Base Step: At 𝑛 = 1:
𝑥1 ≥ 2 and 𝑥2 = 1 + √𝑥1 − 1 ≤
⏟ 1 + 𝑥1 − 1 = 𝑥1
since 𝑥1 −1≥1
⇒ 𝑥1 ≥ 𝑥2 is true
Inductive Step:
 Suppose that the inequality holds at 𝑛 = 𝑘, that is
𝑥𝑘 ≥ 𝑥𝑘+1 ……….①
 At 𝑛 = 𝑘 + 1:
𝑥𝑘+1 = 1 + √𝑥𝑘 − 1 ≥⏟ 1 + √𝑥𝑘+1 − 1 = 𝑥𝑘+2
by ①
𝑥𝑘+1 ≥ 𝑥𝑘+2 is true
∴ 𝑥𝑛 ≥ 𝑥𝑛+1 , ∀𝑛 ∈ ℕ ⇒ (𝑥𝑛 ) is decreasing
 47

Claim 2: (𝒙𝒏 ) is bounded:

 Since (𝑥𝑛 ) is decreasing: 𝑥1 ≥ 𝑥2 ≥ 𝑥3 ≥ ⋯
⇒ 𝑥1 is an UB of (𝑥𝑛 ) ⇒ (𝑥𝑛 ) bounded above by 𝑥1
⇒ 𝑥1 ≥ 𝑥𝑛 , ∀𝑛 ∈ ℕ……②
 We show that 𝑥𝑛 ≥ 2, ∀𝑛 ∈ ℕ (by induction)
Base Step: At 𝑛 = 1:
𝑥1 ≥ 2 is given ⇒ 𝑥1 ≥ 2 is true
Inductive Step: Suppose that the inequalities hold at 𝑛 = 𝑘, that is
𝑥𝑘 ≥ 2……….③
At 𝑛 = 𝑘 + 1:
𝑥𝑘+1 = 1 + √𝑥𝑘 − 1 ⏟ ≥ 1 + √2 − 1 = 2
by ③
⇒ 𝑥𝑘+1 ≥ 2 is true
∴ 𝑥𝑛 ≥ 2, ∀𝑛 ∈ ℕ ………….④
②&④: ⇒ 2 ≤ 𝑥𝑛 ≤ 𝑥1 , ∀𝑛 ∈ ℕ ⇒ (𝑥𝑛 ) is bounded
∴ Claim 1& Claim 2& MCT ⇒ (𝑥𝑛 ) converges.

(2) To find lim 𝑥𝑛 : By part (1) (𝑥𝑛 ) converges so let lim 𝑥𝑛 = 𝑥

𝑛→∞

𝑥𝑛+1 = 1 + √𝑥𝑛 − 1 ⇒ lim 𝑥𝑛+1 = 1 + √ lim 𝑥𝑛 − 1 ⇒ 𝑥 = 1 + √𝑥 − 1

𝑛→∞ 𝑛→∞

⇒ 𝑥 − 1 = √𝑥 − 1 ⇒ (𝑥 − 1) = 𝑥 − 1 ⇒ 𝑥 2 − 2𝑥 + 1 = 𝑥 − 1
2

⇒ 𝑥 2 − 3𝑥 + 2 = 0 ⇒ (𝑥 − 1)(𝑥 − 2) = 0
⇒ 𝑥 = 1 or 𝑥 = 2……….⑤
Observe that: ⏟
2 ≤ 𝑥𝑛 ≤ 𝑥1 , ∀𝑛 ∈ ℕ ⇒ 2 ≤ lim 𝑥𝑛 ≤ 𝑥1
𝑛→∞
the study of boundedness
⇒ 2 ≤ 𝑥 ≤ 𝑥1 ………….⑥
⑤&⑥: ⇒ 𝑥 = 2 ⇒ lim 𝑥𝑛 = 2
𝑛→∞

(3) Since (𝑥𝑛 ) is decreasing that is 𝑥1 ≥ 𝑥2 ≥ 𝑥3 ≥ ⋯

⇒ sup(𝑥𝑛 ) = 𝑥1 and inf(𝑥𝑛 ) = 𝑥∞ = lim 𝑥𝑛 = 2 (by the MCT)
𝑛→∞

Remark 3.3.9: There are other examples in the text book, the students should study them.

Exercises 3.3.10 (page 77): 1,2,12

 48

Section 3.4: Subsequences and the Bolzano-Weierstrass Theorem

Definition 3.4.1: Let 𝑋 = (𝑥𝑛 ) be a sequence and (𝑛1 , 𝑛2 , 𝑛3 , … ) be a sequence of natural numbers
such that 𝑛1 < 𝑛2 < 𝑛3 < ⋯. The sequence 𝑋 ′ = (𝑥𝑛𝑘 ) = (𝑥𝑛1 , 𝑥𝑛2 , 𝑥𝑛3 , … ) is
called a subsequence of 𝑋 = (𝑥𝑛 ).
𝑋 ‫ 𝑋 هي حدود في المتتالية‬′ ‫) من المتتالية 𝑋 إذا كانت جميع حدود‬subsequence( ‫ 𝑋 متتالية جزئية‬′ ‫تكون المتتالية‬
.‫مع المحافظة على الترتيب‬
1 1 1 1 1 1 1 1
Example 3.4.2: Let 𝑋 = (𝑛) = (1, 2 , 3 , 4 , … ) ⇒ 𝑥𝑛 = 𝑛 ⇒ 𝑥1 = 1, 𝑥2 = 2 , 𝑥3 = 3 , 𝑥4 = 4,…
𝑥2 𝑥 4 𝑥6
1 ⏞1 ⏞ 1 ⏞ 1
(1) (2𝑛) = ( 2 , 4 , 6 , … ) is a subsequence of 𝑋.

𝑥3 𝑥5 𝑥2 𝑥6 𝑥24
𝑥1
⏞
1 ⏞ 1 ⏞
1 ⏞1 ⏞ 1
(2) The following are subsequences of 𝑋: ( ⏞
1 , 3 , 5 , … ) , ( 2! , 3! , ,…)
4!

𝑥2
𝑥1
⏞
1 1
1
⏞ 1 1
(3) The following are not subsequences of 𝑋: (0,
⏟ 1, 3 , 5 , … ) , ( 2 , 1 , 4 , 3 , … )
0 is not a term in 𝑋 ⏟
The ordering is
different than that of 𝑋
Remark 3.4.3:
(1) Every sequence is a subsequence of itself.
Proof: Let 𝑋 = (𝑥𝑛 ) = (𝑥1 , 𝑥2 , 𝑥3 , … ) and take the natural numbers:
𝑛1 = 1, 𝑛2 = 2, 𝑛3 = 3,…,𝑛𝑘 = 𝑘,… ⇒ 𝑛1 < 𝑛2 < 𝑛3 < ⋯.
So, 𝑋 ′ = (𝑥𝑛1 , 𝑥𝑛2 , 𝑥𝑛3 , … ) is a subsequence of 𝑋. But 𝑋 ′ = (𝑥1 , 𝑥2 , 𝑥3 , … ) ⇒ 𝑋 ′ = 𝑋.

Every sequence is a subsequence of itself.

(2) If 𝑋 ′ = (𝑥𝑛𝑘 ) is a subsequence of 𝑋 = (𝑥𝑛 ), then 𝑛𝑘 ≥ 𝑘, ∀𝑘 ∈ ℕ that is:

𝑛1 ≥ 1, 𝑛2 ≥ 2, 𝑛3 ≥ 3,…
(3) Any tail of a sequence 𝑋 = (𝑥𝑛 ) is a subsequence 𝑋.
Proof: Let 𝑋 = (𝑥𝑛 ) = (𝑥1 , 𝑥2 , 𝑥3 , … )
The 𝑚-tail: 𝑋𝑚 = (𝑥𝑚+𝑛 ) = (𝑥𝑚+1 , 𝑥𝑚+2 , 𝑥𝑚+3 , … ) ⇒ 𝑋𝑚 is a subsequence of 𝑋
1 1 1 1
For example: Let 𝑋 = (𝑛) = (1, 2 , 3 , 4 , … )
1 1 1 1
The 3-tail of 𝑋 is: 𝑋3 = (𝑥3+𝑛 ) = (3+𝑛) = (4 , 5 , 6 , … )
⇒ 𝑋3 is a subsequence of 𝑋.
 49

(4) A subsequence 𝑋 ′ = (𝑥𝑛𝑘 ) of 𝑋 = (𝑥𝑛 ) need not be a tail of (𝑥𝑛 ).

1 1 1 1 1 1 1 1
For example: Let 𝑋 = (𝑛) = (1, 2 , 3 , 4 , … ) and 𝑋 ′ = (2𝑛) = (2 , 4 , 6 , … )
⇒ 𝑋 ′ is a subsequence of 𝑋 and 𝑋 ′ is not a tail of 𝑋.

Recall that
(𝑥𝑛 ) converges to 𝑥 iff any tail 𝑋𝑚 = (𝑥𝑚+𝑛 ) converges to
that is: lim 𝑥𝑛 = 𝑥 ⇔ lim 𝑥𝑚+𝑛 = 𝑥
𝑛→∞ 𝑛→∞

Theorem 3.4.4: Let 𝑋 = (𝑥𝑛 ) be a sequence of real numbers. Then

(𝑥𝑛 ) converges to 𝑥 iff any subsequence 𝑋 ′ = (𝑥𝑛𝑘 ) converges to 𝑥, that is
lim 𝑥𝑛 = 𝑥 ⇔ lim 𝑥𝑛𝑘 = 𝑥
𝑛→∞ 𝑘→∞
Proof:
Suppose that lim 𝑥𝑛 = 𝑥. We will use the definition of limit to show that lim 𝑥𝑛𝑘 = 𝑥:
𝑛→∞ 𝑘→∞
Let 𝜀 > 0 be given. Want to show that ∃𝐾 ∈ ℕ such that |𝑥𝑛𝑘 − 𝑥| < 𝜀, ∀𝑘 ≥ 𝐾.
Since lim 𝑥𝑛 = 𝑥: ⇒ ∃𝐾 ∈ ℕ such that |𝑥𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾……….①
𝑛→∞
Since 𝑛𝑘 ≥ 𝑘, ∀𝑘 ∈ ℕ, then taking 𝑘 ≥ 𝐾 we have 𝑛𝑘 ≥ 𝑘 ≥ 𝐾
⇒ 𝑛𝑘 ≥ 𝐾 so in ① taking 𝑛 = 𝑛𝑘 we have |𝑥𝑛𝑘 − 𝑥| < 𝜀
∴ |𝑥𝑛𝑘 − 𝑥| < 𝜀, ∀𝑘 ≥ 𝐾.
Suppose that all subsequences 𝑋 ′ of 𝑋 converges to 𝑥
Take 𝑋 ′ = 𝑋 is a subsequence of 𝑋 ⇒ 𝑋 converges to 𝑥

Corollary 3.4.5 (Contraposition of Theorem 3.4.4):

A sequence (𝑥𝑛 ) diverges iff at least one of the following holds:
(1) There exist a subsequence (𝑥𝑛𝑘 ) of 𝑋 such that (𝑥𝑛𝑘 ) diverges
(2) There exist two convergent subsequences (𝑥𝑛𝑘 ) and (𝑥𝑚𝑘 ) such that lim 𝑥𝑛𝑘 ≠ lim 𝑥𝑚𝑘 .
𝑘→∞ 𝑘→∞
1 1 1
Example 3.4.6: Show that the sequence 𝑋 = (−1,1,1, −1,2, , −1,3, , −1,4, , … ) diverges.
2 3 4

1st Solution: Let 𝑋 ′ = (1,2,3,4, … ) ⇒ 𝑋 ′ is a divergent subsequence of 𝑋

⇒ 𝑋 diverges.
1 1
2nd Solution: Let 𝑋 ′ = (−1, −1, −1, … ) and 𝑋 ′′ = (1, 2 , 3 , … ) ⇒ 𝑋 ′ and 𝑋 ′′ are subsequence of 𝑋

But lim 𝑋 ′ = −1 and lim 𝑋 ′′ = 0 ⇒ lim 𝑋 ′ ≠ lim 𝑋 ′′ ⇒ 𝑋 diverges.

 50

Example 3.4.7:
(1) Show that the sequence (sin(𝑛)) diverges.
(2) Show that the sequence (cos(𝑛)) diverges.

Solution: We prove part (1), the proof of part (2) is an exercise.

(1) To show that (sin(𝑛)) diverges we will use the proof by contradiction:
Assume that (sin(𝑛)) converges to 𝑥:
⇒ All subsequences of (sin(𝑛)) converges to 𝑥………①

To have a contradiction, we will construct two subsequences of (sin(𝑛)) with different limits
as follows:
𝜋 5𝜋
 Let 𝐼𝑘 = [ 6 + 2(𝑘 − 1)𝜋, 6 + 2(𝑘 − 1)𝜋] , 𝑘 ∈ ℕ
4𝜋 2𝜋
⇒ Length of 𝐼𝑘 = 6 = 3 > 2, ∀𝑘 ∈ ℕ
⇒ ∃𝑛𝑘 ∈ ℕ such that 𝑛𝑘 ∈ 𝐼𝑘 , ∀𝑘 ∈ ℕ
⇒ (sin(𝑛𝑘 )) is a subsequence of (sin(𝑛)) and by ①⇒ lim sin(𝑛𝑘 ) = 𝑥 ………②
𝑘→∞
1
Since ≤ sin(𝑛𝑘 ) ≤ 1, ∀𝑘 ∈ ℕ
2
1 1
⇒ 2 ≤ lim sin(𝑛𝑘 ) ≤ 1 and by ②: 2 ≤ 𝑥 ≤ 1 ………③
𝑘→∞

7𝜋 11𝜋
 Let 𝐽𝑘 = [ + 2(𝑘 − 1)𝜋, + 2(𝑘 − 1)𝜋] , 𝑘 ∈ ℕ
6 6
4𝜋 2𝜋
⇒ Length of 𝐽𝑘 = 6 = 3 > 2, ∀𝑘 ∈ ℕ
⇒ ∃𝑚𝑘 ∈ ℕ such that 𝑚𝑘 ∈ 𝐽𝑘 , ∀𝑘 ∈ ℕ
⇒ (sin(𝑚𝑘 )) is a subsequence of (sin(𝑛)) and by ①: lim sin(𝑚𝑘 ) = 𝑥 ………④
𝑘→∞
1
Since −1 ≤ sin(𝑚𝑘 ) ≤ − 2 , ∀𝑘 ∈ ℕ
1 1
⇒−1 ≤ lim sin(𝑚𝑘 ) ≤ − 2 and by ②: −1 ≤ 𝑥 ≤ − 2 ………⑤
𝑘→∞

③&⑤:
1 1
⇒ 2 ≤ 𝑥 ≤ 1 and −1 ≤ 𝑥 ≤ − 2 which is impossible (since the limit of a sequence is unique)
⇒ We have a contradiction ⇒ Our assumption is false
⇒ (sin(𝑛)) diverges.
 51

Example 3.4.8: Let 𝑋 = (𝑥𝑛 ) be a sequence of real numbers. There are two important subsequences
of 𝑋 which are (𝑥2𝑛 ) = (𝑥2 , 𝑥4 , 𝑥6 , … ) and (𝑥2𝑛−1 ) = (𝑥1 , 𝑥3 , 𝑥5 , … ).

Theorem 3.4.9: A sequence (𝑥𝑛 ) converges to 𝑥 iff [(𝑥2𝑛 ) and (𝑥2𝑛−1 ) both converge to 𝑥].
That is lim 𝑥𝑛 = 𝑥 iff [ lim 𝑥2𝑛 = 𝑥 and lim 𝑥2𝑛−1 = 𝑥]
𝑛→∞ 𝑛→∞ 𝑛→∞

Proof (Optional):
Suppose that (𝑥𝑛 ) converges to 𝑥 ⇒ All subsequences of 𝑋 converges to 𝑥 (by Theorem 3.4.4)
⇒ 𝑋 ′ = (𝑥2𝑛 ) and 𝑋 ′′ = (𝑥2𝑛−1 ) both converges to 𝑥
Suppose that both 𝑋 ′ = (𝑥2𝑛 ) and 𝑋 ′′ = (𝑥2𝑛−1 ) converges to 𝑥
To show that 𝑋 = (𝑥𝑛 ) converges to 𝑥 we use definition:
Let 𝜀 > 0 be given. Want to show that ∃𝐾 ∈ ℕ such that |𝑥𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾
lim 𝑥2𝑛 = 𝑥 ⇒ ∃𝐾1 ∈ ℕ such that |𝑥2𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾1
lim 𝑥2𝑛−1 = 𝑥 ⇒ ∃𝐾2 ∈ ℕ such that |𝑥2𝑛−1 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾2
Let 𝐾 = max(2𝐾1 , 2𝐾2 − 1).
If 𝑛 ≥ 𝐾 ⇒ 𝑛 ≥ 2𝐾1 & 𝑛 ≥ 2𝐾2 − 1
But 𝑛 is even or 𝑛 is odd ⇒ 𝑛 = 2𝑘1 or 𝑛 = 2𝑘1 − 1 for some 𝑘1 , 𝑘2 ∈ ℕ
 If 𝑛 = 2𝑘1 ⇒ 2𝑘1 ≥ 2𝐾1 ⇒ 𝑘1 ≥ 𝐾1 ⇒ |𝑥2𝑘1 − 𝑥| < 𝜀
⇒ |𝑥𝑛 − 𝑥| < 𝜀
 If 𝑛 = 2𝑘2 − 1 ⇒ 2𝑘2 − 1 ≥ 2𝐾2 − 1 ⇒ 𝑘2 ≥ 𝐾2
⇒ |𝑥2𝑘2 −1 − 𝑥| < 𝜀 ⇒ |𝑥𝑛 − 𝑥| < 𝜀
In either case we have |𝑥𝑛 − 𝑥| < 𝜀, ∀𝑛 ≥ 𝐾 ⇒ (𝑥𝑛 ) converges to 𝑥.

Corollary 3.4.10:
(1) A sequence (𝑥𝑛 ) diverges iff [(𝑥2𝑛 ) or (𝑥2𝑛−1 ) diverges].
(2) A sequence (𝑥𝑛 ) diverges iff lim 𝑥2𝑛 ≠ lim 𝑥2𝑛−1.
𝑛→∞ 𝑛→∞

Monotone Subsequence Theorem (MST) 3.4.11: Every sequence has a monotone subsequence.

The Bolzano-Weierstass Theorem (BWT) 3.4.12:

Every bounded sequence has a convergent subsequence.
Proof: Let (𝑥𝑛 ) be a bounded sequence.
MST ⇒ (𝑥𝑛 ) has a monotone subsequence (𝑥𝑛𝑘 ).
But (𝑥𝑛 ) is bounded ⇒ (𝑥𝑛𝑘 ) is bounded ⇒ (𝑥𝑛𝑘 ) is a monotone bounded sequence
MCT ⇒ (𝑥𝑛𝑘 ) converges ⇒ (𝑥𝑛 ) has a convergent subsequence.

Definition 3.4.13: Let 𝑋 = (𝑥𝑛 ) be a bounded sequence of real numbers.

(1) The limit inferior of (𝑥𝑛 ) (denoted by lim 𝑥𝑛 or lim inf 𝑥𝑛 ) is sup {𝑤𝑚 : 𝑚 ∈ ℕ}. That is
𝑛→∞ 𝑛→∞

lim 𝑥𝑛 = sup inf 𝑥𝑚 that is lim 𝑥𝑛 = sup inf {𝑥𝑛 , 𝑥𝑛+1 , 𝑥𝑛+2 , … }
𝑛→∞ 𝑛∈ℕ 𝑚≥𝑛 𝑛→∞ 𝑛∈ℕ
 52

(2) The limit superior of (𝑥𝑛 ) (denoted by lim 𝑥𝑛 or lim sup 𝑥𝑛 ) is defined by
𝑛→∞ 𝑛→∞

lim 𝑥𝑛 = inf sup 𝑥𝑚 that is lim 𝑥𝑛 = inf sup {𝑥𝑛 , 𝑥𝑛+1 , 𝑥𝑛+2 , … }
𝑛→∞ 𝑛∈ℕ 𝑚≥𝑛 𝑛→∞ 𝑛∈ℕ

Remark 3.4.15: Let 𝑋 = (𝑥𝑛 ) be a sequence and let 𝑋𝑚 = (𝑥𝑚+𝑛 ) be the 𝑚-tail of 𝑋,𝑚 = 0,1,2, …,
where 𝑋0 = 𝑋. Then
 lim 𝑥𝑛 = sup {inf 𝑋0 , inf 𝑋1 , inf 𝑋2 , … }
𝑛→∞
 lim 𝑥𝑛 = inf {sup 𝑋0 , sup 𝑋1 , sup 𝑋2 , … }
𝑛→∞
Theorem 3.4.16: Let (𝑥𝑛 ) be a bounded sequence of real numbers and let 𝑆 be the set of all
subsequential limits of (𝑥𝑛 ). Then
(𝑥𝑛 ) converges to 𝑥 iff lim 𝑥𝑛 = lim 𝑥𝑛 = 𝑥.
𝑛→∞ 𝑛→∞
1 1 1 1 1 1
Example 3.4.17: Let (𝑥𝑛 ) = (−1,1, − 2 , 2 , − 3 , 3 , − 4 , 4 , … )
(1) Find lim 𝑥𝑛 and lim 𝑥𝑛
𝑛→∞ 𝑛→∞
(2) Using part (1) determine whether the sequence (𝑥𝑛 ) converges or diverges. If it converges
find its limit.
Solution:
(1)
inf 𝑥𝑚 = inf 𝑋𝑚 sup 𝑥𝑚 = sup 𝑋𝑚
𝑚≥𝑛 𝑚≥𝑛

1 1 1 1 1 1
𝑋0 = (−1,1, − , , − , , − , , … ) −1 1
2 2 3 3 4 4
1 1 1 1 1 1 1
𝑋1 = (1, − , , − , , − , , … ) − 1
2 2 3 3 4 4 2
1 1 1 1 1 1 1 1
𝑋2 = (− , , − , , − , , … ) −
2 2 3 3 4 4 2 2
1 1 1 1 1 1 1
𝑋3 = ( , − , , − , , … ) −
2 3 3 4 4 3 2
1 1 1 1 1 1
𝑋4 = (− , , − , , … ) −
3 3 4 4 3 3
1 1 1 1 1
𝑋5 = ( , − , , … ) −
3 4 4 4 3
1 1 1 1
lim 𝑥𝑛 = sup {inf 𝑋0 , inf 𝑋1 , inf 𝑋2 , … } = sup {−1, − , − , − , − , … } = 0
𝑛→∞ 2 2 3 3
1 1 1 1
lim 𝑥𝑛 = inf {sup 𝑋0 , sup 𝑋1 , sup 𝑋2 , … } = inf {1,1, , , , , … } = 0
𝑛→∞ 2 2 3 3
(2) Since lim 𝑥𝑛 = lim 𝑥𝑛 = 0 ⇒ The sequence (𝑥𝑛 ) converges to 0, that .
𝑛→∞ 𝑛→∞
 53

1 1 1
Example 3.4.18: Let (𝑥𝑛 ) = (−1,1, −1, 2 , −1, 3 , −1, 4 , … )
(1) Find lim 𝑥𝑛 and lim 𝑥𝑛
𝑛→∞ 𝑛→∞
(2) Use part (1) to show that the sequence (𝑥𝑛 ) diverges.
Solution:
(3)
inf 𝑥𝑚 = inf 𝑋𝑚 sup 𝑥𝑚 = sup 𝑋𝑚
𝑚≥𝑛 𝑚≥𝑛

1 1 1 1 −1 1
𝑋0 = (−1,1, −1, , −1, , −1, , −1, , … )
2 3 4 5
1 1 1 1 −1 1
𝑋1 = (1, −1, , −1, , −1, , −1, , … )
2 3 4 5
1 1 1 1 1
𝑋2 = (−1, , −1, , −1, , −1, , … ) −1
2 3 4 5 2
1 1 1 1 1
𝑋3 = ( , −1, , −1, , −1, , … ) −1
2 3 4 5 2
1 1 1 1
𝑋4 = (−1, , −1, , −1, , … ) −1
3 4 5 3
1 1 1 1
𝑋5 = ( , −1, , −1, , … ) −1
3 4 5 3

lim 𝑥𝑛 = sup {inf 𝑋0 , inf 𝑋1 , inf 𝑋2 , … } = sup {−1, −1, −1, … } = −1

𝑛→∞
1 1 1 1
lim 𝑥𝑛 = inf {sup 𝑋0 , sup 𝑋1 , sup 𝑋2 , … } = inf {1,1, , , , , … } = 0
𝑛→∞ 2 2 3 3
(4) Since lim 𝑥𝑛 ≠ lim 𝑥𝑛 ⇒ The sequence (𝑥𝑛 ) diverges.
𝑛→∞ 𝑛→∞

Example 3.4.19:
(1) Find lim (−1)𝑛 and lim (−1)𝑛
𝑛→∞ 𝑛→∞
(2) Use part (1) to determine to show that the sequence ((−1)𝑛 ) diverges.
Solution:
(1)
inf 𝑥𝑚 = inf 𝑋𝑚 sup 𝑥𝑚 = sup 𝑋𝑚
𝑚≥𝑛 𝑚≥𝑛

𝑋0 = (−1,1, −1,1, −1,1, … ) −1 1

𝑋1 = (1, −1,1, −1,1, −1,1, … ) −1 1

𝑋2 = (−1,1, −1,1, −1,1, … ) −1 1

𝑋3 = (1, −1,1, −1,1, … ) −1 1
𝑋4 = (−1,1, −1,1, −1,1, … ) −1 1
𝑋5 = (1, −1,1, −1,1, … ) −1 1
 54

lim 𝑥𝑛 = sup {inf 𝑋0 , inf 𝑋1 , inf 𝑋2 , … } = sup {−1, −1, −1, … } = −1

𝑛→∞
lim 𝑥𝑛 = inf {sup 𝑋0 , sup 𝑋1 , sup 𝑋2 , … } = inf {1,1,1, … } = 1
𝑛→∞
(2) Since lim (−1)𝑛 ≠ lim ⇒ The sequence ((−1)𝑛 ) diverges
𝑛→∞ 𝑛→∞

2𝑛2 +1 2𝑛2 +1
Example 3.4.20: Find lim and lim .
𝑛→∞ 𝑛2 +𝑛−1 𝑛2 +𝑛−1
𝑛→∞

2𝑛2 +1 4𝑛 4
Solution: Since lim = lim = lim = 2 we have:
𝑛→∞ 𝑛2 +𝑛−1 𝑛→∞ 2𝑛+1 𝑛→∞ 2
2𝑛2 +1
The sequence (𝑛2 +𝑛−1) converges to 2 and so

2𝑛2 +1 2𝑛2 +1 2𝑛2 +1

lim 𝑛2 +𝑛−1
= lim 𝑛2 +𝑛−1
= lim = 2.
𝑛→∞ 𝑛→∞ 𝑛→∞ 𝑛2 +𝑛−1

Exercises 3.4.21 (pages 84+85): 1,3,4,9,11

 55

Section 3.5: The Cauchy Criterion

Definition 3.5.1:
1) A sequence 𝑋 = (𝑥𝑛 ) of real numbers is said to be a Cauchy sequence
① ② ③

∀𝜀 > 0, ⏞
if ⏞ ∃𝐻(𝜀) ∈ ℕ such that ⏞|𝑥𝑚 − 𝑥𝑛 | < 𝜀, ∀𝑚 > 𝑛 ≥ 𝐻(𝜀), where 𝑚, 𝑛 ∈ ℕ.
2) A sequence 𝑋 = (𝑥𝑛 ) of real numbers is not a Cauchy sequence
① ② ③

if ⏞ ∀𝐻(𝜀) ∈ ℕ and ⏞
∃𝜀0 > 0, ⏞ ∃𝑚0 > 𝑛0 ≥ 𝐻(𝜀) such that |𝑥𝑚0 − 𝑥𝑛0 | ≥ 𝜀0 .

Note: The value of the natural number 𝐻(𝜀) is usually depends on the value of 𝜀. For simplicity,
we will write 𝐻 instead of writing 𝐻(𝜀)

Example 3.5.2: Using the definition of Cauchy sequences, show that the sequence
1
𝑋 = (𝑛) is a Cauchy sequence.
1
Solution: Let 𝜀 > 0 be given and let 𝑥𝑛 = 𝑛.

Want to find 𝐻 ∈ ℕ such that |𝑥𝑚 − 𝑥𝑛 | < 𝜀, ∀𝑚 > 𝑛 ≥ 𝐻

1 1 1 1 1 1
Let 𝑚 > 𝑛: |𝑥𝑚 − 𝑥𝑛 | = | − | ≤
⏟ |𝑚| + |𝑛| = + ………..①
𝑚 𝑛 𝑚 𝑛
by the triangle inequality

1 1 𝜀 1 𝜀
Since lim 𝑛 = 0: ∃𝐾 ∈ ℕ such that |𝑛| < 2 , ∀𝑛 ≥ 𝐾 ⇒ 𝑛 < 2 , ∀𝑛 ≥ 𝐾 ……….②

Take 𝐻 = 𝐾: Let 𝑚 > 𝑛 ≥ 𝐻 ⇒ 𝑚 ≥ 𝐾 and 𝑛 ≥ 𝐾 ⇒ 𝑚 and 𝑛 satisfy ②

1 𝜀 1 𝜀
⇒ 𝑚 < 2 and 𝑛 < 2 ………….③
1 1 𝜀 𝜀
①&③: ⇒ |𝑥𝑚 − 𝑥𝑛 | ≤ 𝑚 + 𝑛 < 2 + 2 = 𝜀, ∀𝑚 > 𝑛 ≥ 𝐻
1
⇒ |𝑥𝑚 − 𝑥𝑛 | < 𝜀, ∀𝑚 > 𝑛 ≥ 𝐻 ⇒ (𝑛) is a Cauchy sequence.

Example 3.5.3: Using the definition of Cauchy sequences, show that the sequence
𝑋 = ((−1)𝑛 ) is not a Cauchy sequence.
Solution: Let 𝜀0 = 1 > 0 be given and 𝐻 ∈ ℕ. Want to find
𝑚0 > 𝑛0 ≥ 𝐻 such that |(−1)𝑚0 − (−1)𝑛0 | ≥ 𝜀0 = 1
Take 𝑚0 = 2𝐻 and 𝑛0 = 2𝐻 − 1 ⇒ 𝑚0 > 𝑛0 ≥ 𝐻 and
|(−1)𝑚0 − (−1)𝑛0 | = |(−1)2𝐻 − (−1)2𝐻−1 | = |1 − (−1)| = 2 > 1 = 𝜀0
⇒ ∃𝑚0 > 𝑛0 ≥ 𝐻 such that |(−1)𝑚0 − (−1)𝑛0 | ≥ 𝜀0
∴ 𝑋 = ((−1)𝑛 ) is not a Cauchy sequence.
 56

Remark 3.5.4:
(1) A sequence (𝑥𝑛 ) is bounded iff ∃𝑀 > 0 such that |𝑥𝑚 | ≤ 𝑀, ∀𝑚 ∈ ℕ.
(2) The triangle inequality: |𝑎 − 𝑏| ≤ |𝑎| + |𝑏|

Theorem 3.5.5: Every Cauchy sequence is bounded.

Proof: Let (𝑥𝑛 ) be a Cauchy sequence.

In the definition of Cauchy sequence: Take 𝜀 = 1 ⇒ ∃𝐻 ∈ ℕ such that
|𝑥𝑚 − 𝑥𝑛 | < 1, ∀𝑚 > 𝑛 ≥ 𝐻…………①
In ①: take 𝑛 = 𝐻 ⇒ |𝑥𝑚 − 𝑥𝐻 | < 1, ∀𝑚 ≥ 𝐻………②
Let 𝑀 = max{ |𝑥1 |, |𝑥2 |, … , |𝑥𝐻 |, 1 + |𝑥𝐻 |}
⇒ |𝑥𝑚 | ≤ 𝑀, ∀𝑚 = 1,2, … , 𝐻 ……….③
⇒ |𝑥𝑚 | = |𝑥𝑚 − 𝑥𝐻 + 𝑥𝐻 |
≤ |𝑥𝑚 − 𝑥𝐻 | + |𝑥𝐾 |
⏟
by the triangle inequality
⏟ 1 + |𝑥𝐻 | , ∀𝑚 > 𝐻
<
by ②
≤ 𝑀, ∀𝑚 > 𝐻
⇒ |𝑥𝑚 | < 𝑀, ∀𝑚 > 𝐻 …..④

③&④: ⇒ |𝑥𝑚 | ≤ 𝑀, ∀𝑚 ∈ ℕ ⇒ (𝑥𝑛 ) is bounded.

Bolzano-Weierstrass Theorem 3.5.6:

Every bounded sequence has a convergent subsequence.

Cauchy Convergence Criterion 3.5.7:

Let (𝑥𝑛 ) be a sequence of real numbers. Then
(𝑥𝑛 ) converges iff (𝑥𝑛 ) is a Cauchy sequence.
Proof:
Suppose that (𝑥𝑛 ) converges to 𝑥 that is lim 𝑥𝑛 = 𝑥. To show that (𝑥𝑛 ) is a Cauchy sequence:
Let 𝜀 > 0 be given.
Want to show that ∃𝐻 ∈ ℕ such that |𝑥𝑚 − 𝑥𝑛 | < 𝜀, ∀𝑚 > 𝑛 ≥ 𝐻.
Let 𝑚 > 𝑛: |𝑥𝑚 − 𝑥𝑛 | = |𝑥𝑚 − 𝑥 + 𝑥 − 𝑥𝑛 | ≤ |𝑥𝑚 − 𝑥| + |𝑥 − 𝑥𝑛 |……….①
𝜀
Since lim 𝑥𝑛 = 𝑥: ⇒ ∃𝐾 ∈ ℕ such that |𝑥𝑛 − 𝑥| < 2 , ∀𝑛 ≥ 𝐾……….②
𝑛→∞
Take 𝐻 = 𝐾: Let 𝑚 > 𝑛 ≥ 𝐻 ⇒ 𝑚 ≥ 𝐾 and 𝑛 ≥ 𝐾 ⇒ 𝑚 and 𝑛 satisfy ②
𝜀 𝜀
⇒ |𝑥𝑚 − 𝑥| < 2 and |𝑥𝑛 − 𝑥| < 2 ………….③
𝜀 𝜀
①&③: ⇒ |𝑥𝑚 − 𝑥𝑛 | ≤ |𝑥𝑚 − 𝑥| + |𝑥 − 𝑥𝑛 | < 2 + 2 = 𝜀, ∀𝑚 > 𝑛 ≥ 𝐻

⇒ |𝑥𝑚 − 𝑥𝑛 | < 𝜀, ∀𝑚 > 𝑛 ≥ 𝐻

∴ (𝑥𝑛 ) is a Cauchy sequence.
 57

Suppose that (𝑥𝑛 ) is a Cauchy sequence. Want to show that (𝑥𝑛 ) converges.
Let (𝑥𝑛 ) be a Cauchy sequence ⇒ (𝑥𝑛 ) is bounded (by Theorem 4)
Bolzano-Weierstrass Theorem ⇒ 𝑋 = (𝑥𝑛 ) has a convergent subsequence 𝑋 ′ = (𝑥𝑛𝑘 )
that is lim 𝑥𝑛𝑘 = 𝑥 ⇒ Now, we show that lim 𝑥𝑛 = 𝑥
𝑘→∞ 𝑛→∞
Let 𝜀 > 0 be given.
Since (𝑥𝑛 ) is Cauchy seq.: ∃𝐻 ∈ ℕ such that
𝜀
|𝑥𝑚 − 𝑥𝑛 | < , ∀𝑚 > 𝑛 ≥ 𝐻………④
3
𝜀
Since lim 𝑥𝑛𝑘 = 𝑥: ⇒ ∃𝐾1 ∈ ℕ such that |𝑥𝑛𝑘 − 𝑥| < 3 , ∀𝑘 ≥ 𝐾1 …..⑤
𝑘→∞
Let 𝐾 = max{𝐻, 𝐾1 } and let 𝑚 ≥ 𝐾 ⇒ 𝑚 ≥ 𝐻 and 𝑚 ≥ 𝐾1
𝜀
In ④: Take 𝑛 = 𝐾⇒ |𝑥𝑚 − 𝑥𝐾 | < 3……….⑥
𝜀
In ⑤: Take 𝑘 = 𝐾⇒ |𝑥𝑛𝐾 − 𝑥| < 3………⑦
Also, since 𝑛𝐾 ≥ 𝐾 so in ④: taking 𝑚 = 𝑛𝐾 and 𝑛 = 𝐾 we have:
𝜀
|𝑥𝑛𝐾 − 𝑥𝐾 | < 3 ………⑧
Now,
|𝑥𝑚 − 𝑥| = |𝑥𝑚 − 𝑥𝐾 + 𝑥𝐾 − 𝑥𝑛𝐾 + 𝑥𝑛𝐾 − 𝑥|
⏟ |𝑥𝑚 − 𝑥𝐾 | + |𝑥𝐾 − 𝑥𝑛𝐾 | + |𝑥𝑛𝐾 − 𝑥|
≤
by the triangle inequality
= |𝑥𝑚 − 𝑥𝐾 | + ⏟
|𝑥𝑛𝐾 − 𝑥𝐾 | + |𝑥𝑛𝐾 − 𝑥|
since |𝑎|=|−𝑎|
𝜀 𝜀 𝜀
< + +
⏟
3 ⏟
3 ⏟
3
by ⑥ by ⑧ by ⑦
=𝜀
⇒ |𝑥𝑚 − 𝑥| < 𝜀, ∀𝑚 ≥ 𝐾 ⇒ 𝑋 converges.

Example 3.5.8: Which of the following is a Cauchy sequence:

𝑛 2 3𝑛
1) ((−1)𝑛 ) 2) (𝑛+1) 3) ((1 − 𝑛) ) 4) (−√𝑛)

Solution:
1) ((−1)𝑛 ) diverges ⇒ ((−1)𝑛 ) is not a Cauchy sequence.
𝑛 𝑛
2) (𝑛+1) converges to 1 ⇒ (𝑛+1) is a Cauchy sequence.
2 3𝑛 2 3𝑛
3) ((1 − 𝑛) ) converges to 𝑒 −6 ⇒ ((1 − 𝑛) ) is a Cauchy sequence.
4) (−√𝑛) diverges ⇒ (−√𝑛) is not a Cauchy sequence.

Definition 3.5.9: Let 𝑋 = (𝑥𝑛 ) be a sequence of real numbers. We say that a sequence 𝑋 of real
numbers is contractive if ∃𝐶 ∈ (0,1) such that
|𝑥𝑛+2 − 𝑥𝑛+1 | ≤ 𝐶|𝑥𝑛+1 − 𝑥𝑛 |, ∀𝑛 ∈ ℕ.
The number 𝐶 is called the constant of the contractive sequence.
 58

Theorem 3.5.10: Every contractive sequence is a Cauchy sequence, and therefore is convergent.

Example 3.5.11: Let (𝑥𝑛 ) be defined by:

𝑥𝑛−1 +𝑥𝑛−2
𝑥1 = 1, 𝑥2 = 2 and 𝑥𝑛 = ,𝑛 ≥ 2
2

(1) Show that (𝑥𝑛 ) is contractive.

(2) Show that (𝑥𝑛 ) converges
2 1
(3) Find the limit point of (𝑥𝑛 ) (Hint: 𝑥2𝑛+1 = 1 + 3 (1 − 4𝑛) , 𝑛 ∈ ℕ)

Solution:
𝑥 +𝑥
1) |𝑥𝑛+2 − 𝑥𝑛+1 | = | 𝑛+12 𝑛 − 𝑥𝑛+1 |
𝑥𝑛+1 + 𝑥𝑛 2𝑥𝑛+1
=| − |
2 2
𝑥𝑛 − 𝑥𝑛+1 1
=| | = |𝑥𝑛+1 − 𝑥𝑛 |, ∀𝑛 ∈ ℕ
2 2
1
(𝑥𝑛 ) is contractive with 𝐶 = 2 ∈ (0,1).

2) Since (𝑥𝑛 ) is contractive ⇒ (𝑥𝑛 ) is a Cauchy seq. ⇒ (𝑥𝑛 ) converges.

3) Let lim 𝑥𝑛 = 𝑥. Want to find 𝑥:
𝑛→∞
𝑥𝑛−1 +𝑥𝑛−2 lim 𝑥𝑛−1 + lim 𝑥𝑛−2
𝑥𝑛 = , ∀𝑛 ≥ 2 ⇒ lim 𝑥𝑛 = 𝑛→∞ 𝑛→∞
2 2
𝑥+𝑥
⇒𝑥= = 𝑥 (we have no information 😡😡😡)
2
We use subsequences: Let 𝑋 ′ = (𝑥2𝑛+1 ) be a subsequence 𝑋.
Since (𝑥𝑛 ) converges to 𝑥 ⇒ 𝑋 ′ converges to 𝑥 ⇒ lim 𝑥2𝑛+1 = 𝑥
2 1
Using the hint: 𝑥2𝑛+1 = 1 + 3 (1 − 4𝑛)
2 1 2 5
⇒ lim 𝑥2𝑛+1 = lim (1 + 3 (1 − 4𝑛 )) = 1 + 3 (1 − 0) = 3
𝑛→∞ 𝑛→∞
5 5
⇒ 𝑥 = 3 ⇒ limit point of (𝑥𝑛 ) is 3.
 59

Example 3.5.12: We consider the sequence (𝑥𝑛 ) of Fibonacci fractions that is

𝑓
𝑥𝑛 = 𝑓 𝑛 , 𝑛 ∈ ℕ, where 𝑓1 = 𝑓2 = 1 and 𝑓𝑛 = 𝑓𝑛−1 + 𝑓𝑛−2 , 𝑛 ≥ 2.
𝑛+1

1
(1) Show that 𝑥𝑛+1 = 1+𝑥 , 𝑛 ∈ ℕ.
𝑛
1
(2) Show that 2 ≤ 𝑥𝑛 ≤ 1, 𝑛 ∈ ℕ.
(3) Show that (𝑥𝑛 ) is contractive.
(4) Show that (𝑥𝑛 ) converges
(5) Find lim 𝑥𝑛 .
𝑛→∞
Solution:
𝑓𝑛+1
𝑓𝑛+1 𝑓𝑛+1 𝑓𝑛+1 1
1) 𝑥𝑛+1 = 𝑓 =𝑓 = 𝑓𝑛+1 𝑓 = 1+𝑥 , 𝑛 ∈ ℕ
𝑛+2 𝑛+1 +𝑓𝑛 + 𝑛 𝑛
𝑓𝑛+1 𝑓𝑛+1
1
2) To show that 2 ≤ 𝑥𝑛 ≤ 1, 𝑛 ∈ ℕ, we use the proof by induction:
1
Base step: At 𝑛 = 1: 𝑥1 = 1 ⇒ 2 ≤ 𝑥1 ≤ 1 is true
Inductive Step:
 Suppose that the inequalities is true at 𝑛 = 𝑘 that is
1
≤ 𝑥𝑘 ≤ 1………..①
2
1
 At 𝑛 = 𝑘 + 1: 𝑥𝑘+1 = 1+𝑥 (by part (1))
𝑘
⇒ 𝑥𝑘+1 < 1……..②
1 1 1
Also, 𝑥𝑘 ≤ 1 ⇒ 1 + 𝑥𝑘 ≤ 2 ⇒ 1+𝑥 ≥ 2 ⇒ 𝑥𝑘+1 ≥ 2………③
𝑘
1
②&③ ⇒ 2 ≤ 𝑥𝑘+1 ≤ 1 ⇒ the inequalities is true at 𝑛 = 𝑘 + 1
1
∴ 2 ≤ 𝑥𝑛 ≤ 1, 𝑛 ∈ ℕ
1 1 1+𝑥 −(1+𝑥 )
3) |𝑥𝑛+2 − 𝑥𝑛+1 | = |1+𝑥 − 1+𝑥 | = |(1+𝑥𝑛 )(1+𝑥
𝑛+1
|
𝑛+1 𝑛 𝑛+1 𝑛)
𝑥𝑛 − 𝑥𝑛+1
=| |
(1 + 𝑥𝑛+1 )(1 + 𝑥𝑛 )
|𝑥𝑛 −𝑥𝑛+1 | |𝑥𝑛+1 −𝑥𝑛 |
= = , ∀𝑛 ∈ ℕ…….①
(1+𝑥𝑛+1 )(1+𝑥𝑛 ) (1+𝑥𝑛+1 )(1+𝑥𝑛 )

1 3
Part (2) ⇒ 𝑥𝑛 ≥ 2 , ∀𝑛 ∈ ℕ ⇒ 1 + 𝑥𝑛 ≥ 2 , ∀𝑛 ∈ ℕ
1 2
⇒ 1+𝑥 ≤ 3 , ∀𝑛 ∈ ℕ……..②
𝑛

1 1
①&② ⇒ |𝑥𝑛+2 − 𝑥𝑛+1 | = 1+𝑥 |𝑥𝑛+1 − 𝑥𝑛 |
𝑛+1 1+𝑥𝑛

22
≤ |𝑥 − 𝑥𝑛 |
3 3 𝑛+1
4
= |𝑥 − 𝑥𝑛 |, ∀𝑛 ∈ ℕ
9 𝑛+1
4
(𝑥𝑛 ) is contractive with 𝐶 = 9 ∈ (0,1).

4) Since (𝑥𝑛 ) is contractive ⇒ (𝑥𝑛 ) is a Cauchy seq. ⇒ (𝑥𝑛 ) converges.

 60

5) To find lim 𝑥𝑛 : Let lim 𝑥𝑛 = 𝑥. Want to find 𝑥:

𝑛→∞
1 1 1
𝑥𝑛+1 = 1+𝑥 , ∀𝑛 ∈ ℕ ⇒ lim 𝑥𝑛+1 = 1+lim 𝑥 ⇒ 𝑥 = 1+𝑥
𝑛 𝑛

2 −1±√1−4(1)(−1) −1±√5
⇒ 𝑥(1 + 𝑥) = 1 ⇒ 𝑥 + 𝑥 − 1 = 0 ⇒ 𝑥 = =
2 2
−1+√5 −1−√5
⇒𝑥= or 𝑥 =
2 2
1 1 1
But 2 ≤ 𝑥𝑛 ≤ 1, 𝑛 ∈ ℕ ⇒ ≤ lim 𝑥𝑛 ≤ 1, 𝑛 ∈ ℕ ⇒ ≤ 𝑥 ≤ 1, 𝑛 ∈ ℕ
2 2
−1+√5
∴𝑥= 2

The Golden Ratio (‫ )النسبة الذهبية‬3.5.13:

1 2 1+√5
The ratio 𝜑 = 𝑥 = −1+√5 = = 1.618043 … … .. is called the golden ratio or the
2
1 1 𝑥
golden number, where 𝑥 satisfies 𝑥 = 1+𝑥 ⇔ 𝑥 = 1−𝑥
𝑎
If 𝑎, 𝑏 are positive real numbers such that = 1.618043 we have a perfect situation
𝑏
between 𝑎 and 𝑏.

Nice videos that talk about relation between the Fibonacci sequence and the Golden ratio and the
golden spiral are with the following links:
1) https://youtu.be/qTw_qay54WI
2) https://youtu.be/SjSHVDfXHQ4

Exercises 3.5.14 (page 91): 1,2(a),3(a),4,12,13

 61

Section 3.6: Properly Divergent Sequences

For certain purposes it is convenient to define what is meant for a sequence (𝑥𝑛 ) of real numbers
to ‘‘tend to ±∞.’’
Definition 3.6.1: Let (𝑥𝑛 ) be a sequence of real numbers.
① ②

(1) ⏞ > 0, ⏞
We say that (𝑥𝑛 ) tends to +∞, written as lim 𝑥𝑛 = +∞ , if 𝛼 ∃𝐾(𝛼) ∈ ℕ such
𝑛→∞
③

that ⏞
𝑥𝑛 > 𝛼, ∀𝑛 ≥ 𝐾(𝛼).
① ②

(2) ⏞ > 0, ⏞
We say that (𝑥𝑛 ) tends to −∞, written as lim 𝑥𝑛 = −∞ , if 𝛽 ∃𝐾(𝛽) ∈ ℕ such
𝑛→∞
③

that ⏞
𝑥𝑛 < 𝛽, ∀𝑛 ≥ 𝐾(𝛽).
(3) We say that (𝑥𝑛 ) is properly divergent if either: lim 𝑥𝑛 = +∞ or lim 𝑥𝑛 = −∞.
𝑛→∞ 𝑛→∞

Remark 3.6.2:
1) lim 𝑥𝑛 = −∞ iff lim (−𝑥𝑛 ) = +∞
𝑛→∞ 𝑛→∞
So, we will focus our study on sequences (𝑥𝑛 ) with lim 𝑥𝑛 = +∞
𝑛→∞
2) The value of the natural number 𝐾(𝛼) is usually depends on the value of 𝛼. For
simplicity, we will write 𝐾 instead of writing 𝐾(𝛼).
3) To find the value of 𝐾 ∈ ℕ, we start as follows:
Let 𝛼 > 0 be given
Study 𝑥𝑛 by: 𝑥𝑛 > ⋯ > 𝑎𝑛𝑏 or 𝑎(𝑏 𝑛 ) with 𝑎, 𝑏 > 0 (‫)نقوم بعمل تصغير للمقدار‬
 Take 𝑎(𝑏 𝑛 ) > 𝛼 or 𝑎𝑛𝑏 > 𝛼 ⇒ 𝑛 > 𝑀
Apply the Archimedean Property: ∃𝐾 ∈ ℕ such that 𝐾 > 𝑀
⇒ 𝑥𝑛 > 𝛼 ∀𝑛 ≥ 𝐾

Example 3.6.3:
7𝑛4 −3𝑛+1
(1) Using definition, show that lim = ∞.
𝑛→∞ 𝑛+1
(2) Based on your calculations in Part (1), Find the smallest value of 𝐾 that satisfies the
7𝑛4 −3𝑛+1
definition of lim = +∞ by taking 𝛼 = 54.
𝑛→∞ 𝑛+1
Solution:
7𝑛4 −3𝑛+1
(1) Let 𝛼 > 0 be given. Let 𝑥𝑛 = want to show that: 𝑥𝑛 > ⋯ > 𝑎𝑛𝑏 :
𝑛+1

7𝑛4 −3𝑛+1 7𝑛4 −3𝑛 7𝑛4 −3𝑛4 4𝑛4 4𝑛4 4𝑛4

𝑥𝑛 = ≥ ≥ = 𝑛+1 ≥ 𝑛+𝑛 = = 2𝑛3
𝑛+1 𝑛+1 𝑛+1 2𝑛

⇒ 𝑥𝑛 ≥ 2𝑛3 ……..①
3 𝛼
Take 2𝑛3 > 𝛼 ⇔ 𝑛 > √ 2 ………②
 62

3 𝛼
By the Archimedean Property: ∃𝐾 ∈ ℕ such that 𝐾 > √ 2 .

3 𝛼
If 𝑛 ≥ 𝐾 ⇒ 𝑛 > √ 2 ⇒ 2𝑛3 > 𝛼 (by ②) ⇒ 𝑥𝑛 > 𝛼 (by ①)

⇒ 𝑥𝑛 > 𝛼 ∀𝑛 ≥ 𝐾
7𝑛4 −3𝑛+1
⇒ lim = +∞.
𝑛→∞ 𝑛+1
3 𝛼 3 54
(2) By Part (1), 𝐾 ∈ ℕ such that 𝐾 > √ 2 ⇒ 𝐾 > √ 2 = 3 ⇒ 𝐾 > 3
⇒ The smallest value of 𝐾 is 𝐾 = 4.

Example 3.6.4:
(1) Using definition, show that lim 𝑛2 = ∞.
𝑛→∞
(2) Based on your calculations in Part (1), Find the smallest value of 𝐾 that satisfies the
definition of lim 𝑛2 = +∞ by taking 𝛼 = 16.
𝑛→∞
Solution:
(1) Let 𝛼 > 0 be given. Let 𝑥𝑛 = 𝑛2 want to show that: 𝑥𝑛 > ⋯ > 𝑎𝑛𝑏 :
𝑥𝑛 = 𝑛2 ……..①

Take 𝑛2 > 𝛼 ⇔ 𝑛 > √𝛼 ………②

By the Archimedean Property: ∃𝐾 ∈ ℕ such that 𝐾 > √𝛼 .

If 𝑛 ≥ 𝐾 ⇒ 𝑛 > √𝛼 ⇒ 𝑛2 > 𝛼 (by ②) ⇒ 𝑥𝑛 > 𝛼 (by ①)

⇒ 𝑥𝑛 > 𝛼 ∀𝑛 ≥ 𝐾
⇒ lim 𝑛2 = ∞.
𝑛→∞
(2) By Part (1), 𝐾 ∈ ℕ such that 𝐾 > √𝛼 ⇒ 𝐾 > √16 = 4 ⇒ 𝐾 > 4
⇒ The smallest value of 𝐾 is 𝐾 = 5.

Example 3.6.5: Using definition, show that if 𝑐 > 1, then lim 𝑐 𝑛 = ∞.

𝑛→∞

Solution: Let 𝛼 > 0 be given. Let 𝑥𝑛 = 𝑐 𝑛 :

ln 𝛼
𝑥𝑛 > 𝛼 ⇔ 𝑐 𝑛 > 𝛼 ⇔ ln 𝑐 𝑛 > ln 𝛼 ⇔ 𝑛 ln 𝑐 > ln 𝛼 ⇔ 𝑛 > ……..①
ln 𝑐
ln 𝛼
By the Archimedean Property: ∃𝐾 ∈ ℕ such that 𝐾 > .
ln 𝑐

ln 𝛼
If 𝑛 ≥ 𝐾 ⇒ 𝑛 > ⇒ 𝑥𝑛 > 𝛼 (by ①) ⇒ 𝑥𝑛 > 𝛼 ∀𝑛 ≥ 𝐾
ln 𝑐

⇒ lim 𝑐 𝑛 = ∞.
𝑛→∞
 63

Theorem 3.6.6: Let (𝑥𝑛 ) be a monotone sequence of real numbers. Then

(𝑥𝑛 ) is properly divergent iff (𝑥𝑛 ) is unbounded.
(a) If (𝑥𝑛 ) is an unbounded increasing sequence, then lim 𝑥𝑛 = +∞.
𝑛→∞
(b) If (𝑥𝑛 ) is an unbounded decreasing sequence, then lim 𝑥𝑛 = −∞
𝑛→∞
Proof: Let (𝑥𝑛 ) is a monotone sequence.
Suppose that (𝑥𝑛 ) is properly divergent
Assume that (𝑥𝑛 ) is bounded.
⇒ (𝑥𝑛 ) is bounded and monotone
The MCT ⇒ (𝑥𝑛 ) converges which is a contradiction since (𝑥𝑛 ) is properly divergent
∴ our assumption is false ⇒ (𝑥𝑛 ) is unbounded

Suppose that (𝑥𝑛 ) is unbounded.

Let 𝛼 > 0 be given.
Since (𝑥𝑛 ) is monotone, we have two cases:
(𝑥𝑛 ) is increasing or (𝑥𝑛 ) is decreasing:

Case 1: If (𝑥𝑛 ) is increasing sequence ⇒ (𝑥𝑛 ) increasing and unbounded.

(𝑥𝑛 ) is an increasing ⇒ 𝑥1 ≤ 𝑥2 ≤ 𝑥3 ≤ ⋯ ⇒ 𝑥1 is a LB of (𝑥𝑛 ).
(𝑥𝑛 ) is unbounded ⇒ (𝑥𝑛 ) has no UB
Since 𝛼 > 0 we have: ∃𝐾 ∈ ℕ such that 𝑥𝐾 > 𝛼……….①
But (𝑥𝑛 ) is an increasing ⇒ 𝑥𝑛 ≥ 𝑥𝐾 , ∀𝑛 ≥ 𝐾……..②
①&② ⇒ 𝑥𝑛 > 𝛼, ∀𝑛 ≥ 𝐾
⇒ lim 𝑥𝑛 = +∞
𝑛→∞
∴ (𝑥𝑛 ) is properly divergent

Case 2: Let (𝑥𝑛 ) be sequence ⇒ (𝑥𝑛 ) is decreasing and unbounded

⇒ (−𝑥𝑛 ) is increasing and unbounded
Case 1 ⇒ lim (−𝑥𝑛 ) = +∞ ⇒ − lim 𝑥𝑛 = +∞ ⇒ lim 𝑥𝑛 = −∞
𝑛→∞ 𝑛→∞ 𝑛→∞
⇒ (𝑥𝑛 ) is properly divergent

Theorem 3.6.7: Let (𝑥𝑛 ) be a sequence of positive numbers. Then

1
(1) lim 𝑥𝑛 = 0 iff lim 𝑥 = +∞
𝑛→∞ 𝑛→∞ 𝑛
1
(2) lim = 0 iff lim 𝑥𝑛 = +∞
𝑛→∞ 𝑥𝑛 𝑛→∞
Proof:
(1)
Suppose that lim 𝑥𝑛 = 0.
𝑛→∞
1
Want to show that lim = +∞ by using definition: Let 𝛼 > 0 be given
𝑛→∞ 𝑥𝑛
1
In the definition of lim 𝑥𝑛 = 0: Take 𝜀 = 𝛼 ⇒ ∃𝐾 ∈ ℕ such that
𝑛→∞
1 1 1
|𝑥𝑛 − 0| < , ∀𝑛 ≥ 𝐾 ⇒ |𝑥𝑛 | < , ∀𝑛 ≥ 𝐾 ⇒ 𝑥𝑛 < , ∀𝑛 ≥ 𝐾
𝛼 𝛼 𝛼
1 1
⇒ > 𝛼, ∀𝑛 ≥ 𝐾 ⇒ lim = +∞
𝑥𝑛 𝑛→∞ 𝑥𝑛
1
Suppose that lim = +∞.
𝑛→∞ 𝑥𝑛
 64

Want to show that lim 𝑥𝑛 = 0 by using definition: Let 𝜀 > 0 be given

𝑛→∞
1 1
In the definition of lim = +∞: take 𝛼 = 𝜀 ⇒ ∃𝐾 ∈ ℕ such that
𝑥𝑛
1 1
|𝑥𝑛 | < 𝜀, ∀𝑛 ≥ 𝐾
> 𝛼 = 𝜀 , ∀𝑛 ≥ 𝐾 ⇒ 𝑥𝑛 < 𝜀, ∀𝑛 ≥ 𝐾 ⇒ ⏟
𝑥𝑛
since 𝑥𝑛 >0,∀𝑛∈ℕ
⇒ |𝑥𝑛 − 0| < 𝜀, ∀𝑛 ≥ 𝐾 ⇒ lim 𝑥𝑛 = 0
𝑛→∞
1 1
(2) lim = 0 iff lim 1 = +∞ iff lim 𝑥𝑛 = +∞
𝑛→∞ 𝑥𝑛 𝑛→∞ (𝑥 ) 𝑛→∞
⏟ 𝑛
by Part (1)

Recall that 3.6.8:

1) If lim 𝑥𝑛 = 𝐿, then lim |𝑥𝑛 | = |𝐿|
𝑛→∞ 𝑛→∞
2) lim 𝑥𝑛 = 0 iff lim |𝑥𝑛 | = 0
𝑛→∞ 𝑛→∞
3) If lim 𝑥𝑛 = +∞ or lim 𝑥𝑛 = −∞, then lim |𝑥𝑛 | = +∞
𝑛→∞ 𝑛→∞ 𝑛→∞

Theorem 3.6.9: Let (𝑥𝑛 ) and (𝑦𝑛 ) be a sequences of nonzero real numbers such that (𝑥𝑛 ) is
properly divergent and lim 𝑥𝑛 𝑦𝑛 = 𝐿 ∈ ℝ. Then lim 𝑦𝑛 = 0
𝑛→∞ 𝑛→∞

Proof: Since (𝑥𝑛 ) is properly divergent ⇒ lim 𝑥𝑛 = +∞ or lim 𝑥𝑛 = −∞

𝑛→∞ 𝑛→∞
1
⇒ lim |𝑥𝑛 | = +∞ ⇒ lim |𝑥 | = 0 (by Theorem 4)
𝑛→∞ 𝑛→∞ 𝑛
|𝑥𝑛 𝑦𝑛 | 1 1
⇒ lim |𝑦𝑛 | = lim = lim |𝑥𝑛 𝑦𝑛 | (|𝑥 |) = ( lim |𝑥𝑛 𝑦𝑛 |) ( lim ) = |𝐿| × 0 = 0
𝑛→∞ 𝑛→∞ |𝑥𝑛 | 𝑛→∞ 𝑛 𝑛→∞ 𝑛→∞ |𝑥𝑛 |
⇒ lim |𝑦𝑛 | = 0 ⇒ lim 𝑦𝑛 = 0
𝑛→∞ 𝑛→∞

Comparison Test 3.6.10: Let (𝑥𝑛 ) and (𝑦𝑛 ) be two sequences of real numbers such that
𝑥𝑛 ≤ 𝑦𝑛 , ∀𝑛 ≥ 𝐾.
(1) If lim 𝑥𝑛 = +∞, then lim 𝑦𝑛 = +∞
𝑛→∞ 𝑛→∞
(2) If lim 𝑦𝑛 = −∞, then lim 𝑥𝑛 = −∞
𝑛→∞ 𝑛→∞
Proof:
(1) Suppose that lim 𝑥𝑛 = +∞. Want to show that lim 𝑦𝑛 = +∞ by using definition:
𝑛→∞ 𝑛→∞
Let 𝛼 > 0 be given
Since lim 𝑥𝑛 = +∞ ⇒ ∃𝐾 ∈ ℕ such that 𝑥𝑛 > 𝛼, ∀𝑛 ≥ 𝐾…①
𝑛→∞
But 𝑦𝑛 ≥ 𝑥𝑛 , ∀𝑛 ∈ ℕ ⇒ ⏟
𝑦𝑛 > 𝛼 , ∀𝑛 ≥ 𝐾
by ①
∴ lim 𝑦𝑛 = +∞
𝑛→∞
(2) Let lim 𝑦𝑛 = −∞ ⇒ lim (−𝑦𝑛 ) = +∞……….②
𝑛→∞ 𝑛→∞
𝑥𝑛 ≤ 𝑦𝑛 , ∀𝑛 ∈ ℕ ⇒ −𝑦𝑛 ≤ −𝑥𝑛 , ∀𝑛 ∈ ℕ……..③
Part (1) and ②+③: ⇒ lim (−𝑥𝑛 ) = +∞ ⇒ lim 𝑥𝑛 = −∞
𝑛→∞ 𝑛→∞
 65

Limit Comparison Test 3.6.11:

Let (𝑥𝑛 ) and (𝑦𝑛 ) be two sequences of positive real numbers.
𝑥
(1) Let lim 𝑦𝑛 = 𝐿 with 0 < 𝐿 < ∞. Then lim 𝑥𝑛 = +∞ iff lim 𝑦𝑛 = +∞
𝑛→∞ 𝑛 𝑛→∞ 𝑛→∞
𝑥𝑛
(2) Let lim = 0.
𝑛→∞ 𝑦𝑛
(a) If lim 𝑥𝑛 = +∞, then lim 𝑦𝑛 = +∞.
𝑛→∞ 𝑛→∞
(b) If (𝑦𝑛 ) is bounded, then lim 𝑥𝑛 = 0.
𝑛→∞
𝑥𝑛
(3) Let lim = +∞.
𝑛→∞ 𝑦𝑛
(a) If lim 𝑥𝑛 = 0, then lim 𝑦𝑛 = 0.
𝑛→∞ 𝑛→∞
(b) If (𝑥𝑛 ) is bounded, then lim 𝑦𝑛 = 0.
Proof:
𝑥𝑛
(1) Since lim = 𝐿 with 0 < 𝐿 < ∞:
𝑛→∞ 𝑦𝑛
𝐿 𝑥 𝐿
Take 𝜀 = 2 ⇒ ∃𝐾 ∈ ℕ such that |𝑦𝑛 − 𝐿| < 2 , ∀𝑛 ≥ 𝐾
𝑛
𝐿 𝑥𝑛 𝐿 𝐿 𝑥 3𝐿
⇒ − 2 < 𝑦 − 𝐿 < 2 , ∀𝑛 ≥ 𝐾 ⇒ 2 < 𝑦𝑛 < , ∀𝑛 ≥ 𝐾
𝑛 𝑛 2
𝐿 3𝐿
⇒ 2⏟𝑦𝑛 < 𝑥𝑛 <
⏟ 𝑦𝑛 , ∀𝑛 ≥ 𝐾
2
① ②
Suppose that lim 𝑥𝑛 = +∞.
𝑛→∞
3𝐿
Since 𝑥
⏟𝑛 < 𝑦 , ∀𝑛 ≥ 𝐾 ⇒ lim 𝑦𝑛 = +∞ (by the comparison Test 3.6.10)
2 𝑛
②
Suppose that lim 𝑦𝑛 = +∞.
𝑛→∞
𝐿
Since ⏟𝑦 < 𝑥𝑛 , ∀𝑛 ≥ 𝐾 ⇒ lim 𝑥𝑛 = +∞ (by the comparison Test 3.6.10)
2 𝑛 𝑛→∞
①
𝑥𝑛 𝑥
(2) Since lim = 0: Take 𝜀 = 1 ⇒ ∃𝐾 ∈ ℕ such that |𝑦𝑛 − 0| < 1, ∀𝑛 ≥ 𝐾
𝑛→∞ 𝑦𝑛 𝑛
𝑥𝑛 𝑥
⇒ −1 < 𝑦 < 1, ∀𝑛 ≥ 𝐾 ⇒ 𝑦𝑛 < 1, ∀𝑛 ≥ 𝐾 ⇒ 𝑥𝑛 < 𝑦𝑛 , ∀𝑛 ≥ 𝐾….①
𝑛 𝑛
(a) Suppose that lim 𝑥𝑛 = +∞.
①& The Comparison Test 6 ⇒ lim 𝑦𝑛 = +∞
𝑛→∞
(b) Suppose that (𝑦𝑛 ) is bounded ⇒ |𝑦𝑛 | ≤ 𝑀, ∀𝑛 ∈ ℕ…..②
Since 𝑦𝑛 > 0, ∀𝑛 ∈ ℕ we have ⏟ 0 < 𝑦𝑛 ≤ 𝑀, ∀𝑛 ∈ ℕ ………③
by ②
To show that lim 𝑥𝑛 = 0 we use definition: Let 𝜀 > 0 be given
𝑥 𝜀
In the definition of lim 𝑦𝑛 = 0: take 𝜀 to be 𝑀 ⇒ ∃𝐾 ∈ ℕ such that
𝑛→∞ 𝑛
𝑥𝑛 𝜀 𝜀 𝑥 𝜀 𝑥 𝜀
|𝑦 − 0| < 𝑀 , ∀𝑛 ≥ 𝐾 ⇒ − 𝑀 < 𝑦𝑛 < 𝑀 , ∀𝑛 ≥ 𝐾 ⇒ 𝑦𝑛 < 𝑀 , ∀𝑛 ≥ 𝐾
𝑛 𝑛 𝑛
𝜀 𝜀
⇒ 𝑥𝑛 < 𝑀 𝑦𝑛 ≤ 𝑀 𝑀 = 𝜀, ∀𝑛 ≥ 𝐾 ⇒ 𝑥𝑛 < 𝜀, ∀𝑛 ≥ 𝐾
|𝑥𝑛 − 0| < 𝜀, ∀𝑛 ≥ 𝐾 ⇒ lim 𝑥𝑛 = 0
⇒⏟
𝑛→∞
since 𝑥𝑛 >0,∀𝑛∈ℕ
 66

(3)
𝑥𝑛
(a) Suppose that lim = +∞ and lim 𝑥𝑛 = 0
𝑛→∞ 𝑦𝑛
1 1
⇒ lim = 0 and lim = +∞ (by Theorem 3.6.7)
𝑛→∞ 𝑥𝑛 /𝑦𝑛 𝑛→∞ 𝑥𝑛
1/𝑥𝑛 1
⇒ lim = 0 and lim = +∞
𝑛→∞ 1/𝑦𝑛 𝑛→∞ 𝑥𝑛
1
Part (2) (a) ⇒ lim = +∞ ⇒ lim 𝑦𝑛 = 0 (by Theorem 3.6.7)
𝑛→∞ 𝑦𝑛 𝑛→∞

𝑥𝑛
(b) Suppose that lim = +∞ and (𝑥𝑛 ) is bounded
𝑛→∞ 𝑦𝑛
𝑦𝑛
Theorem 4 ⇒ lim = 0 and (𝑥𝑛 ) is bounded
𝑛→∞ 𝑥𝑛
Part (2)(b) ⇒ lim 𝑦𝑛 = 0
𝑛→∞

Example 3.6.12: Let (𝑥𝑛 ) be a sequence of positive real numbers.

𝑥
(1) If lim 𝑛𝑛 = 𝐿 with 0 < 𝐿 < ∞, then find lim 𝑥𝑛 .
𝑛→∞ 𝑛→∞
𝑥𝑛
(2) If lim = 0, then find lim 𝑥𝑛 .
𝑛→∞ sin(𝑛) 𝑛→∞
|cos(𝑛)|
(3) If lim = +∞, then find lim 𝑥𝑛 .
𝑛→∞ 𝑥𝑛 𝑛→∞
1
|sin( )|
𝑛
(4) If lim = +∞, then find lim 𝑥𝑛 .
𝑛→∞ 𝑥𝑛 𝑛→∞

Solution:
𝑥𝑛
(1) Since lim = 𝐿 with 0 < 𝐿 < ∞, then we apply The Limit Comparison Test Part (1). Let
𝑛→∞ 𝑛
𝑦𝑛 = 𝑛 we have lim 𝑦𝑛 = lim 𝑛 = +∞ ⇒ lim 𝑥𝑛 = +∞
𝑛→∞ 𝑛→∞ 𝑛→∞
𝑥𝑛 𝑛 𝑥 |𝑥𝑛 | 𝑥𝑛
(2) lim = 0 ⇒ lim |sin(𝑛)| = 0 ⇒ lim = 0 ⇒ lim =0
𝑛→∞ sin(𝑛) 𝑛→∞ 𝑛→∞ |sin(𝑛)| 𝑛→∞ |sin(𝑛)|
Now, we apply The Limit Comparison Test Part (2):
Let 𝑦𝑛 = |sin(𝑛)| ⇒ (𝑦𝑛 ) is bounded (since 0 ≤ |sin(𝑛)| ≤ 1)
⇒ lim 𝑥𝑛 = 0
𝑛→∞
|cos(𝑛)| 𝑥𝑛
(3) Since lim = +∞ ⇒ lim |cos(𝑛)|
= 0, then we apply The Limit Comparison Test Part
𝑛→∞ 𝑥𝑛 𝑛→∞
(2): Let 𝑦𝑛 = |cos(𝑛)|
⇒ (𝑦𝑛 ) is bounded (since 0 ≤ |cos(𝑛)| ≤ 1)
⇒ lim 𝑥𝑛 = 0.
𝑛→∞
1
|sin( )|
𝑛
(4) Since lim = +∞, then we apply The Limit Comparison Test Part (3): Let 𝑦𝑛 =
𝑛→∞ 𝑥𝑛
1 1
|sin(𝑛)| we have lim 𝑦𝑛 = lim |sin(𝑛)| = 0 ⇒ lim 𝑥𝑛 = 0
𝑛→∞ 𝑛→∞ 𝑛→∞
1
|sin( )| 𝑥𝑛
𝑛
Another Solution for Part (4): lim = +∞ ⇒ lim 1 =0
𝑛→∞ 𝑥𝑛 𝑛→∞ |sin(𝑛)|
Now, we apply The Limit Comparison Test Part (2):
1 1
Let 𝑦𝑛 = |sin(𝑛)| ⇒ (𝑦𝑛 ) is bounded (since 0 ≤ |sin(𝑛)| ≤ 1)
⇒ lim 𝑥𝑛 = 0
𝑛→∞
 67

Remark 3.6.13: Observe that when we say that (𝑥𝑛 ) diverges means:
(𝑥𝑛 ) is properly diverges or diverges but not properly
For Example:
 ((−1)𝑛 ) diverges but it is not properly diverges.
 (sin(𝑛)) diverges but it is not properly diverges.
 (𝑛) diverges and it is properly diverges.

Exercises 3.6.14 (pages 93+94): 1,2,4,5,8

 68

Section 3.7: Introduction to Infinite Series

In elementary texts, an infinite series is sometimes ‘‘defined’’ to be ‘‘an expression of the form:
𝑥1 + 𝑥2 + 𝑥3 + ⋯ + 𝑥𝑛 + ⋯

‫نظرا لعدم القدرة على اضافة أي قيمة في نهاية مجموعة الرموز هذه والمكتوبة على‬ ً ‫هذا "التعريف" يفتقر إلى الوضوح وذلك‬
.‫شكل صف حيث أن هذا التعريف يتطلب عددًا ال نهائي من عمليات الجمع التي يتعين القيام بها‬

Definition 3.7.1: If 𝑋 = (𝑥𝑛 ) is a sequence in ℝ, then the infinite series (or simply the series)
generated by 𝑋 is the sequence 𝑆 = (𝑠𝑛 ) defined by:
𝑠1 = 𝑥1
𝑠2 = 𝑠1 + 𝑥2 = 𝑥1 + 𝑥2
𝑠𝑘 = 𝑠𝑘−1 + 𝑥𝑛 = 𝑥1 + 𝑥2 + ⋯ + 𝑥𝑘
⋮
The numbers 𝑥1 , 𝑥2 , … are called the terms of the series and the numbers 𝑠1 , 𝑠2 , … are called the
partial sums of this series.

Remark 3.7.2:
(1) If lim 𝑠𝑛 exists, we say that this series is convergent and call this limit the sum or the value
𝑛→∞
of this series
(2) If this limit does not exist, we say that the series 𝑆 is divergent.
(3) It is convenient to use symbols such as:
∑(𝑥𝑛 ) or ∑ 𝑥𝑛 or ∑∞ 𝑛=1 𝑥𝑛 … … … … . . ①
to denote both the infinite series 𝑆 generated by the sequence 𝑋 = (𝑥𝑛 ) and also to denote
the value lim 𝑠𝑛 , in case this limit exists, that is ∑ 𝑥𝑛 = lim 𝑠𝑛 .
𝑛→∞ 𝑛→∞
(4) Just as a sequence may be indexed such that its first element is not 𝑥1 , but is 𝑥0 , or 𝑥5 or
𝑥99 , we will denote the series having these numbers as their first element by the symbols:
a) If 𝑥0 is the first term of the sequence 𝑋 = (𝑥𝑛 )∞ 𝑛=0 = (𝑥0 , 𝑥1 , 𝑥2 , … ):
The series 𝑆 = (𝑠𝑛 ) = (𝑠0 , 𝑠1 , 𝑠2 , … ) is denoted as ∑∞
𝑛=0 𝑥𝑛 , where
𝑠0 = 𝑥0 , 𝑠1 = 𝑥0 + 𝑥1 , … , 𝑠𝑘 = 𝑠𝑘−1 + 𝑥𝑘 , 𝑘 ≥ 1, …
b) If 𝑥5 is the first term of the sequence 𝑋 = (𝑥𝑛 )∞ 𝑛=5 = (𝑥5 , 𝑥6 , 𝑥7 , … )
The series 𝑆 = 𝑛 = (𝑠5 , 𝑠6 , 𝑠7 , … ) is denoted as ∑∞
(𝑠 ) 𝑛=5 𝑥𝑛 , where:
𝑠5 = 𝑥5 , 𝑠6 = 𝑥5 + 𝑥6 , … , 𝑠𝑘 = 𝑠𝑘−1 + 𝑥𝑘 , 𝑘 ≥ 6, …

c) If 𝑥99 is the first term of the sequence 𝑋 = (𝑥𝑛 )∞

𝑛=99 = (𝑥99 , 𝑥100 , 𝑥101 , … )
The series 𝑆 = 𝑛 = (𝑠99 , 𝑠100 , 𝑠101 , … ) is denoted as ∑∞
(𝑠 ) 𝑛=99 𝑥𝑛 ,
where 𝑠99 = 𝑥99 , 𝑠100 = 𝑥99 + 𝑥100 , … , 𝑠𝑘 = 𝑠𝑘−1 + 𝑥𝑘 , 𝑘 ≥ 100, …

Definition 3.7.3: Consider the sequence 𝑋 = (𝑟 𝑛 )∞

𝑛=0 , where 𝑟 ∈ ℝ, which generates the series
∑∞ 𝑛 2 𝑛
𝑛=0 𝑟 = 1 + 𝑟 + 𝑟 + ⋯ + 𝑟 + ⋯ which is called the geometric series.
 69

Example 3.7.4:
1
(1) Show that if |𝑟| < 1, then the geometric series converges to 1−𝑟 , that is:
1
∑∞ 𝑛
𝑛=0 𝑟 = 1−𝑟
(2) Show that if |𝑟| ≥ 1, then the geometric series ∑∞ 𝑛
𝑛=0 𝑟 diverges.

1
Solution: Want to show that the sequence (𝑠𝑛 ) = (𝑠0 , 𝑠1 , 𝑠2 , … ) of partial sums converges to 1−𝑟
1
that is lim 𝑠𝑛 = 1−𝑟.
𝑛→∞
(𝑥𝑛 ) = (1, 𝑟, 𝑟 2 , … , 𝑟 𝑛 , … )
2 𝑛−1
𝑠𝑛 = 1 + 𝑟 + 𝑟 + ⋯ + 𝑟 + 𝑟𝑛
𝑟𝑠𝑛 = 𝑟 + 𝑟 2 + ⋯ + 𝑟 𝑛 + 𝑟 𝑛+1
𝑠𝑛 − 𝑟𝑠𝑛 = (1 + 𝑟 + 𝑟 2 + ⋯ + 𝑟 𝑛−1 + 𝑟 𝑛 ) − (𝑟 + 𝑟 2 + ⋯ + 𝑟 𝑛 + 𝑟 𝑛+1 )
= 1 − 𝑟 𝑛+1
1−𝑟 𝑛+1
⇒ 𝑠𝑛 (1 − 𝑟) = 1 − 𝑟 𝑛+1 ⇒ 𝑠𝑛 = (Closed Form of 𝑠𝑛 )
1−𝑟
1−𝑟 𝑛+1 1− lim 𝑟 𝑛+1
⇒ lim 𝑠𝑛 = lim = 𝑛→∞
……….①
𝑛→∞ 𝑛→∞ 1−𝑟 1−𝑟
1−𝑟 𝑛+1 1−0 1
(1) If |𝑟| < 1, then lim 𝑟 𝑛+1 = 0. ①: ⇒ lim 𝑠𝑛 = lim = 1−𝑟 = 1−𝑟
𝑛→∞ 𝑛→∞ 𝑛→∞ 1−𝑟
1 1
⇒ the series converges to 1−𝑟 that is ∑∞ 𝑛
𝑛=0 𝑟 = lim 𝑠𝑛 = 1−𝑟
𝑛→∞

(2) If |𝑟| ≥ 1, then lim 𝑟 𝑛+1 does not exist ⇒ lim 𝑠𝑛 does not exists
𝑛→∞ 𝑛→∞
⇒ (𝑠𝑛 ) diverges ⇒ ∑∞ 𝑛
𝑛=0 𝑟 diverges.

Remark 3.7.5: Let (𝑠𝑛 ) be the sequence of partial sums of the geometric series
∑∞ 𝑛
𝑛=𝑛0 𝑟 = 𝑟
𝑛0
+ 𝑟 1+𝑛0 + 𝑟 2+𝑛0 + ⋯ + 𝑟 𝑛+𝑛0 + ⋯
𝑟 𝑛0 −𝑟 𝑛+1 First Term−Term after last
The closed form of 𝑠𝑛 is: 𝑠𝑛 = 1−𝑟 (= )
1−Base
(1) If |𝑟| < 1, then
∑∞ 𝑛 𝑛0 2 𝑛
𝑛=𝑛0 𝑟 = 𝑟 (1 + 𝑟 + 𝑟 + ⋯ + 𝑟 + ⋯ )
1
= 𝑟 𝑛0 (∑∞ 𝑛
𝑛=0 𝑟 ) = 𝑟
𝑛0
( )
1−𝑟
𝑟 𝑛0 First Term
⇒ ∑∞
𝑛=𝑛0 𝑟 = 1−𝑟 (=
𝑛
)
1−Base
1 1 35 −1
1 1 1 ( 3 )−( 8 ) 35 −1 242
3 3 38
a) + 34 + ⋯ + 37 = 1 = 2 = 2(37 ) = 4374
33 1−
3 3
1 1 1
1 1 1 ( 𝑛)−( 𝑚 ) ( 𝑛) 1
2 2 2
b) Let 𝑚 > 𝑛. Then + 2𝑛+1 + ⋯ + 2𝑚−1 = 1 < 1 = 2𝑛−1
2𝑛 1− 1−
2 2
6 3
6 First Term ( 3) 3
24
c) ∑∞
𝑛=5 2𝑛−1 = = 1 = 2
1 =4
1−Base 1− ( )
2 2
 70

1
Example 3.7.6: Determine whether the series ∑∞
𝑛=1 𝑛(𝑛+1) converges or diverges. If it converges
find its sum.
1
Solution: Let (𝑠𝑛 ) be the sequence of partial sums of ∑∞
𝑛=1 𝑛(𝑛+1). We must find a closed form of 𝑠𝑛 :
1 1 1 1 1
𝑠𝑛 = 1(2) + 2(3) + 3(4) + ⋯ + (𝑛−1)𝑛 + 𝑛(𝑛+1)
1 1 1 1 1 1 1 1 1 1
= (1 − 2) + (2 − 3) + (3 − 4) + ⋯ + (𝑛−1 − 𝑛) + (𝑛 − 𝑛+1) = 1 − 𝑛+1 (Closed Form)
1 1 1
⇒ lim 𝑠𝑛 = lim (1 − 𝑛+1) = 1 ⇒ ∑∞ ∞
𝑛=1 𝑛(𝑛+1) = lim 𝑠𝑛 = 1 ⇒ ∑𝑛=1 𝑛(𝑛+1) converges to 1
𝑛→∞ 𝑛→∞ 𝑛→∞

Cauchy Criterion for Series 3.7.7:

The series ∑ 𝑥𝑛 converges iff ∀𝜀 > 0, ∃𝐻 ∈ ℕ with ⏟
|𝑥𝑛+1 + 𝑥𝑛+2 + ⋯ + 𝑥𝑚 | < 𝜀, ∀𝑚 > 𝑛 ≥ 𝐻.
𝑚−𝑛 terms

Proof: Let (𝑠𝑛 ) be the sequence of partial sums of the series ∑ 𝑥𝑛 .

∑ 𝑥𝑛 converges iff (𝑠𝑛 ) converges iff (𝑠𝑛 ) is a Cauchy sequence
iff ∀𝜀 > 0, ∃𝐻 ∈ ℕ such that |𝑠𝑚 − 𝑠𝑛 | < 𝜀, ∀𝑚 > 𝑛 ≥ 𝐻
iff ∀𝜀 > 0, ∃𝐻 ∈ ℕ such that
|𝑥1 + 𝑥2 + ⋯ + 𝑥𝑛 + 𝑥𝑛+1 + ⋯ + 𝑥𝑚 − (𝑥1 + 𝑥2 + ⋯ + 𝑥𝑛 )| < 𝜀, ∀𝑚 > 𝑛 ≥ 𝐻
iff ∀𝜀 > 0, ∃𝐻 ∈ ℕ such that |𝑥𝑛+1 + ⋯ + 𝑥𝑚 | < 𝜀, ∀𝑚 > 𝑛 ≥ 𝐻

The Cauchy Criterion for Series (Contraposition Form) 3.7.8:

∑ 𝑥𝑛 diverges iff ∃𝜀0 > 0, ∀𝐻 ∈ ℕ, and ∃𝑚0 > 𝑛0 ≥ 𝐻 with |𝑥𝑛0 +1 + 𝑥𝑛+2 + ⋯ + 𝑥𝑚0 | ≥ 𝜀0 .

(−1)𝑛
Example 3.7.9: Using the Cauchy Criterion, show that the series ∑∞
𝑛=1 converges.
𝑛!
1 1
(Hint: 𝑛! ≤ 2𝑛−1 , ∀𝑛 ∈ ℕ)
Solution: We will apply the Cauchy Criterion for series:
Let 𝜀 > 0 be given and let 𝑚 > 𝑛:
(−1)𝑛+1 (−1)𝑚 (−1)𝑛+1 (−1)𝑚 1 1
| (𝑛+1)! + ⋯ + | ≤ | (𝑛+1)! | + ⋯ + | | = (𝑛+1)! + ⋯ + 𝑚!
⏟ 𝑚! ⏟ 𝑚! ⏟
𝑚−𝑛 terms by the triangle inequality 𝑚−𝑛 terms
1 1 1
≤ 2𝑛 + 2𝑛+1 + ⋯ + 2𝑚−1 (by the Hint)
1 1 1
( 𝑛)−( 𝑚 ) ( 𝑛) 1
2 2 2
= 1 < 1 = 2𝑛−1
1− 1−
2 2
(−1)𝑛+1 (−1)𝑚 1
⇒ | (𝑛+1)! + ⋯ + | < 2𝑛−1 ………①
𝑚!
1 1
Since lim = 0: Applying the definition of lim = 0, ∃𝐾 ∈ ℕ such that
𝑛→∞ 2𝑛−1 𝑛→∞ 2𝑛−1
1 1
|2𝑛−1 − 0| < 𝜀, ∀𝑛 ≥ 𝐾 ⇒ 2𝑛−1 < 𝜀, ∀𝑛 ≥ 𝐾……………②
1
Take 𝐻 = 𝐾 and let 𝑚 > 𝑛 ≥ 𝐻 ⇒ 𝑛 ≥ 𝐾 ⇒ ②: 2𝑛−1 < 𝜀, ∀𝑛 ≥ 𝐻………..③
(−1)𝑛+1 (−1)𝑚 1
① ⇒ | (𝑛+1)! + ⋯ + | < 2𝑛−1 < 𝜀, ∀𝑚 > 𝑛 ≥ 𝐻
𝑚!
(−1)𝑛
The Cauchy Criterion for series ⇒ ∑∞
𝑛=1 converges
𝑛!
1 1 1 1
Definition 3.7.10: The series ∑∞
𝑛=1 𝑛 = 1 + 2 + 3 + ⋯ + 𝑛 + ⋯ is called the Harmonic series.
 71

1
Example 3.7.11: Show that the Harmonic series ∑∞ 𝑛=1 𝑛 diverges.
Proof: We will show that the Contrapositive of the Cauchy criterion for series holds:
1
Take 𝜀0 = 2, and let 𝐻 ∈ ℕ.
Take 𝑚0 , 𝑛0 ∈ ℕ such that 𝑛0 = 𝐻 and 𝑚0 = 2𝑛0 ⇒ 𝑚0 > 𝑛0 ≥ 𝐻
𝑛0 −terms

1 1 ⏞1 1 1 1
|𝑥𝑛0 +1 + 𝑥𝑛0 +2 + ⋯ + 𝑥𝑚0 | = |
⏟ + ⋯+ |= + ⋯+ ≥ 𝑛0 ( )=
𝑛0 + 1 2𝑛0 𝑛0 + 1 2𝑛0 2𝑛0 2
𝑚0 −𝑛0 =𝑛0
1
⇒ ∃𝑚0 > 𝑛0 ≥ 𝐻 such that |𝑥𝑛0 +1 + 𝑥𝑛0 +1 + ⋯ + 𝑥𝑚0 | ≥ 𝜀0 = 2
1 1
The Contrapositive of the Cauchy criterion for the series ∑∞ ∞
𝑛=1 𝑛 holds ⇒ ∑𝑛=1 𝑛 diverges.

The 𝒏𝒕𝒉 Term Test 3.7.12: If ∑ 𝑥𝑛 converges, then lim 𝑥𝑛 = 0.

𝑛→∞

Proof: Suppose that ∑ 𝑥𝑛 converges to 𝑥.

Let (𝑠𝑛 ) be the sequence of partial sums of ∑ 𝑥𝑛 ⇒ lim 𝑠𝑛 = 𝑥
𝑛→∞

⇒ 𝑥𝑛 = 𝑠𝑛 − 𝑠𝑛−1
⇒ lim 𝑥𝑛 = lim (𝑠𝑛 − 𝑠𝑛−1 ) = lim 𝑠𝑛 − lim 𝑠𝑛−1 = 𝑥 − 𝑥 = 0 lim 𝑥𝑛 = 0
𝑛→∞ 𝑛→∞ 𝑛→∞ 𝑛→∞ 𝑛→∞

The 𝒏𝒕𝒉 Term Test (Contraposition Form) 3.7.13: If lim 𝑥𝑛 ≠ 0, then ∑ 𝑥𝑛 diverges

Example 3.7.14:
𝑛
(1) Show that ∑∞𝑛=1 2𝑛+1 diverges
(2) Consider the series generated by ((−1)𝑛 ) that is ∑∞ 𝑛
𝑛=1(−1) . Show that the series
diverges.

Solution: We apply the 𝑛𝑡ℎ term test (Contrapositive Form):

𝑛 1
(1) lim 2𝑛+1 = 2 ≠ 0
𝑛→∞
𝑛
The Contrapositive of the 𝑛𝑡ℎ Term Test of the series ∑∞
𝑛=1 2𝑛+1 holds
𝑛
⇒ ∑∞
𝑛=1 2𝑛+1 diverges
(2) Observe that:
lim (−1)2𝑛 = lim 1 = 1
𝑛→∞ 𝑛→∞
2𝑛+1 } ⇒ lim (−1)2𝑛 ≠ lim (−1)2𝑛+1
lim (−1) = lim (−1) = −1 𝑛→∞ 𝑛→∞
𝑛→∞ 𝑛→∞
⇒ lim (−1)𝑛 does not exist ⇒ lim (−1)𝑛 ≠ 0.
𝑛→∞ 𝑛→∞
The Contrapositive of the 𝑛𝑡ℎ Term Test of the series ∑∞ 𝑛
𝑛=1(−1) holds
∞ 𝑛
∴ ∑𝑛=1(−1) diverges
 72

Comparison Test for series 3.7.15: Let (𝑥𝑛 ) and (𝑦𝑛 ) be sequences of real numbers such that:
0 ≤ 𝑥𝑛 ≤ 𝑦𝑛 , ∀𝑛 ≥ 𝐾 for some 𝐾 ∈ ℕ.
(1) If ∑ 𝑦𝑛 converges, then ∑ 𝑥𝑛 converges.
(2) If ∑ 𝑥𝑛 diverges, then ∑ 𝑦𝑛 diverges.
Proof:
(1) Suppose that ∑ 𝑦𝑛 converges.
To show that ∑ 𝑥𝑛 converges we apply the Cauchy Criterion for series:
Let 𝜀 > 0 be given.
∑ 𝑦𝑛 converges ⇒ By the Cauchy Criterion applied to ∑ 𝑦𝑛 : ∃𝐻 ∈ ℕ such that
|𝑦𝑛+1 + 𝑦𝑛+2 + ⋯ + 𝑦𝑚 | < 𝜀, ∀𝑚 > 𝑛 ≥ 𝐻……①
⇒ 𝑦𝑛+1 + 𝑦𝑛+2 + ⋯ + 𝑦𝑚 < 𝜀, ∀𝑚 > 𝑛 ≥ 𝐻 (since 𝑦𝑛 ≥ 0, ∀𝑛 ∈ ℕ)
⇒ |𝑥𝑛+1 + 𝑥𝑛+2 + ⋯ + 𝑥𝑚 | = 𝑥𝑛+1 + 𝑥𝑛+2 + ⋯ + 𝑥𝑚
≤ 𝑦𝑛+1 + 𝑦𝑛+2 + ⋯ + 𝑦𝑚
< 𝜀, ∀𝑚 > 𝑛 ≥ 𝐻
⇒ |𝑥𝑛+1 + 𝑥𝑛+2 + ⋯ + 𝑥𝑚 | < 𝜀, ∀𝑚 > 𝑛 ≥ 𝐻
Cauchy Criterion for series ⇒ ∑ 𝑥𝑛 converges
(2) The contrapositive of part (1): If ∑ 𝑥𝑛 diverges, then ∑ 𝑦𝑛 diverges

Theorem 3.7.16: Let (𝑥𝑛 ) be a sequence of positive real numbers and let (𝑠𝑛 ) be the sequence of
partial sums of ∑ 𝑥𝑛 . Then the following are equivalent:
(1) ∑ 𝑥𝑛 converges
(2) The sequence (𝑠𝑛 ) is bounded.
(3) There is a bounded subsequence of (𝑠𝑛 ).
In this case: ∑ 𝑥𝑛 = lim 𝑠𝑛 = sup{𝑠𝑛 : 𝑛 ∈ ℕ}
𝑛→∞

Proof: Since (𝑥𝑛 ) be a sequence of positive real numbers, then (𝑠𝑛 ) is an increasing sequence:
𝑠𝑛 = 𝑠𝑛−1 + 𝑥𝑛 ≥ 𝑠𝑛−1 ⇒ 𝑠𝑛 ≥ 𝑠𝑛−1 , ∀𝑛 ∈ ℕ

(1) (2) Suppose that ∑ 𝑥𝑛 converges ⇒ (𝑠𝑛 ) converges ⇒ (𝑠𝑛 ) bounded

(2) (1) Suppose that (𝑠𝑛 ) is bounded ⇒ (𝑠𝑛 ) bounded and increasing
Monotone Convergence Theorem MCT: ⇒ (𝑠𝑛 ) converges ⇒ ∑ 𝑥𝑛 converges

(2) (3) is an exercise (optional)

1 1 1 1
Definition 3.7.17: The series ∑∞
𝑛=1 𝑛𝑝 = 1 + 2𝑝 + 3𝑝 + ⋯ + 𝑛𝑝 + ⋯ is called the 𝒑-series.
 73

Example 3.7.18:
1
(1) Show that if 𝑝 > 1, then ∑∞
𝑛=1 𝑛𝑝 converges
1
(2) Show that if 𝑝 ≤ 1, then ∑∞
𝑛=1 𝑛𝑝 diverges
Solution:
1
(1) Let 𝑝 > 1 and let (𝑠𝑛 ) be the sequence of partial sums of ∑∞
𝑛=1 𝑛𝑝 .
1
Since the series ∑∞
𝑛=1 𝑛𝑝 has positive terms, then we will show that Part (3) of Theorem 3.7.16
holds, that is we will find a bounded subsequence of (𝑠𝑛 ):
Let 𝑘𝑗 = 2𝑗 − 1, 𝑗 ∈ ℕ. Then:
𝑠𝑘1 = 𝑠1 = 1
1 1 1 1 1
𝑠𝑘2 = 𝑠3 = 1 + 2𝑝 + 3𝑝 ≤ 1 + 2𝑝 + 2𝑝 = 1 + 2𝑝−1
1 1 1 1 1 1 1 4 1 1
𝑠𝑘3 = 𝑠7 = 1
⏟+ 2𝑝 + 3𝑝 + ⏟ + 5𝑝 + 6𝑝 + 7𝑝 ≤ 1 + 2𝑝−1 + 4𝑝 = 1 + 2𝑝−1 + (2𝑝−1 )2
4𝑝
𝑠3 =𝑠𝑘2
1 1
⇒ 𝑠𝑘3 ≤ 1 + 2𝑝−1 + (2𝑝−1 )2
⋮
1 1 1 1 1
𝑠𝑘𝑛 ≤ 1 + 2𝑝−1 + (2𝑝−1 )2 + ⋯ + (2𝑝−1 )𝑛−1 ≤ ∑∞
𝑛=0 = 1
⏟ ⏟ (2𝑝−1 )𝑛 1− 𝑝−1
2
Finite geometric series 1
𝑟= 𝑝−1 ⇒0<𝑟<1 since 𝑝>1
2

1
⇒ 0 ≤ 𝑠𝑘𝑛 ≤ 1 , ∀𝑛 ∈ ℕ ⇒ (𝑠𝑘𝑛 ) is bounded
1− 𝑝−1
2
⇒ There is a bounded subsequence of (𝑠𝑛 ) ⇒ Part (3) of Theorem 3.7.16 holds
1
⇒ Part (1) of Theorem 3.7.16 holds ⇒ ∑∞𝑛=1 𝑝 converges 𝑛
𝑝 1 1
(2) Let 𝑝 ≤ 1. Then 𝑛 ≤ 𝑛, ∀𝑛 ∈ ℕ ⇒ ≤ 𝑛𝑝 , ∀𝑛 ∈ ℕ
𝑛
1 1
⇒ 0 ≤ 𝑛 ≤ 𝑛𝑝 , ∀𝑛 ∈ ℕ……..①
1
But the harmonic series ∑ diverges
𝑛
1
By ① and the Comparison Test: ⇒ ∑ 𝑛𝑝 diverges

Remark 3.7.19: Recall Theorem 3.6.7: Let (𝑥𝑛 ) be a sequence of positive numbers. Then
1
lim 𝑥𝑛 = ∞ iff lim 𝑥 = 0
𝑛→∞ 𝑛→∞ 𝑛
 74

Limit Comparison Test for series 3.7.20: Let (𝑥𝑛 ) and (𝑦𝑛 ) be a sequence of positive real
𝑥
numbers such that lim 𝑦 𝑛 = 𝐿.
𝑛→∞ 𝑛
(1) If 0 < 𝐿 < ∞, then ∑ 𝑥𝑛 converges iff ∑ 𝑦𝑛 converges.
(2) If 𝐿 = 0 and ∑ 𝑦𝑛 converges, then ∑ 𝑥𝑛 converges.
(3) If 𝐿 = ∞ and ∑ 𝑥𝑛 converges, then ∑ 𝑦𝑛 converges
Proof:
𝑥𝑛 𝐿
(1) Suppose that 0 < 𝐿 < ∞. By the definition of the limit lim = 𝐿: Take 𝜀 = 2
𝑛→∞ 𝑦𝑛
𝑥𝑛 𝐿 𝐿 𝑥𝑛 𝐿
⇒ ∃𝐾 ∈ ℕ such that | 𝑦 − 𝐿| < 2 , ∀𝑛 ≥ 𝐾 ⇒ − 2 < − 𝐿 < 2 , ∀𝑛 ≥ 𝐾
𝑛 𝑦𝑛
𝐿 𝑥𝑛 3𝐿 𝐿 3𝐿
≤ ≤ , ∀𝑛 ≥ 𝐾 ⇒ (2) 𝑦𝑛 ≤ 𝑥𝑛 ≤ ( 2 ) 𝑦𝑛 , ∀𝑛 ≥ 𝐾………①
2 𝑦𝑛 2
Suppose that ∑ 𝑥𝑛 converges:
2
①: ⇒ 0 ≤ 𝑦𝑛 ≤ 𝐿 𝑥𝑛 , ∀𝑛 ≥ 𝐾……….②
②& ∑ 𝑥𝑛 converges & The Comparison Test for series 3.7.15: ⇒ ∑ 𝑦𝑛 converges
Suppose that ∑ 𝑦𝑛 converges:
3𝐿
①: ⇒ 0 ≤ 𝑥𝑛 ≤ ( 2 ) 𝑦𝑛 , ∀𝑛 ≥ 𝐾………③
③& ∑ 𝑦𝑛 converges & The Comparison Test for series 3.7.15: ⇒ ∑ 𝑥𝑛 converges
(2) Suppose that 𝐿 = 0 and ∑ 𝑦𝑛 converges.
𝑥
In the definition of the limit lim 𝑦 𝑛 = 0: Take 𝜀 = 1 ⇒ ∃𝐾 ∈ ℕ such that
𝑛→∞ 𝑛
𝑥𝑛 𝑥𝑛
| 𝑦 − 0| < 1, ∀𝑛 ≥ 𝐾 ⇒ < 1, ∀𝑛 ≥ 𝐾 ⇒ 𝑥𝑛 < 𝑦𝑛 , ∀𝑛 ≥ 𝐾
𝑛 𝑦𝑛
⇒ 0 ≤ 𝑥𝑛 < 𝑦𝑛 , ∀𝑛 ≥ 𝐾……④
④& ∑ 𝑦𝑛 converges & The Comparison Test for series 3.7.15: ⇒ ∑ 𝑥𝑛 converges
𝑥𝑛
(3) Suppose that 𝐿 = ∞ and ∑ 𝑥𝑛 converges. Since lim = ∞ we have from Theorem 3.6.7
𝑛→∞ 𝑦𝑛
1 𝑦𝑛
that lim 𝑥 = 0 ⇒ lim = 0.
𝑛→∞ ( 𝑦 𝑛) 𝑛→∞ 𝑥𝑛
𝑛
Since ∑ 𝑥𝑛 converges, it follows from Part (2) that ∑ 𝑦𝑛 converges.

The Contrapositive of the Limit Comparison Test for series 3.7.21:

𝑥𝑛
Let (𝑥𝑛 ) and (𝑦𝑛 ) be a sequence of positive real numbers such that lim = 𝐿.
𝑛→∞ 𝑦𝑛
(1) If 0 < 𝐿 < ∞, then ∑ 𝑥𝑛 diverges iff ∑ 𝑦𝑛 diverges.
(2) If 𝐿 = 0 and ∑ 𝑥𝑛 diverges, then ∑ 𝑦𝑛 diverges.
(3) If 𝐿 = ∞ and ∑ 𝑦𝑛 diverges , then ∑ 𝑥𝑛 converges.

Exercises 3.7.22 (pages 100+101): 3,4,6,9-13

 75

Chapter 4: Limits

Section 4.1: Limits of Functions

Definition 4.1.1: Let 𝐴 ⊆ ℝ. A point 𝑐 ∈ ℝ is a cluster point of 𝑨 if for every 𝛿 > 0 there exists
at least one point 𝑎 ∈ 𝐴, 𝑎 ≠ 𝑐 such that |𝑎 − 𝑐| < 𝛿, that is:
𝑐 is a cluster point of 𝐴 iff ∃𝑎 ∈ 𝐴, 𝑎 ≠ 𝑐 such that 𝑎 ∈ 𝑉𝛿 (𝑐) = (𝑐 − 𝛿, 𝑐 + 𝛿),∀𝛿 > 0
Remark 4.1.2:
1
(1) Taking 𝛿 = 𝑛 , 𝑛 ∈ ℕ we have:
(2) 𝑐 is a cluster point of 𝐴 iff there is a sequence (𝑎𝑛 ) in 𝐴 with lim 𝑎𝑛 = 𝑐 and 𝑎𝑛 ≠ 𝑐, ∀𝑛 ∈ ℕ
𝑛→∞
(3) To find the set of all cluster points of a set 𝐴: Take all sequences (with different terms) in 𝐴 that
are convergent. The set of all limit points of these sequences is the set of all cluster points of 𝐴.
Example 4.1.3:
(1) Let 𝐴 = {𝑎1 , … , 𝑎𝑛 } be a finite set ⇒ 𝐴 does not contain any sequence that converges
⇒ 𝐴 has no cluster points ⇒ The set of all cluster points of 𝐴 is the empty set ∅ = { }.

(2) The infinite set ℕ does not contain any sequence that converges ⇒ ℕ has no cluster points
⇒ The set of all cluster points of ℕ is the empty set ∅ = { }.

1 1
(3) Let 𝐴 = {𝑛 : 𝑛 ∈ ℕ} ⇒ 𝐴 contains sequence (𝑛) that converges to 0 ⇒ 𝐴 has only one cluster
point, which is 0 ⇒ The set of all cluster points of 𝐴 is {0}.

(4) The set of all cluster points of any one of the intervals [𝑎, 𝑏], (𝑎, 𝑏], [𝑎, 𝑏), (𝑎, 𝑏) is [𝑎, 𝑏].

𝜺-𝜹 Definition 4.1.4: Let 𝐴 ⊆ ℝ and 𝑐 ∈ ℝ be a cluster point of 𝐴, and let 𝑓: 𝐴 → ℝ be a function.
A real number 𝐿 is said to be a limit of 𝑓 at 𝑐, denoted by lim 𝑓(𝑥) = 𝐿 if ∀𝜀 > 0, ∃𝛿 > 0 such
𝑥→𝑐
that if 𝑥 ∈ 𝐴 and 0 < |𝑥 − 𝑐| < 𝛿, then |𝑓(𝑥) − 𝐿| < 𝜀.
That is lim 𝑓(𝑥) = 𝐿 iff ∀𝜀 > 0, ∃𝛿 > 0 such that if 𝑥 ∈ 𝑉𝛿 (𝑐) − {𝑐} ⇒ 𝑓(𝑥) ∈ 𝑉𝜀 (𝐿)
𝑥→𝑐
 76

Example 4.1.5: Using the graph of the given function, if 𝜀 > 0 is given as in the figure then
approximate the larges possible value of 𝛿 that satisfies
the definition of the limit lim 𝑓 = 1
𝑥→0.5

Solution: lim 𝑓 = 1.
𝑥→0.5
𝜀
From the figure the largest possible value of 𝛿 is 𝛿 ≅ 3

Notations 4.1.6:
(1) lim 𝑓(𝑥) = 𝐿 may be written as:
𝑥→𝑐
 lim 𝑓 = 𝐿
𝑥→𝑐
 𝑓 converges to 𝐿 at 𝑐
 𝑓(𝑥) approaches 𝐿 as 𝑥 approaches 𝑐.’’
(2) The definition: lim 𝑓(𝑥) = 𝐿 ∈ ℝ is equivalent to the following 3 conditions:
𝑥→𝑐
① ∀𝜀 > 0
② ∃𝛿 > 0
③ 𝑥 ∈ 𝐴 and 0 < |𝑥 − 𝑐| < 𝛿 ⇒ |𝑓(𝑥) − 𝐿| < 𝜀

𝑥 𝑐
Example 4.1.7: Using definition, show that if 𝑐 ∈ ℝ, then lim 2 = 2.
𝑥→𝑐
𝑥 𝑐
Solution: Let 𝜀 > 0 be given. 𝑓(𝑥) = 2 , 𝐿 = 2
𝑥 𝑐 |𝑥−𝑐|
|𝑓(𝑥) − 𝐿| = | − | = …….①
2 2 2
|𝑥−𝑐|
< 𝜀 ⇔ |𝑥 − 𝑐| < 2𝜀
2
Take 0 < 𝛿 ≤ 2𝜀.
|𝑥−𝑐|
Let 0 < |𝑥 − 𝑐| < 𝛿 ⇒ 0 < |𝑥 − 𝑐| < 2𝜀 ⇒ < 𝜀 ⇒ |𝑓(𝑥) − 𝐿| < 𝜀 (by ①)
2
0 < |𝑥 − 𝑐| < 𝛿 ⇒ |𝑓(𝑥) − 𝐿| < 𝜀
𝑥 𝑐
∴ lim 2 = 2.
𝑥→𝑐

Note: Recall that if 𝑎, 𝑏 ∈ [0, ∞), then: |√𝑎 − √𝑏| ≤ √|𝑎 − 𝑏|

Example 4.1.8: Using definition, show that if 𝑐 ∈ [0, ∞), then lim √𝑥 = √𝑐.
𝑥→𝑐
Solution: Let 𝜀 > 0 be given. 𝑓(𝑥) = √𝑥, 𝐿 = √𝑐
|𝑓(𝑥) − 𝐿| = |√𝑥 − √𝑐| ≤ √|𝑥 − 𝑐|…….①
√|𝑥 − 𝑐| < 𝜀 ⇔ |𝑥 − 𝑐| < 𝜀 2
Take 0 < 𝛿 ≤ 𝜀 2 .
Let 0 < |𝑥 − 𝑐| < 𝛿 ⇒ |𝑥 − 𝑐| < 𝜀 2 ⇒ √|𝑥 − 𝑐| < 𝜀 ⇒ |𝑓(𝑥) − 𝐿| < 𝜀 (by ①)
So, 0 < |𝑥 − 𝑐| < 𝛿 ⇒ |𝑓(𝑥) − 𝐿| < 𝜀
∴ lim √𝑥 = √𝑐.
𝑥→𝑐
 77

3𝑥 2 −𝑥+2
Example 4.1.9: Using definition, show that lim = 4.
𝑥→1 𝑥 2 −𝑥+1
3𝑥 2 −𝑥+2 3𝑥 2 −𝑥+2
Solution: lim = 2 ⇒ 𝑐 = 1, 𝑓(𝑥) = and 𝐿 = 4 ⇒ |𝑥 − 𝑐| = |𝑥 − 1|
𝑥→1 𝑥 2 −𝑥+1 𝑥 2 −𝑥+1
Let 𝜀 > 0 be given.
3𝑥 2 −𝑥+2 3𝑥 2 −𝑥+2 4(𝑥 2 −𝑥+1) −𝑥 2 +3𝑥−2 𝑥 2 −3𝑥+2 (𝑥−1)(𝑥−2)
|𝑓(𝑥) − 𝐿| = | − 4| = | 𝑥 2 −𝑥+1 − |=| | = | 𝑥 2 −𝑥+1 | = | |
𝑥 2 −𝑥+1 𝑥 2 −𝑥+1 𝑥 2 −𝑥+1 𝑥 2 −𝑥+1
𝑥−2 𝑥−2
= |𝑥 − 𝑐| |⏟ | ⇒ |𝑓(𝑥) − 𝐿| = |𝑥 − 𝑐| |⏟ | …….①
𝑥 2 −𝑥+1 𝑥 2 −𝑥+1
‫يجب التخلص منه‬ ‫يجب التخلص منه‬
𝑥−2
Let 𝑔(𝑥) = 𝑥 2 −𝑥+1 be continuous on [𝑐 − 𝛿, 𝑐 + 𝛿] = [1 − 𝛿, 1 + 𝛿]:
‫مقترحة‬
⏞= 1 and let 𝑔(𝑥) = 𝑥 𝑥
Take 𝛿 , 𝑥 ∈ [𝑐 − 𝛿, 𝑐 + 𝛿] ⇒ 𝑔(𝑥) = 𝑥 2 −𝑥+1 , 𝑥 ∈ [−2,0]
𝑥 2 −𝑥+1
Now we find the absolute maximum value of |𝑔(𝑥)| on the interval [−2,0]
1−𝑥 2
 𝑔′ (𝑥) = (𝑥 2 −𝑥+1)2 = 0 ⇒ 1 − 𝑥 2 = 0 ⇒ 𝑥 = ±1
⇒ |𝑔(𝑥)| ≤ max(|𝑔(−2)|, |𝑔(−1)|, |𝑔(0)|)
2 1 1
= max (7 , 3 , 0) = 3
1
⇒ |𝑔(𝑥)| ≤ 3 , ∀𝑥 ∈ [−2,0]………..②
1
① |𝑓(𝑥) − 𝐿| = |𝑥 − 𝑐| |𝑔(𝑥)|
⏟ ≤ 3 |𝑥 − 𝑐| (by ②)
‫يجب التخلص منه‬
1
⇒ |𝑓(𝑥) − 𝐿| ≤ 3 |𝑥 − 𝑐|…………③
1
Now, 3 |𝑥 − 𝑐| < 𝜀 ⇔ |𝑥 − 𝑐| < 3𝜀
‫مقترحة‬ ‫محسوبة‬

Take 0 < 𝛿 ≤ min ( ⏞

1 ,⏞
3𝜀 )

1
Let 0 < |𝑥 − 𝑐| < 𝛿 ⇒ 0 < |𝑥 − 𝑐| < min(1,3𝜀) ⇒ 0 < |𝑥 − 𝑐| < 3𝜀 ⇒ 3 |𝑥 − 𝑐| < 𝜀
⇒ |𝑓(𝑥) − 𝐿| < 𝜀 (by ③). So, 0 < |𝑥 − 𝑐| < 𝛿 ⇒ |𝑓(𝑥) − 𝐿| < 𝜀
3𝑥 2 −𝑥+2
∴ lim =4
𝑥→1 𝑥 2 −𝑥+1

𝑥 3 +3 5
Example 4.1.10: Using definition, show that lim = − 4.
𝑥→−2 𝑥 2 −𝑥−2
𝑥 3 +3 5
Solution: Let 𝑐 = −2, 𝑓(𝑥) = 𝑥 2 −𝑥−2, and 𝐿 = − 4 ⇒ |𝑥 − 𝑐| = |𝑥 + 2|
Let 𝜀 > 0 be given
𝑥3 + 3 5 𝑥3 + 3 5
|𝑓(𝑥) − 𝐿| = | 2
− (− )| = | 2
+ |
𝑥 −𝑥−2 4 𝑥 −𝑥−2 4

4𝑥 3 + 5𝑥 2 − 5𝑥 + 2
4(𝑥 3 +3)+5(𝑥 2 −𝑥−2) 4𝑥 3 +5𝑥 2 −5𝑥+2 4 5 −5 2
=| |=| |
4(𝑥 2 −𝑥−2) 4(𝑥 2 −𝑥−2) −8 6 −2 −2
(𝑥+2)(4𝑥 2 −3𝑥+1)
=| | 4 −3 1 0
4(𝑥 2 −𝑥−2)
2
4𝑥 − 3𝑥 + 1
|𝑥+2| 4𝑥 2 −3𝑥+1 |𝑥−𝑐| 4𝑥 2 −3𝑥+1
⇒ |𝑓(𝑥) − 𝐿| = | |= | |
4 𝑥 2 −𝑥−2 4 𝑥 2 −𝑥−2
 78

|𝑥−𝑐| 4𝑥 2 −3𝑥+1
⇒ |𝑓(𝑥) − 𝐿| = |⏟𝑥 2 −𝑥−2 | ……①
4
‫نريد التخلص منه‬
‫بالحصول على القيمة‬
‫العظمى المطلقة له‬
4𝑥 2 −3𝑥+1
Let 𝑔(𝑥) = be continuous on [𝑐 − 𝛿, 𝑐 + 𝛿] = [−2 − 𝛿, −2 + 𝛿]
𝑥 2 −𝑥−2

‫مقترحة‬
⏞ 1
Take 𝛿 = 2……..② (why we cannot take 𝛿 = 1?)
1 1
⇒ 𝑔(𝑥) is continuous on [𝑐 − 𝛿, 𝑐 + 𝛿] = [−2 − , −2 + ] = [−2.5, −1.5]
2 2
Want to find the absolute maximum of |𝑔(𝑥)| on the interval [−2.5, −1.5]:
−𝑥 2 −18𝑥+7
 𝑔′ (𝑥) = (𝑥 2 −𝑥−2)2
(How?)
18±√352
𝑔′ (𝑥) = 0 ⇒ −𝑥 2 + 18𝑥 + 7 = 0 ⇒ 𝑥 2 − 18𝑥 − 7 = 0 ⇒ 𝑥 = 2
18−√352 18+√352
⇒ = 9 − √88 ≅ −0.4 and = 9 + √88 ≅ 18.4
2 2

⇒ 𝑔 is decreasing on [−2.5, −1.5]

134 58 58
⇒ |𝑔(𝑥)| ≤ max( |𝑔(−2.5)|, |𝑔(−1.5)|) = max ( 27 , 7 ) = 7 …….③

|𝑥−𝑐| |𝑥−𝑐| 58 29
①&③: ⇒ |𝑓(𝑥) − 𝐿| = |𝑔(𝑥)| ≤ ( 7 ) = 14 |𝑥 − 𝑐|, ∀𝑥 ∈ [−2.5, −1.5]
4 4
29
⇒ |𝑓(𝑥) − 𝐿| ≤ 14 |𝑥 − 𝑐|, ∀𝑥 ∈ [−2.5, −1.5] ……④
29 14𝜀
Now, take 14 |𝑥 − 𝑐| < 𝜀 ⇔ |𝑥 − 𝑐| < 29
‫مقترحة‬ ‫محسوبة‬
⏞
1 ⏞
14𝜀
Take 0 < 𝛿 ≤ min ( , ). So, 0 < |𝑥 − 𝑐| < 𝛿 ⇒ |𝑓(𝑥) − 𝐿| < 𝜀
2 29

𝑥 3 +3 5
∴ lim = − 4.
𝑥→−2 𝑥 2 −𝑥−2

Execise 4.1.11: Using definition, show that

(1) lim 𝑥 2 = 1 (2) lim 𝑥 2 = 4
𝑥→−1 𝑥→2
𝑥 3 +3 5 1
(3) lim = −3 (4) lim = −1
𝑥→−2 2𝑥 2 +𝑥−3 𝑥→−1 𝑥
1 1
(5) lim 𝑥 = 2 (6) lim (𝑥 3 − 1) = 0
𝑥→2 𝑥→1
3
(7) lim (𝑥 + 1) = 0 (8) lim(𝑥 3 + 1) = 2
𝑥→−1 𝑥→1
 79

Sequential Criterion for Limits 4.1.12:

Let 𝑓: 𝐴 → ℝ be a function and let 𝑐 ∈ ℝ be a cluster point of 𝐴. Then the following are
equivalent.
(a) lim 𝑓 = 𝐿
𝑥→𝑐
(b) For every sequence (𝑥𝑛 ) in 𝐴 that converges to 𝑐 such that 𝑥𝑛 ≠ 𝑐, ∀𝑛 ∈ ℕ, the sequence
(𝑓(𝑥𝑛 )) converges to L, that is:
∀(𝑥𝑛 ) ⊆ 𝐴 with 𝑥𝑛 ≠ 𝑐, ∀𝑛 ∈ ℕ and 𝑥𝑛 → 𝑐 ⇒ lim 𝑓(𝑥𝑛 ) = 𝐿.
𝑛→∞
Proof:
(a) (b) Suppose that part (a) holds that is lim 𝑓 = 𝐿 and let (𝑥𝑛 ) ⊆ 𝐴 with 𝑥𝑛 ≠ 𝑐, ∀𝑛 ∈ ℕ and
𝑥→𝑐
lim 𝑥𝑛 = 𝑐. Want to show that lim 𝑓(𝑥𝑛 ) = 𝐿 using definition:
𝑛→∞ 𝑛→∞

Let 𝜀 > 0 be given. By the definition of lim 𝑓 = 𝐿: ∃𝛿 > 0 such that ∀𝑥 ∈ 𝐴 we have
𝑥→𝑐

0 < |𝑥 − 𝑐| < 𝛿 ⇒ |𝑓(𝑥) − 𝐿| < 𝜀………①

Applying the definition of lim 𝑥𝑛 = 𝑐 for 𝛿 > 0: ⇒ ∃𝐾 ∈ ℕ such that
𝑛→∞
|𝑥𝑛 − 𝑐| < 𝛿, ∀𝑛 ≥ 𝐾…………②
Since 𝑥𝑛 ≠ 𝑐, ∀𝑛 ≥ 𝐾 and by ② we have 0 < |𝑥𝑛 − 𝑐| < 𝛿, ∀𝑛 ≥ 𝐾…………③
①&③: ⇒ |𝑓(𝑥𝑛 ) − 𝐿| < 𝜀, ∀𝑛 ≥ 𝐾.
∴ lim 𝑓(𝑥𝑛 ) = 𝐿.
𝑛→∞
(b) (a) We will prove its contraposition, that is we will prove that:

not (a) ⇒ not (b):

1
Suppose that (a) is false ⇒ lim 𝑓 ≠ 𝐿: Take 𝜀 = 1 and 𝛿 = 𝑛 , 𝑛 ∈ ℕ ⇒ ∃𝑥𝑛 ∈ 𝐴 such
𝑥→𝑐
1
that 0 < |𝑥𝑛 − 𝑐| < 𝛿 = 𝑛 and |𝑓(𝑥𝑛 ) − 𝐿| ≥ 1…………….④
1
0 < |𝑥𝑛 − 𝑐| < 𝛿 ⇒ 0 < |𝑥𝑛 − 𝑐| < 𝑛 ……..⑤
Letting 𝑛 → ∞ in ⑤ and using the Comparison test for sequences we have: 𝑥𝑛 → 𝑐
So, ∃(𝑥𝑛 ) ⊆ 𝐴 with 𝑥𝑛 ≠ 𝑐, ∀𝑛 ∈ ℕ and 𝑥𝑛 → 𝑐 and |𝑓(𝑥𝑛 ) − 𝐿| ≥ 1, , ∀𝑛 ∈ ℕ
⇒ lim 𝑓(𝑥𝑛 ) ≠ 𝐿 ⇒ (b) is false. So we prove that:
𝑛→∞
not (a) ⇒ not(b).
∴ (b) ⇒ (a).

The contraposition of the Sequential Criterion for Limits implies the following divergence criteria:
Divergence Criteria 4.1.13: Let 𝑓: 𝐴 → ℝ be a function and let 𝑐, 𝐿 ∈ ℝ such that 𝑐 is a cluster
point of 𝐴. Then
(1) lim 𝑓 ≠ 𝐿 iff ∃(𝑥𝑛 ) ⊆ 𝐴 with 𝑥𝑛 ≠ 𝑐, ∀𝑛 ∈ ℕ and 𝑥𝑛 → 𝑐 such that lim 𝑓(𝑥𝑛 ) ≠ 𝐿.
𝑥→𝑐 𝑛→∞

(2) lim 𝑓 does not exist iff ∃(𝑥𝑛 ) ⊆ 𝐴 with 𝑥𝑛 ≠ 𝑐, ∀𝑛 ∈ ℕ and 𝑥𝑛 → 𝑐 such that (𝑓(𝑥𝑛 )) diverges.
𝑥→𝑐
(3) lim 𝑓 does not exist iff ∃(𝑥𝑛 ) ⊆ 𝐴, ∃(𝑦𝑛 ) ⊆ 𝐴 with 𝑥𝑛 ≠ 𝑐, 𝑦𝑛 ≠ 𝑐, ∀𝑛 ∈ ℕ and 𝑥𝑛 → 𝑐,
𝑥→𝑐
𝑦𝑛 → 𝑐 such that lim 𝑓(𝑥𝑛 ) ≠ lim 𝑓(𝑦𝑛 ).
𝑛→∞ 𝑛→∞
 80

1
Example 4.1.14: Using the divergence criterion, show that lim 𝑥 does not exist
𝑥→0
1 1
Proof: Let 𝑓(𝑥) = 𝑥 , 𝑥 ∈ ℝ − {0} and let 𝑐 = 0 and 𝑥𝑛 = 𝑛 , 𝑛 ∈ ℕ
1 1 1
⇒ (𝑛) ⊆ ℝ − {0}, 𝑛 ≠ 0, ∀𝑛 ∈ ℕ and 𝑛 → 0
1 1
Now, (𝑓(𝑥𝑛 )) = (𝑥 ) = ( 1 ) = (𝑛) is properly diverges (diverges to +∞)
𝑛
𝑛
1
Part (b) of the Divergence Criteria ⇒ lim 𝑥 does not exist
𝑥→0

1
Example 4.1.15: Using the divergence criterion, show that lim sin (𝑥) does not exist
𝑥→0
1 1 1
Proof: Let 𝑥𝑛 = 𝑛 , 𝑛 ∈ ℕ ⇒ 𝑛 ≠ 0 and 𝑛 → 0
1
Now, (sin (𝑥 )) = (sin(𝑛)) diverges (See Example 3.4.7)
𝑛
1
Divergence Criteria Part (b): ⇒ lim sin (𝑥) does not exist
𝑥→0

Example 4.1.16: Show that lim sgn(𝑥) does not exist, where sgn(𝑥) is the signum function
𝑥→0
1 ,𝑥 > 0
defined by sgn(𝑥) = { 0 , 𝑥 = 0}
−1 , 𝑥 < 0
1 1
Solution: We will apply Part (c) of the Divergence Criteria: Let 𝑥𝑛 = 𝑛 and 𝑦𝑛 = − 𝑛, 𝑛 ∈ ℕ
⇒ 𝑥𝑛 ≠ 0, 𝑦𝑛 ≠ 0, ∀𝑛 ∈ ℕ and 𝑥𝑛 → 0, 𝑦𝑛 → 0.
1
(sgn(𝑥𝑛 )) = (sgn (𝑛)) = (1) = (1,1,1, … ) ⇒ lim sgn(𝑥𝑛 ) = 1
1
(sgn(𝑦𝑛 )) = (sgn (− 𝑛)) = (−1) = (−1, −1, −1, … ) ⇒ lim sgn(𝑦𝑛 ) = −1
⇒ lim sgn(𝑥𝑛 ) ≠ lim sgn(𝑦𝑛 )
Part (c) of the Divergence Criteria ⇒ lim sgn(𝑥) does not exist
𝑥→0

1 1
Exercise 4.1.17: Let 𝑓: ℝ → ℝ be a bounded function, let 𝐼𝑛 = (− , ) , 𝑛 ∈ ℕ, and let
𝑛 𝑛
𝑎𝑛 = sup 𝑓(𝐼𝑛 ) and 𝑏𝑛 = inf 𝑓(𝐼𝑛 ) , 𝑛 ∈ ℕ.
(1) Show that (𝑎𝑛 ) is a decreasing sequence
(2) Show that (𝑏𝑛 ) is an increasing sequence
(3) Show that if lim 𝑎𝑛 = lim 𝑏𝑛 , then lim 𝑓 exists
𝑛→∞ 𝑛→∞ 𝑥→0

Exercises of Section 4.1 (page 110): 1,2,3,6,7,9-12,14

 81

Section 4.2: Limit Theorems

Definition 4.2.1: Let 𝐴 ⊆ ℝ, 𝑓: 𝐴 → ℝ is a function and let 𝑐 ∈ ℝ be a cluster point of 𝐴. The function
𝑓 is called bounded on a neighborhood of 𝑐 if there exists a 𝛿-neighborhood 𝑉𝛿 (𝑐) = (𝑐 − 𝛿, 𝑐 + 𝛿)
and a constant 𝑀 > 0 such that:
|𝑓(𝑥)| ≤ 𝑀, ∀𝑥 ∈ 𝐴 ∩ 𝑉𝛿 (𝑐).

Lemma 4.2.2: Let 𝐴 ⊆ ℝ, 𝑓: 𝐴 → ℝ is a function and let 𝑐, 𝐿 ∈ ℝ. If lim 𝑓 exists, then 𝑓 is bounded
𝑥→𝑐

on some neighborhood of 𝑐, that is ∃𝛿 > 0 such that 𝑓 is bounded on 𝐴 ∩ 𝑉𝛿 (𝑐).

Remark 4.2.3 (The Contraposition of Lemma 4.2.2):

Let 𝐴 ⊆ ℝ, 𝑓: 𝐴 → ℝ is a function and let 𝑐 ∈ ℝ be a cluster point of 𝐴. If 𝑓(𝑥) is unbounded on 𝐴 ∩
𝑉𝛿 (𝑐), ∀𝛿 > 0, then lim 𝑓 does not exist.
𝑥→𝑐
1
Example 4.2.4: lim+ 𝑥 does not exist
𝑥→0
1 1
Solution: 𝑓(𝑥) = 𝑥 is unbounded on (0, 𝛿), ∀𝛿 > 0 ⇒ lim+ 𝑥 does not exist
𝑥→0

Definition 4.2.5: Let 𝐴 ⊆ ℝ, 𝑓: 𝐴 → ℝ, 𝑔: 𝐴 → ℝ be functions and let 𝑏 ∈ ℝ.

(1) Define the multiple function 𝑏𝑓: 𝐴 → ℝ by (𝑏𝑓)(𝑥) = 𝑏𝑓(𝑥), 𝑥 ∈ 𝐴.
(2) Define the sum function 𝑓 + 𝑔: 𝐴 → ℝ by (𝑓 + 𝑔)(𝑥) = 𝑓(𝑥) + 𝑔(𝑥), 𝑥 ∈ 𝐴.
(3) Define the difference function 𝑓 − 𝑔: 𝐴 → ℝ by
(𝑓 − 𝑔)(𝑥) = 𝑓(𝑥) − 𝑔(𝑥), 𝑥 ∈ 𝐴.
(4) Define the product function 𝑓𝑔: 𝐴 → ℝ by (𝑓𝑔)(𝑥) = 𝑓(𝑥)𝑔(𝑥), 𝑥 ∈ 𝐴.
𝑓
(5) If 𝑔(𝑥) ≠ 0, ∀𝑥 ∈ 𝐴, define the quotient function 𝑔 : 𝐴 → ℝ by
𝑓 𝑓(𝑥)
(𝑔) (𝑥) = 𝑔(𝑥) , 𝑥 ∈ 𝐴.
 82

Theorem 4.2.6: Let 𝐴 ⊆ ℝ, 𝑓: 𝐴 → ℝ, 𝑔: 𝐴 → ℝ be functions and let 𝑏, 𝑐 ∈ ℝ such that 𝑐 is a cluster

point of 𝐴. Let lim 𝑓 = and lim 𝑔 both exist. Then
𝑥→𝑐 𝑥→𝑐

(1) Multiple Rule: lim 𝑏𝑓 = 𝑏 lim 𝑓

𝑥→𝑐 𝑥→𝑐

(2) Sum Rule: lim(𝑓 + 𝑔) = lim 𝑓 + lim 𝑔

𝑥→𝑐 𝑥→𝑐 𝑥→𝑐

(3) Difference Rule: lim(𝑓 − 𝑔) = lim 𝑓 − lim 𝑔

𝑥→𝑐 𝑥→𝑐 𝑥→𝑐

(4) Product Rule: lim 𝑓𝑔 = (lim 𝑓) (lim 𝑔)

𝑥→𝑐 𝑥→𝑐 𝑥→𝑐

𝑓 lim 𝑓 1 1
(5) Quotient Rule: If lim 𝑔 ≠ 0, then lim 𝑔 = 𝑥→𝑐 . In particular, lim 𝑔 =
𝑥→𝑐 lim 𝑔𝑥→𝑐 𝑥→𝑐 lim 𝑔
𝑥→𝑐 𝑥→𝑐

Proof: Let lim 𝑓 = 𝐿1 and lim 𝑔 = 𝐿2 .

𝑥→𝑐 𝑥→𝑐

We will use the Sequential Criterion for Limits 4.1.12. So, let (𝑥𝑛 ) be a sequence in 𝐴 satisfying:
𝑥𝑛 ≠ 𝑐, ∀𝑛 ∈ ℕ and 𝑥𝑛 → 𝑐.
Since lim 𝑓 = 𝐿1 and lim 𝑔 = 𝐿2 , then the Sequential Criterion for Limits 4.1.12 implies that:
𝑥→𝑐 𝑥→𝑐

lim 𝑓(𝑥𝑛 ) = 𝐿1 ………..①

𝑛→∞

and
lim 𝑔(𝑥𝑛 ) = 𝐿2 ……….②
𝑛→∞

(1) lim (𝑏𝑓)(𝑥𝑛 ) = lim 𝑏𝑓(𝑥𝑛 ) = 𝑏 lim 𝑓(𝑥𝑛 ) = 𝑏𝐿1 ⇒ lim (𝑏𝑓)(𝑥𝑛 ) = 𝑏𝐿1
𝑛→∞ 𝑛→∞ 𝑛→∞ 𝑛→∞

∴ The Sequential Criterion for Limits 4.1.12 ⇒ lim 𝑏𝑓 = 𝑏𝐿1 = 𝑏 lim 𝑓

𝑥→𝑐 𝑥→𝑐

(2) lim (𝑓 + 𝑔)(𝑥𝑛 ) = lim (𝑓(𝑥𝑛 ) + 𝑔(𝑥𝑛 )) = lim 𝑓(𝑥𝑛 ) + lim 𝑔(𝑥𝑛 ) = 𝐿1 + 𝐿2
𝑛→∞ 𝑛→∞ 𝑛→∞ 𝑛→∞

⇒ lim (𝑓 + 𝑔)(𝑥𝑛 ) = 𝐿1 + 𝐿2
𝑛→∞

∴ The Sequential Criterion for Limits 4.1.12⇒ lim(𝑓 + 𝑔) = 𝐿1 + 𝐿2 = lim 𝑓 + lim 𝑔

𝑥→𝑐 𝑥→𝑐 𝑥→𝑐

(3) Similar to Part (2)

(4) lim (𝑓𝑔)(𝑥𝑛 ) = lim (𝑓(𝑥𝑛 )𝑔(𝑥𝑛 )) = ( lim 𝑓(𝑥𝑛 )) ( lim 𝑔(𝑥𝑛 )) = 𝐿1 𝐿2
𝑛→∞ 𝑛→∞ 𝑛→∞ 𝑛→∞

⇒ lim (𝑓𝑔)(𝑥𝑛 ) = 𝐿1 𝐿2
𝑛→∞

∴ The Sequential Criterion for Limits 4.1.12⇒ lim(𝑓𝑔) = 𝐿1 𝐿2 = (lim 𝑓) (lim 𝑔)

𝑥→𝑐 𝑥→𝑐 𝑥→𝑐

𝑓 𝑓(𝑥 ) lim 𝑓(𝑥𝑛 ) 𝐿 𝑓 𝐿

(5) lim (𝑔) (𝑥𝑛 ) = lim (𝑔(𝑥𝑛 )) = 𝑛→∞
= 𝐿1 ⇒ lim (𝑔) (𝑥𝑛 ) = 𝐿1
𝑛→∞ 𝑛→∞ 𝑛 lim 𝑔(𝑥𝑛 ) 2 𝑛→∞ 2
𝑛→∞

𝑓 𝐿 lim 𝑓
𝑥→𝑐
∴ The Sequential Criterion for Limits 4.1.12⇒ lim 𝑔 = 𝑀 =
𝑥→𝑐 lim 𝑔
𝑥→𝑐
 83

Remark 4.2.7: Let 𝑐 ∈ ℝ be a cluster point of set 𝐴 ⊆ ℝ, 𝑓: 𝐴 → ℝ and let 𝑓𝑗 : 𝐴 → ℝ be functions

for 𝑗 = 1, … , 𝑛 such that lim 𝑓 exist and lim 𝑓𝑗 exist ∀𝑗 = 1, … , 𝑛 . Then
𝑥→𝑐 𝑥→𝑐

(1) lim(𝑓1 + ⋯ + 𝑓𝑛 ) = lim 𝑓1 + ⋯ + lim 𝑓𝑛

𝑥→𝑐 𝑥→𝑐 𝑥→𝑐

(2) lim 𝑓1 ∙ 𝑓2 ∙∙∙ 𝑓𝑛 = (lim 𝑓1 ) ∙∙∙ (lim 𝑓𝑛 )

𝑥→𝑐 𝑥→𝑐 𝑥→𝑐
𝑛
(3) lim 𝑓 𝑛 = (lim 𝑓)
𝑥→𝑐 𝑥→𝑐

Example 4.2.8: Using Theorems:

(1) show that lim 𝑥 𝑛 = 𝑐 𝑛 , ∀𝑛 ∈ ℕ
𝑥→𝑐

(2) If 𝑐 ∈ ℝ∗ , show that lim 𝑥 −𝑛 = 𝑐 −𝑛 , ∀𝑛 ∈ ℕ

𝑥→𝑐

Solution: Observe that lim𝑥 = 𝑐 exists

𝑥→𝑐
𝑛
(1) lim 𝑥 𝑛 = (lim𝑥) = 𝑐 𝑛 , ∀𝑛 ∈ ℕ
⏟
𝑥→𝑐 𝑥→𝑐
by the Product Rule

1 lim 1 1
(2) lim 𝑥 −𝑛 = lim 𝑥 𝑛 = 𝑥→𝑐
⏟= 𝑐 𝑛 = 𝑐 −𝑛
𝑥→𝑐 ⏟ 𝑥→𝑐 lim 𝑥 𝑛
𝑥→𝑐
by Part (1)
by the Quotient Rule

Example 4.2.9: Using Theorems, show that:

𝑝(𝑥) 𝑝(𝑐)
(1) If 𝑝(𝑥) and 𝑞(𝑥) are polynomials and 𝑞(𝑐) ≠ 0, then lim 𝑞(𝑥) = 𝑞(𝑐)
𝑥→𝑐
1 1
(2) If 𝑐 ∈ ℝ∗ , then lim 𝑥 = 𝑐
𝑥→𝑐
𝑥 2 −4 1
(3) lim 𝑥 3 −8 = 3
𝑥→2
1
(4) lim 𝑥 does not exist
𝑥→0

Solution:
(1) Let 𝑝(𝑥) = 𝑎𝑛 𝑥 𝑛 + 𝑎𝑛−1 𝑥 𝑛−1 + ⋯ + 𝑎1 𝑥 + 𝑎0 and
𝑞(𝑥) = 𝑏𝑚 𝑥 𝑚 + 𝑏𝑚−1 𝑥 𝑚−1 + ⋯ + 𝑏1 𝑥 + 𝑏0
lim𝑞(𝑥) = lim(𝑏𝑚 𝑥 𝑚 + 𝑏𝑚−1 𝑥 𝑚−1 + ⋯ + 𝑏1 𝑥 + 𝑏0 )
𝑥→𝑐 𝑥→𝑐

= lim 𝑏𝑚 𝑥 𝑚 + lim 𝑏𝑚−1 𝑥 𝑚−1 + ⋯ + lim 𝑏1 𝑥 + lim 𝑏0

⏟ 𝑥→𝑐 𝑥→𝑐 𝑥→𝑐 𝑥→𝑐
by the Sum Rule
 84

= lim 𝑥 𝑚 + 𝑏𝑚−1 lim 𝑥 𝑚−1 + ⋯ + 𝑏1 lim 𝑥 + 𝑏0

⏟ 𝑏𝑚 𝑥→𝑐 𝑥→𝑐 𝑥→𝑐
by the Multiple Rule

= 𝑏𝑚 𝑐 𝑚 + 𝑏𝑚−1 𝑐 𝑚−1 + ⋯ + 𝑏1 𝑐 + 𝑏0
⏟
by Example 7

= 𝑞(𝑐)
≠0
Similarly, lim𝑝(𝑥) = 𝑝(𝑐).
𝑥→𝑐

𝑝(𝑥) lim 𝑝(𝑥) 𝑝(𝑐)

Now, lim 𝑞(𝑥) = 𝑥→𝑐 = 𝑞(𝑐)
⏟
𝑥→𝑐 lim 𝑞(𝑥)
𝑥→𝑐
by the Quotient Rule

1 lim 1 1
(2) Observe that lim𝑥 = 𝑐 ≠ 0 ⇒ lim 𝑥 = 𝑥→𝑐 =𝑐
𝑥→𝑐 ⏟
𝑥→𝑐 lim 𝑥
𝑥→𝑐
by the Quotient Rule

𝑥 2 −4 0
(3) lim 𝑥 3 −8 = 0 ⇒ We cannot apply the Quotient Rule
𝑥→2
𝑥 2 −4 (𝑥−2)(𝑥+2) 𝑥+2 4 1
lim = lim (𝑥−2)(𝑥 2 +2𝑥+4) = ⏟
lim = 12 = 3
𝑥→2 𝑥 3 −8 𝑥→2 𝑥→2 𝑥 2 +2𝑥+4
by the Quotient Rule

1 1
(4) Since lim 𝑥 = 0 ⇒ We cannot apply the Quotient Rule
𝑥→0
1
Let 𝛿 > 0 and let 𝑓(𝑥) = 𝑥 , 𝑥 ∈ ℝ∗
1
⇒ is unbounded on ℝ∗ ∩ 𝑉𝛿 (0), ∀𝛿 > 0
𝑥 -3 -2 -1 1 2

1
⇒ is unbounded on (−𝛿, 0) ∪ (0, 𝛿), ∀𝛿 > 0
𝑥

By the Contraposition of Lemma 4.2.2 we have:

1 1
lim 𝑥 does not exist ⇒ lim 𝑥 does not exist
𝑥→0 𝑥→0
 85

Theorem 4.2.10: Let 𝑐 ∈ ℝ be a cluster point of set 𝐴 ⊆ ℝ, 𝑓: 𝐴 → ℝ is a function such that

lim 𝑓 exists. Then
𝑥→𝑐

(1) lim|𝑓| = |lim𝑓|.

𝑥→𝑐 𝑥→𝑐

(2) If lim 𝑓 = 0 iff lim|𝑓| = 0

𝑥→𝑐 𝑥→𝑐

(3) If 𝑓(𝑥) ≥ 0, ∀𝑥 ∈ 𝐴, then lim √𝑓 = √lim𝑓

𝑥→𝑐 𝑥→𝑐

Proof: Since lim 𝑓 exists ⇒ Let lim 𝑓 = 𝐿.

𝑥→𝑐 𝑥→𝑐

We will use the sequential Criterion for Limits 4.1.12 to prove Parts (1)&(2):
Let (𝑥𝑛 ) be a sequence in 𝐴 satisfying: 𝑥𝑛 ≠ 𝑐, ∀𝑛 ∈ ℕ and 𝑥𝑛 → 𝑐.
Since lim 𝑓 = 𝐿………..①
𝑥→𝑐

①& The Sequential Criterion for Limits 4.1.12 ⇒ lim 𝑓(𝑥𝑛 ) = 𝐿……..②
𝑛→∞

(1) lim|𝑓(𝑥𝑛 )| = |lim 𝑓(𝑥𝑛 )| = |𝐿| ⇒ lim |𝑓(𝑥𝑛 )| = |𝐿|……③

𝑛→∞

③& The Sequential Criterion for Limits 4.1.12 ⇒ lim |𝑓(𝑥)| = |𝐿|
𝑥→𝑐

∴ lim|𝑓| = |lim𝑓|
𝑥→𝑐 𝑥→𝑐

(2)
lim|𝑓| = |lim𝑓| = |0| = 0 ⇒ lim|𝑓| = 0
Suppose that lim 𝑓 = 0 ⇒ ⏟
𝑥→𝑐 𝑥→𝑐 𝑥→𝑐 𝑥→𝑐
by Part (1)
Suppose that lim|𝑓| = 0 ⇒ 0 |𝑓| = |lim𝑓| ⇒ |lim 𝑓| = 0 ⇒ lim𝑓 = 0
𝑥→𝑐 ⏟= lim 𝑥→𝑐 𝑥→𝑐 𝑥→𝑐 𝑥→𝑐
by Part (1)

(3) lim √𝑓(𝑥𝑛 ) = √ lim 𝑓(𝑥𝑛 ) = √𝐿 ⇒ lim √𝑓 (𝑥𝑛 ) = √𝐿 ……④

𝑛→∞ 𝑛→∞ 𝑛→∞

④& The Sequential Criterion for Limits 4.1.12 ⇒ lim √𝑓 (𝑥) = √𝐿

𝑥→𝑐

∴ lim √𝑓 = √lim𝑓
𝑥→𝑐 𝑥→𝑐
 86

Theorem 4.2.11: Let 𝑐 ∈ ℝ be a cluster point of 𝐴 ⊆ ℝ and 𝑓: 𝐴 → ℝ. If 𝑎 ≤ 𝑓(𝑥) ≤ 𝑏

(or 𝑎 < 𝑓(𝑥) < 𝑏), ∀𝑥 ∈ 𝐴, 𝑥 ≠ 𝑐 and if lim 𝑓 exists, then
𝑥→𝑐

𝑎 ≤ lim 𝑓 ≤ 𝑏.
𝑥→𝑐

Proof: Since lim 𝑓 exists ⇒ let lim 𝑓 = 𝐿. We will use the sequential Criterion for Limits 4.1.12:
𝑥→𝑐 𝑥→𝑐

Let (𝑥𝑛 ) be a sequence in 𝐴 satisfying: 𝑥𝑛 ≠ 𝑐, ∀𝑛 ∈ ℕ and 𝑥𝑛 → 𝑐.

Since lim 𝑓 = 𝐿……..①
𝑥→𝑐

①& the Sequential Criterion for Limits 4.1.12⇒ lim 𝑓(𝑥𝑛 ) = 𝐿………..②
𝑛→∞

Since 𝑥𝑛 ∈ 𝐴, ∀𝑛 ∈ ℕ and 𝑥𝑛 ≠ 𝑐, ∀𝑛 ∈ ℕ ⇒ 𝑎 ≤ 𝑓(𝑥𝑛 ) ≤ 𝑏, ∀𝑛 ∈ ℕ

⇒ 𝑎 ≤ lim 𝑓(𝑥𝑛 ) ≤ 𝑏 ⇒ 𝑎 ≤ 𝐿 ≤ 𝑏 (by ①) ⇒ 𝑎 ≤ lim 𝑓 ≤ 𝑏.
𝑛→∞ 𝑥→𝑐

Squeez Theorem for Functions 4.2.12:

Let 𝐴 ⊆ ℝ, 𝑓, 𝑔, ℎ are functions on 𝐴, and let 𝑐 ∈ ℝ be a cluster point of 𝐴.
If 𝑓(𝑥) ≤ 𝑔(𝑥) ≤ ℎ(𝑥), ∀𝑥 ∈ 𝐴, 𝑥 ≠ 𝑐; and if lim 𝑓 and lim ℎ both exist and
𝑥→𝑐 𝑥→𝑐

lim 𝑓 = lim ℎ, then lim 𝑔 = lim 𝑓 = lim ℎ.

𝑥→𝑐 𝑥→𝑐 𝑥→𝑐 𝑥→𝑐 𝑥→𝑐

Proof: Since lim 𝑓 and lim ℎ both exist and lim 𝑓 = lim ℎ ⇒ let lim 𝑓 = lim ℎ = 𝐿. We will use
𝑥→𝑐 𝑥→𝑐 𝑥→𝑐 𝑥→𝑐 𝑥→𝑐 𝑥→𝑐

the sequential Criterion:

Let (𝑥𝑛 ) be a sequence in 𝐴 satisfying: 𝑥𝑛 ≠ 𝑐, ∀𝑛 ∈ ℕ and 𝑥𝑛 → 𝑐.
Since lim 𝑓 = lim ℎ = 𝐿, then the Sequential Criterion for Limits 4.1.12 implies that:
𝑥→𝑐 𝑥→𝑐

lim 𝑓(𝑥𝑛 ) = lim ℎ(𝑥𝑛 ) = 𝐿………..①

𝑛→∞ 𝑥→𝑐

Since 𝑥𝑛 ∈ 𝐴, ∀𝑛 ∈ ℕ and 𝑥𝑛 ≠ 𝑐, ∀𝑛 ∈ ℕ ⇒ 𝑓(𝑥𝑛 ) ≤ 𝑔(𝑥𝑛 ) ≤ ℎ(𝑥𝑛 ), ∀𝑛 ∈ ℕ

⇒ lim 𝑓(𝑥𝑛 ) ≤ lim 𝑔(𝑥𝑛 ) ≤ lim ℎ(𝑥𝑛 ), ∀𝑛 ∈ ℕ
𝑛→∞ 𝑛→∞ 𝑛→∞

⇒ 𝐿 ≤ lim 𝑔(𝑥𝑛 ) ≤ 𝐿 (by ①) ⇒ lim 𝑔(𝑥𝑛 ) = 𝐿

𝑛→∞ 𝑛→∞

The Sequential Criterion for Limits 4.1.12 ⇒ lim 𝑔 = 𝐿 ⇒ lim 𝑔 = lim 𝑓 = lim ℎ
𝑥→𝑐 𝑥→𝑐 𝑥→𝑐 𝑥→𝑐
 87

Example 4.2.13:
(1) Show that lim sin 𝑥 = 0 (Hint: |sin 𝑥| ≤ |𝑥| for all 𝑥 ∈ ℝ)
𝑥→0
𝑥2
(2) Show that lim cos 𝑥 = 1 (Hint: 1 − ≤ cos 𝑥 ≤ 1 for all 𝑥 ∈ ℝ)
𝑥→0 2

cos 𝑥−1 𝑥2
(3) Show that lim = 0 (Hint: 1 − ≤ cos 𝑥 ≤ 1 for all 𝑥 ∈ ℝ)
𝑥→0 𝑥 2

sin 𝑥 𝑥2 sin 𝑥
(4) Show that lim = 1 (Hint: 1 − ≤ ≤ 1 for all 𝑥 ∈ ℝ∗ )
𝑥→0 𝑥 6 𝑥

Proof:
(1) |sin 𝑥| ≤ |𝑥| for 𝑥 ∈ ℝ ⇒ 0 ≤ |sin 𝑥| ≤ |𝑥| for 𝑥 ∈ ℝ
The squeeze theorem ⇒ lim 0 ≤ lim |sin 𝑥| ≤ lim|𝑥| ⇒ 0 ≤ lim|sin 𝑥| ≤ 0
𝑥→0 𝑥→0 𝑥→0 𝑥→0

⇒ lim |sin 𝑥| = 0 ⇒ lim sin 𝑥 = 0

𝑥→0 𝑥→0
𝑥2
(2) 1− ≤ cos 𝑥 ≤ 1 for 𝑥 ∈ ℝ
2
𝑥2
The squeeze theorem ⇒ lim (1 − ) ≤ lim cos 𝑥 ≤ lim 1
𝑥→0 2 𝑥→0 𝑥→0

⇒ 1 ≤ lim cos 𝑥 ≤ 1 ⇒ lim cos 𝑥 = 1

𝑥→0 𝑥→0

𝑥2 𝑥2
(3) 1− ≤ cos 𝑥 ≤ 1 for 𝑥 ∈ ℝ ⇒ − ≤ cos 𝑥 − 1 ≤ 0 for 𝑥 ∈ ℝ
2 2
𝑥2 𝑥2
⇒ − ≤ cos 𝑥 − 1 ≤ for 𝑥 ∈ ℝ
2 2

𝑥2 |𝑥|2
⇒ |cos 𝑥 − 1| ≤ = for 𝑥 ∈ ℝ (dividing by |𝑥|)
2 2
|cos 𝑥−1| |𝑥| cos 𝑥−1 |𝑥|
⇒ |𝑥|
≤ for 𝑥 ∈ ℝ∗ ⇒ 0 ≤ | |≤ for 𝑥 ∈ ℝ∗
2 𝑥 2

cos 𝑥−1 |𝑥|

The squeeze theorem ⇒ lim 0 ≤ lim | | ≤ lim
𝑥→0 𝑥→0 𝑥 𝑥→0 2

cos 𝑥−1 cos 𝑥−1 cos 𝑥−1

⇒ 0 ≤ lim | |≤0 ⇒ lim | | = 0 ⇒ lim =0
𝑥→0 𝑥 𝑥→0 𝑥 𝑥→0 𝑥

(4) Exercise
 88

Solving Problems:
Q 4.2.14)
(1) Let 𝑐 ∈ ℝ be a cluster point of 𝐴 ⊆ ℝ and 𝑓, 𝑔 are functions on 𝐴.
Show that if lim 𝑓 = 0 and 𝑔 is bounded on 𝐴, then lim 𝑓𝑔 = 0.
𝑥→𝑐 𝑥→𝑐
1
(2) Using Part (1), show that lim 𝑥sin (𝑥) = 0.
𝑥→0
𝑥−1
(3) Using Part (1), show that lim 𝑥 2 cos (𝑥ln(𝑥)) = 0
𝑥→0
Solution:
(1) lim 𝑓 = 0 ⇒ lim |𝑓(𝑥)| = 0
𝑥→𝑐 𝑥→𝑐
𝑔 is bounded on 𝐴 ⇒ ∃𝑀 > 0 such that |𝑔(𝑥)| ≤ 𝑀, ∀𝑥 ∈ 𝐴 multiply by |𝑓(𝑥)|
⇒ |𝑓(𝑥)||𝑔(𝑥)| ≤ 𝑀|𝑓(𝑥)|, ∀𝑥 ∈ 𝐴 ⇒ |𝑓(𝑥)𝑔(𝑥)| ≤ 𝑀|𝑓(𝑥)|, ∀𝑥 ∈ 𝐴
⇒ 0 ≤ |𝑓(𝑥)𝑔(𝑥)| ≤ 𝑀|𝑓(𝑥)|, ∀𝑥 ∈ 𝐴
⇒ lim 0 ≤ lim|𝑓(𝑥)𝑔(𝑥)| ≤ lim 𝑀|𝑓(𝑥)|
𝑥→𝑐 𝑥→𝑐 𝑥→𝑐
⇒ 0 ≤ lim|𝑓(𝑥)𝑔(𝑥)| ≤ 𝑀lim |𝑓(𝑥)| ⇒ 0 ≤ lim|𝑓(𝑥)𝑔(𝑥)| ≤ 𝑀(0) = 0
𝑥→𝑐 𝑥→𝑐 𝑥→𝑐
⇒ lim|𝑓(𝑥)𝑔(𝑥)| = 0 ⇒ lim𝑓(𝑥)𝑔(𝑥) = 0 ⇒ lim 𝑓𝑔 = 0
𝑥→𝑐 𝑥→𝑐 𝑥→𝑐
(2) lim 𝑥 = 0
𝑥→0
1 1
|sin (𝑥)| ≤ 1, ∀𝑥 ∈ ℝ∗ ⇒ sin (𝑥) is bounded on ℝ∗
1
Part (1) ⇒ lim 𝑥sin ( ) = 0
𝑥→0 𝑥

(3) lim 𝑥 2 = 0
𝑥→0
𝑥−1 𝑥−1
|cos (𝑥ln(𝑥))| ≤ 1, ∀𝑥 ∈ ℝ − {0,1} ⇒ cos (𝑥ln(𝑥)) is bounded on ℝ − {0,1}
𝑥−1
Part (1) ⇒ lim 𝑥 2 cos (𝑥ln(𝑥)) = 0
𝑥→0

Q 4.2.15) Let 𝑓: 𝐴 → ℝ be a function and.

(1) If lim 𝑓 > 0, then there exists a neighborhood 𝑉𝛿 (𝑐) such that
𝑥→𝑐
𝑓(𝑥) > 0, ∀𝑥 ∈ 𝐴 ∩ 𝑉𝛿 (𝑐), 𝑥 ≠ 𝑐
(2) If lim 𝑓 < 0, then there exists a neighborhood 𝑉𝛿 (𝑐) such that
𝑥→𝑐

𝑓(𝑥) < 0, ∀𝑥 ∈ 𝐴 ∩ 𝑉𝛿 (𝑐), 𝑥 ≠ 𝑐

Proof:
(1) Suppose that lim 𝑓 > 0 and let lim 𝑓 = 𝐿 ⇒ 𝐿 > 0
𝑥→𝑐 𝑥→𝑐
𝐿
Applying the definition of lim 𝑓 by taking 𝜀0 = 2: ∃𝛿 > 0 such that if 𝑥 satisfies 𝑥 ∈ 𝐴 and 0 <
𝑥→𝑐
𝐿
|𝑥 − 𝑐| < 𝛿, then 𝑓(𝑥) satisfies |𝑓(𝑥) − 𝐿| < 𝜀0 =
2
𝐿 𝐿 𝐿 3𝐿
⇒ − 2 < 𝑓(𝑥) − 𝐿 < 2 ⇒ < 𝑓(𝑥) <
2 2
𝐿
⇒ 𝑓(𝑥) > 2 > 0 for all 𝑥 ∈ 𝐴 and 0 < |𝑥 − 𝑐| < 𝛿
⇒ 𝑓(𝑥) > 0, ∀𝑥 ∈ 𝐴 ∩ 𝑉𝛿 (𝑐), 𝑥 ≠ 𝑐
 89

(2) lim 𝑓 < 0 ⇒ lim(−𝑓) > 0

𝑥→𝑐 𝑥→𝑐
Part (1) ⇒ there exists a neighborhood 𝑉𝛿 (𝑐) such that
−𝑓(𝑥) > 0, ∀𝑥 ∈ 𝐴 ∩ 𝑉𝛿 (𝑐), 𝑥 ≠ 𝑐
⇒ there exists a neighborhood 𝑉𝛿 (𝑐) such that
𝑓(𝑥) < 0, ∀𝑥 ∈ 𝐴 ∩ 𝑉𝛿 (𝑐), 𝑥 ≠ 𝑐

Q 4.2.17) Let 𝑓: ℝ → ℝ such that 𝑓(𝑥 + 𝑦) = 𝑓(𝑥) + 𝑓(𝑦), ∀𝑥, 𝑦 ∈ ℝ and let lim 𝑓 be exists.
𝑥→0
(1) Show that lim 𝑓 = 0
𝑥→0
(2) Show that lim 𝑓 = 𝑓(𝑐), ∀𝑐 ∈ ℝ
𝑥→𝑐
1
(Hint: For Part (1): Show that 𝑓(𝑛𝑥) = 𝑛𝑓(𝑥), ∀𝑛 ∈ ℕ and then replace 𝑥 by 𝑛 and
then use the Sequential Criterion for Limits 4.1.12)

Exercises of Section 4.2 (page 116): 1-5,9,10,11

 90

Section 4.3: Some Extensions of the Limit Concept

Part 1: One-Sided Limits

Definition 4.3.1: Let 𝐴 ⊆ ℝ and 𝑓: 𝐴 → ℝ.

(1) If 𝑐 ∈ ℝ is a cluster point of the set 𝐴 ∩ [𝑐, ∞) = {𝑥 ∈ 𝐴: 𝑥 ≥ 𝑐}, then we say that 𝐿 ∈ ℝ is a
right-hand limit of 𝑓 at 𝑐 and we write:
lim+ 𝑓 = 𝐿or lim+ 𝑓(𝑥) = 𝐿
𝑥→𝑐 𝑥→𝑐
if given any 𝜀 > 0, ∃𝛿 > 0 such that
𝑥 ∈ 𝐴 satisfies 0 < 𝑥 − 𝑐 < 𝛿 ⇒ |𝑓(𝑥) − 𝐿| < 𝜀.
(2) If 𝑐 ∈ ℝ is a cluster point of the set 𝐴 ∩ (−∞, 𝑐] = {𝑥 ∈ 𝐴: 𝑥 ≤ 𝑐}, then we say that 𝐿 ∈ ℝ is
a left-hand limit of 𝑓 at 𝑐 and we write:
lim− 𝑓 = 𝐿or lim− 𝑓(𝑥) = 𝐿
𝑥→𝑐 𝑥→𝑐
if given any 𝜀 > 0, ∃𝛿 > 0 such that:
𝑥 ∈ 𝐴 satisfies 0 < 𝑐 − 𝑥 < 𝛿 ⇒ |𝑓(𝑥) − 𝐿| < 𝜀.

Remark 4.3.2:
(1) The limits lim+ 𝑓 and lim− 𝑓 are called one-sided limits of 𝑓 at 𝑐.
𝑥→𝑐 𝑥→𝑐
(2) For the values of lim+ 𝑓 and lim− 𝑓, it is possible that:
𝑥→𝑐 𝑥→𝑐
(a) Both of them exist and equal: lim+ 𝑓 = lim− 𝑓 = 𝐿 ∈ ℝ
𝑥→𝑐 𝑥→𝑐
(b) Both of them exist and are different: lim+ 𝑓 ≠ lim− 𝑓
𝑥→𝑐 𝑥→𝑐
(c) Both of them does not exist
(d) One of them may be exist and the other does not exist.

Theorem 4.3.3: Let 𝑐 ∈ ℝ be a cluster point of both the sets 𝐴 ∩ (𝑐, ∞) and
𝐴 ∩ (−∞, 𝑐) and let 𝑓: 𝐴 → ℝ.
Then lim 𝑓 = 𝐿 iff lim+ 𝑓 = 𝐿 lim− 𝑓 = 𝐿
𝑥→𝑐 𝑥→𝑐 𝑥→𝑐
 91

Remark 4.3.4: We want to define the following limits:

lim 𝑓(𝑥) = ∞, 𝑐 ∈ ℝ lim 𝑓(𝑥) = −∞, 𝑐 ∈ ℝ lim 𝑓(𝑥) = ∞, 𝑐 ∈ ℝ

𝑥→𝑐 𝑥→𝑐 𝑥→𝑐 +

lim 𝑓(𝑥) = −∞, 𝑐 ∈ ℝ lim 𝑓(𝑥) = ∞, 𝑐 ∈ ℝ lim 𝑓(𝑥) = −∞, 𝑐 ∈ ℝ

𝑥→𝑐 + 𝑥→𝑐 − 𝑥→𝑐 −

lim 𝑓(𝑥) = 𝐿, 𝐿 ∈ ℝ lim 𝑓(𝑥) = ∞ lim 𝑓(𝑥) = −∞

𝑥→∞ 𝑥→∞ 𝑥→∞

lim 𝑓(𝑥) = 𝐿, 𝐿 ∈ ℝ lim 𝑓(𝑥) = ∞ lim 𝑓(𝑥) = −∞

𝑥→−∞ 𝑥→−∞ 𝑥→−∞

 Observe that when studying lim 𝑓 = −∞ multiplying by −1 we have

𝑥→??
lim 𝑓 = −∞ iff lim (−𝑓) = ∞
𝑥→?? 𝑥→??
 Observe that when studying lim 𝑓(𝑥) =? ? :
𝑥→−∞
Let 𝑡 = −𝑥 so when 𝑥 → −∞ ⇒ 𝑡 → ∞
∴ lim 𝑓(𝑥) = lim 𝑓(−𝑡)
𝑥→−∞ 𝑡→∞

So it is enough to study only 5 types of limits:

lim 𝑓(𝑥) = ∞, 𝑐 ∈ ℝ lim 𝑓(𝑥) = ∞, 𝑐 ∈ ℝ lim 𝑓(𝑥) = ∞, 𝑐 ∈ ℝ

𝑥→𝑐 𝑥→𝑐 + 𝑥→𝑐 −

lim 𝑓(𝑥) = 𝐿, 𝐿 ∈ ℝ lim 𝑓(𝑥) = ∞

𝑥→∞ 𝑥→∞

Definition 4.3.5:
(1) lim 𝑓(𝑥) = ∞ iff ∀𝛼 > 0, ∃𝛿 > 0 such that [𝑥 ∈ 𝐴 and 0 < |𝑥 − 𝑐| < 𝛿 ⇒ 𝑓(𝑥) > 𝛼].
𝑥→𝑐
(2) lim 𝑓(𝑥) = ∞ iff ∀𝛼 > 0, ∃𝛿 > 0 such that [𝑥 ∈ 𝐴 and 0 < 𝑥 − 𝑐 < 𝛿 ⇒ 𝑓(𝑥) > 𝛼].
𝑥→𝑐 +
(3) lim 𝑓(𝑥) = ∞ iff ∀𝛼 > 0, ∃𝛿 > 0 such that [𝑥 ∈ 𝐴 and 0 < 𝑐 − 𝑥 < 𝛿 ⇒ 𝑓(𝑥) > 𝛼].
𝑥→𝑐 −
(4) lim 𝑓(𝑥) = 𝐿 iff ∀𝜀 > 0, ∃𝑀 > 0 such that [𝑥 ∈ 𝐴 and 𝑥 > 𝛼 ⇒ |𝑓(𝑥) − 𝐿| < 𝜀].
𝑥→∞
(5) lim 𝑓(𝑥) = ∞ iff ∀𝛼 > 0, ∃𝑀 > 0 such that [𝑥 ∈ 𝐴 and 𝑥 > 𝑀 ⇒ 𝑓(𝑥) > 𝛼].
𝑥→∞
 92

Example 4.3.6: Using definition, show that

1 1 𝑥2
(1) lim 𝑥 2 = ∞. (2) lim = 0. (3) lim =0
𝑥→0 𝑥→∞ 𝑥 2 𝑥→−∞ 𝑥 4 +1
Proof:
1 1
(1) lim 𝑥 2 = ∞ ⇒ 𝑓(𝑥) = 𝑥 2 and 𝑐 = 0 ⇒ |𝑥 − 𝑐| = |𝑥 − 0| = |𝑥| <?
𝑥→0
Let 𝛼 > 0 be given. Want to find 𝛿 > 0 such that 0 < |𝑥 − 𝑐| < 𝛿 ⇒ 𝑓(𝑥) > 𝛼:
1 1 1 1
𝑓(𝑥) > 𝛼 ⇔ 𝑥 2 > 𝛼 ⇔ √𝑥 2 > √𝛼 ⇔ |𝑥|
> √𝛼 ⇔ |𝑥| < ………..①
√𝛼
1
Take 0 < 𝛿 ≤ . So, 0 < |𝑥| < 𝛿 ⇒ 𝑓(𝑥) > 𝛼
√𝛼
1
∴ lim =∞
𝑥→0 𝑥 2
1 1
(2) lim = 0 ⇒ 𝑓(𝑥) = 𝑥 2 and 𝐿 = 0
𝑥→∞ 𝑥 2
Let 𝜀 > 0 be given. Want to find 𝑀 > 0 such that 𝑥 > 𝑀 ⇒ |𝑓(𝑥) − 𝐿| < 𝜀:
1 1
|𝑓(𝑥) − 𝐿| = | 2 − 0| = 2
𝑥 𝑥
1 1 1
⇒ 𝑥2 < 𝜀 ⇔ |𝑥|
< √𝜀 ⇔ |𝑥| >
√𝜀
1
Take 𝑀 ≥ . So, 𝑥 > 𝑀 ⇒ |𝑓(𝑥) − 𝐿| < 𝜀
√ 𝜀
1
∴ lim =0
𝑥→∞ 𝑥 2
𝑥2 𝑥2
(3) lim = 0 ⇒ 𝑓(𝑥) = 𝑥 4 +1 and 𝐿 = 0
𝑥→−∞ 𝑥 4 +1
Let 𝜀 > 0 be given. Want to find 𝑀 < 0 such that 𝑥 < 𝑀 ⇒ |𝑓(𝑥) − 𝐿| < 𝜀:
𝑥2 𝑥2 𝑥2 1
|𝑓(𝑥) − 𝐿| = | 4 − 0| = 4 ≤ 4= 2
𝑥 +1 𝑥 +1 𝑥 𝑥
1 2 1 1
< 𝜀 ⇔ 𝑥 > 𝜀 ⇔ |𝑥| > 𝜀 ……①
𝑥2 √
1 1
But we want 𝑥 < 𝑀 and 𝑀 < 0 so we have 𝑥 < 0 and ①: |𝑥| > ⇒ −𝑥 >
√𝜀 √𝜀
1
⇒𝑥<−
√𝜀
−1
Take 𝑀 ≤ . So, 𝑥 < 𝑀 ⇒ |𝑓(𝑥) − 𝐿| < 𝜀.
√𝜀
𝑥2
∴ lim =0
𝑥→−∞ 𝑥 4 +1

The Sequential Criteria for One-Sided Limits 4.3.7:

(1) Let 𝑐 ∈ ℝ be a cluster point of the set {𝑥 ∈ 𝐴: 𝑥 > 𝑐} and let 𝑓: 𝐴 → ℝ be a function.
If 𝐿 ∈ ℝ or 𝐿 = ∞ or 𝐿 = −∞, then the following are equivalent:
(a) lim+ 𝑓 = 𝐿
𝑥→𝑐
(b) ∀ sequence (𝑥𝑛 ) in 𝐴 with 𝑥𝑛 → 𝑐 + we have lim 𝑓(𝑥𝑛 ) = 𝐿.
𝑛→∞
(2) Let 𝑐 ∈ ℝ be a cluster point of the set {𝑥 ∈ 𝐴: 𝑥 < 𝑐} and let 𝑓: 𝐴 → ℝ be a function.
If 𝐿 ∈ ℝ or 𝐿 = ∞ or 𝐿 = −∞, then the following are equivalent:
(a) lim− 𝑓 = 𝐿
𝑥→𝑐
(b) ∀ sequence (𝑥𝑛 ) in 𝐴 with 𝑥𝑛 → 𝑐 − we have lim 𝑓(𝑥𝑛 ) = 𝐿.
𝑛→∞
 93

Theorem 4.3.8: Let 𝑐 ∈ ℝ and let 𝑓: (𝑐, ∞) → ℝ such that 𝑓(𝑥) > 0, ∀𝑥 ∈ (𝑐, ∞). Then
1
(1) lim+ 𝑓 = ∞ iff lim+ 𝑓 = 0
𝑥→𝑐 𝑥→𝑐
1
(2) lim+ 𝑓 = 0 iff lim+ 𝑓 = ∞
𝑥→𝑐 𝑥→𝑐
Proof:
(1) lim+ 𝑓 = ∞ iff ∀ sequence (𝑥𝑛 ) in 𝐴 with 𝑥𝑛 → 𝑐 + we have lim 𝑓(𝑥𝑛 ) = ∞
𝑥→𝑐 𝑛→∞
1
iff ∀ sequence (𝑥𝑛 ) in 𝐴 with 𝑥𝑛 → 𝑐 + we have lim =0
𝑛→∞ 𝑓(𝑥𝑛 )
1
iff lim+ 𝑓 = 0
𝑥→𝑐
1 1
(2) lim+ 𝑓 = ∞ ⏟ iff lim 1 = 0 iff lim+ 𝑓 = 0
𝑥→𝑐 𝑥→𝑐 + 𝑓 𝑥→𝑐
by Part (1)

Theorem 4.3.9: Let 𝑓: (0, ∞) → ℝ be a function and 𝐿 ∈ ℝ. Then

1
lim 𝑓(𝑥) = 𝐿 iff lim+ 𝑓 (𝑥) = 𝐿
𝑥→∞ 𝑥→0
Proof: lim 𝑓 = 𝐿 iff ∀ sequence (𝑥𝑛 ) in (0, ∞) with 𝑥𝑛 → ∞ we have lim 𝑓(𝑥𝑛 ) = 𝐿
𝑥→∞ 𝑛→∞
1 +
iff ∀ sequence (𝑥𝑛 ) in (0, ∞) with 𝑥 → 0 we have lim 𝑓(𝑥𝑛 ) = 𝐿
𝑛 𝑛→∞
+ 1
iff ∀ sequence (𝑦𝑛 ) in (0, ∞) with 𝑦𝑛 → 0 we have lim 𝑓 (𝑦 ) = 𝐿
𝑛→∞ 𝑛
1 1
(take 𝑦𝑛 = 𝑥 ⇔ 𝑥𝑛 = 𝑦 )
𝑛 𝑛
1
iff lim 𝑓 (𝑥) = 𝐿
𝑥→𝑐

Theorem 4.3.10: Let 𝑓 and 𝑔 be functions defined on a set 𝐴 ⊆ ℝ such that

0 ≤ 𝑓(𝑥) ≤ 𝑔(𝑥), ∀𝑥 ∈ 𝐴
(1) If lim 𝑔(𝑥) = 0, then lim 𝑓(𝑥) = 0.
𝑥→𝑐 𝑥→𝑐
(2) If lim 𝑓(𝑥) = ∞, then lim 𝑔(𝑥) = ∞.
𝑥→𝑐 𝑥→𝑐
Where 𝑐 may be a cluster point of 𝐴 or 𝑐 = ∞ or 𝑐 = −∞.
Proof:
(1) Suppose that lim 𝑔(𝑥) = 0. Want to show that lim 𝑓(𝑥) = 0 using the sequential criterion.
𝑥→𝑐 𝑥→𝑐
lim 𝑔(𝑥) = 0 ⇒ ∀ sequence (𝑥𝑛 ) in 𝐴 with 𝑥𝑛 → 𝑐 we have lim 𝑔(𝑥𝑛 ) = 0……….①
𝑥→𝑐 𝑛→∞
But 0 ≤ 𝑓(𝑥) ≤ 𝑔(𝑥), ∀𝑥 ∈ 𝐴 ⇒ 0 ≤ 𝑓(𝑥𝑛 ) ≤ 𝑔(𝑥𝑛 ), ∀𝑛 ∈ ℕ……..②
①&②& the squeeze theorem: ⇒ lim 𝑓(𝑥𝑛 ) = 0
𝑛→∞
∴ ∀ sequence (𝑥𝑛 ) in 𝐴 with 𝑥𝑛 → 𝑐 we have lim 𝑓(𝑥𝑛 ) = 0 ⇒ lim 𝑓(𝑥) = 0.
𝑥→𝑐
(2) Suppose that lim 𝑓(𝑥) = ∞. Want to show that lim 𝑔(𝑥) = ∞ using the sequential criterion.
𝑥→𝑐 𝑥→𝑐
lim 𝑓(𝑥) = ∞ ⇒ ∀ sequence (𝑥𝑛 ) in 𝐴 with 𝑥𝑛 → 𝑐 we have lim 𝑓(𝑥𝑛 ) = ∞……….③
𝑥→𝑐 𝑛→∞
But 0 ≤ 𝑓(𝑥) ≤ 𝑔(𝑥), ∀𝑥 ∈ 𝐴 ⇒ 𝑓(𝑥𝑛 ) ≤ 𝑔(𝑥𝑛 ), ∀𝑛 ∈ ℕ……..④
③&④& the Comparison Test 6 in Section 3.6: ⇒ lim 𝑔(𝑥𝑛 ) = ∞
𝑛→∞
∴ ∀ sequence (𝑥𝑛 ) in 𝐴 with 𝑥𝑛 → 𝑐 we have lim 𝑔(𝑥𝑛 ) = ∞
𝑛→∞
⇒ lim 𝑔(𝑥) = ∞.
𝑥→𝑐

Exercises 4.3.11 (page 123): 3,5,7

 94

Chapter 5: Continuous Functions

Section 5.1: Continuous Functions

Definition 5.1.1: Let 𝐴 ⊆ ℝ and 𝑓: 𝐴 → ℝ, and let 𝑐 ∈ 𝐴. We say that 𝑓 is continuous at 𝑐 if

∀𝜀 > 0, ∃𝛿 > 0 such that [𝑥 ∈ 𝐴 with |𝑥 − 𝑐| < 𝛿 ⇒ |𝑓(𝑥) − 𝑓(𝑐)| < 𝜀].
If 𝑓 fails to be continuous at 𝑐, then we say that 𝑓 is discontinuous at 𝑐 that is ∃𝜀0 > 0 and
∀𝛿 > 0 such that [∃𝑥0 ∈ A with |𝑥0 − 𝑐| < 𝛿 ⇒ |𝑓(𝑥0 ) − 𝑓(𝑐)| ≥ 𝜀0 ]
𝑓 is continuous on a set 𝐵 ⊆ 𝐴 if 𝑓 is continuous at every point of 𝐵.

Example 5.1.2: Using the given graph of a function 𝑓, if 𝜀 > 0 is as in the figure then approximate
the larges value of 𝛿 that satisfies the definition of the continuity of the function at 𝑥 = 0.5.

0.5 1 1.5

Solution: 𝑓 is continuous at 𝑥 = 0.5 ⇔ lim 𝑓 = 𝑓(0.5) = 1. From the figure we find that the
𝑥→0.5
𝜀
largest value of 𝛿 is 𝛿 ≅ 3.
Remark 5.1.3: Recall that
𝑎−𝑏 𝑎+𝑏
(1) sin 𝑎 − sin 𝑏 = 2 sin ( ) cos ( )
2 2
𝑎−𝑏 𝑎+𝑏
(2) cos 𝑎 − cos 𝑏 = 2 sin ( ) sin ( )
2 2
 95

Example 5.1.4: Using definition, show that:

(1) 𝑓(𝑥) = sin 𝑥 is continuous on ℝ. (2) 𝑓(𝑥) = cos 𝑥 is continuous on ℝ.
Solution: (1) Let 𝑐 ∈ ℝ and 𝑓(𝑥) = sin 𝑥.
Let 𝜀 > 0 be given. Want to find 𝛿 > 0 such that |𝑥 − 𝑐| < 𝛿 ⇒ |𝑓(𝑥) − 𝑓(𝑐)| < 𝜀:
Now,
𝑥−𝑐 𝑥+𝑐 𝑥−𝑐 𝑥+𝑐 𝑥−𝑐
|𝑓(𝑥) − 𝑓(𝑐)| = |sin 𝑥 − sin 𝑐| = 2 |sin ( ) cos ( )| = 2 |sin ( )| |cos ( )| ≤ 2 | |
2 2 2 2 2
⇒ |𝑓(𝑥) − 𝑓(𝑐)| ≤ |𝑥 − 𝑐|
Letting |𝑥 − 𝑐| < 𝜀. So, take 0 < 𝛿 ≤ 𝜀 we have |𝑥 − 𝑐| < 𝛿 ⇒ |𝑓(𝑥) − 𝑓(𝑐)| < 𝜀
⇒ 𝑓(𝑥) = sin 𝑥 is continuous at 𝑐 and since 𝑐 is arbitrary in ℝ, we have:
𝑓(𝑥) = sin 𝑥 is continuous on ℝ
(1) Let 𝑐 ∈ ℝ and 𝑓(𝑥) = cos 𝑥. Let 𝜀 > 0 be given. Want to find 𝛿 > 0 such that if 𝑥 satisfies
Now,
𝑥−𝑐 𝑥+𝑐 𝑥−𝑐 𝑥+𝑐 𝑥−𝑐
|𝑓(𝑥) − 𝑓(𝑐)| = |cos 𝑥 − cos 𝑐| = 2 |sin ( ) sin ( )| = 2 |sin ( )| |sin ( )| ≤ 2 | |
2 2 2 2 2
⇒ |𝑓(𝑥) − 𝑓(𝑐)| ≤ |𝑥 − 𝑐|
Letting |𝑥 − 𝑐| < 𝜀. So, take 0 < 𝛿 ≤ 𝜀 we have |𝑥 − 𝑐| < 𝛿 ⇒ |𝑓(𝑥) − 𝑓(𝑐)| < 𝜀
⇒ 𝑓(𝑥) = cos 𝑥 is continuous at 𝑐 and since 𝑐 is arbitrary in ℝ, we have:
𝑓(𝑥) = cos 𝑥 is continuous on ℝ

Definition 5.1.5:
(1) a point 𝑐 ∈ ℝ is called a cluster point of a set 𝑨 if ∃(𝑥𝑛 ) in 𝐴 such that 𝑥𝑛 ≠ 𝑐, ∀𝑛 ∈ ℕ
and 𝑥𝑛 → 𝑐.
(2) a point 𝑐 ∈ ℝ is called an isolated point of a set 𝑨 if 𝑐 is not a cluster point of 𝐴.

Remark 5.1.6: 𝐵 is the set of all cluster points of a set 𝐴 iff ℝ − 𝐵 is the set of all isolated points of 𝐴.

Example 5.1.7: Find the sets of all cluster points and the set of all isolated points of the following sets:
(1) 𝐴 = [0,1] (2) 𝐴 = (0,1) (3) 𝐴 = {0} ∪ (1, ∞).
Solution:
(1) The sets of all cluster points of 𝐴 is [0,1] = 𝐴
The set of all isolated points of 𝐴 is ℝ\[0,1] = (−∞, 0) ∪ (1, ∞).
(2) The sets of all cluster points of 𝐴 is [0,1]
The set of all isolated points of 𝐴 is ℝ\[0,1] = (−∞, 0) ∪ (1, ∞).
(3) The sets of all cluster points of 𝐴 is [1, ∞)
The set of all isolated points of 𝐴 is ℝ\[1, ∞) = (−∞, 1).

Theorem 5.1.8: If 𝑐 ∈ 𝐴 is an isolated point of 𝐴 and 𝑓: 𝐴 → ℝ is a function, then 𝑓 is continuous

at 𝑐.

Proof: Let 𝑐 ∈ 𝐴 is an isolated point of 𝐴.

Let 𝜀 > 0 be given. Take 𝛿 = inf {|𝑥 − 𝑐|: 𝑥 ∈ 𝐴, 𝑥 ≠ 𝑐} ⇒ 𝛿 > 0.
𝑥 ∈ 𝐴 with |𝑥 − 𝑐| < 𝛿 ⇒ 𝑥 = 𝑐 only ⇒ |𝑓(𝑥) − 𝑓(𝑐)| = |𝑓(𝑐) − 𝑓(𝑐)| = 0 < 𝜀
So, |𝑥 − 𝑐| < 𝛿 ⇒ |𝑓(𝑥) − 𝑓(𝑐)| < 𝜀 ⇒ 𝑓 is continuous at 𝑐.
 96

1 ,𝑥 = 0
Example 5.1.9: Determine whether 𝑓 is continuous at 𝑥 = 0, where 𝑓(𝑥) = { }
𝑥2 ,𝑥 ≥ 1

4
3
Solution: Observe that 0 ∈ Dom(𝑓) = {0} ∪ (1, ∞)
⇒ 𝑥 = 0 is an isolated point of Dom(𝑓) 2
⇒ 𝑓 continuous at 𝑥 = 0 (by Theorem 4).. 1

1 2

Note
‫) في مجاالت االقترانات هي دراسة ذات اهمية تطبيقية‬isolated points( ‫يجب العلم بأن دراسة االتصال عند النقاط المعزولة‬
‫ أما النقاط غير المعزولة (أي النقاط العنقودية‬،‫ وذلك الى كون االقترانات دائما متصلة عند هذه النقاط‬،‫قليلة‬
‫ لذلك سنركز في دراستنا على االتصال عند النقاط‬،‫)) فيمكن أن يكون االقتران متصال أو غير متصل عندها‬cluster points(
.‫العنقودية في مجاالت االقترانات‬
‫بما أن مفهوم النهاية مرتبط بالنقاط العنقودية وكذلك مرتبط بمفهوم الجوار فسنعمل في النظرية التالية على ربط مفهوم االتصال‬
.‫بمفهوم النهاية عند النقاط العنقودية فقط‬

Theorem 5.1.10: Let 𝑐 ∈ 𝐴 be a cluster point of 𝐴 and let 𝑓: 𝐴 → ℝ be a function.

Then 𝑓 is continuous at 𝑥 = 𝑐 iff lim 𝑓 = 𝑓(𝑐).
𝑥→𝑐

Remark 5.1.11: According to Theorem 6: Let 𝑐 ∈ Dom(𝑓) be a cluster point of 𝐴. Then

𝑓 is continuous at 𝑥 = 𝑐 iff ①: 𝑐 ∈ Dom(𝑓), that is 𝑓(𝑐) is defined
②: lim 𝑓 exists (equals a number)
𝑥→𝑐
③: the numbers 𝑓(𝑐) and lim 𝑓 must be equal
𝑥→𝑐

Remark 5.1.12: To study the continuity of a function 𝑓 at a point 𝑐, we have two possible cases
for 𝑐:
 Case 1: 𝑐 ∈ Domain (𝑓): We have two cases:
 Case 1.1: 𝑐 is an isolated point in Dom (𝑓) ⇒ 𝑓 is continuous at 𝑐.
 Case 1.2: 𝑐 is a cluster point of Dom (𝑓) ⇒ We have to compare the values of
𝑓(𝑐) and lim 𝑓
𝑥→𝑐
 If lim 𝑓 = 𝑓(𝑐) ⇒ 𝑓 is continuous at 𝑐
𝑥→𝑐
 If lim 𝑓 DNE or lim 𝑓 ≠ 𝑓(𝑐) ⇒ 𝑓 is discontinuous at 𝑐
𝑥→𝑐 𝑥→𝑐
 Case 2: 𝑐 ∉ Domain (𝑓), to say 𝑓 is discontinuous at 𝑐 we must have 𝑐 a cluster point of
Dom (𝑓) otherwise we do not talk about the continuity of functions at isolated points not in
its domain (see the following example):
 97

Example 5.1.13: Let 𝑓(𝑥) = √𝑥.

1) Is 𝑓 continuous at 𝑥 = 0?
2) Find the set of all points at which the function 𝑓 is
discontinuous
3) Find the set of all points at which the function 𝑓 is continuous? 1
Solution: Dom(𝑓) = [0, ∞)
1) 𝑥 = 0 is a cluster point of Dom(𝑓): -2 -1 1 2 3
lim 𝑓 = lim+ 𝑓 = 0 = 𝑓(0) ⇒ 𝑓 continuous at 𝑥 = 0
𝑥→0 𝑥→0
2) The set of discontinuities of 𝑓 is 𝜑 = { }
3) 𝑓 is continuous on [0, ∞)
Example 5.1.14: Let 𝑓 be the function given in the figure, 2
find the set of all points at which 𝑓 is continuous.
1
Solution: First we find the domain of 𝑓 from the figure:
Dom(𝑓) = (−∞, −2) ∪ (−2,0.5] ∪ {1} ∪ [1.5, ∞)
-2 -1.5 -1 -0.5 0.5 1 1.5 2
-1

-2

Step 1: Isolated points in 𝐃𝐨𝐦 (𝒇): 𝑥 = 1 is the only one isolated point of Dom (𝑓) which is in
Dom (𝑓) ⇒ 𝑓 is continuous at 𝑥 = 1.
Step 2: The remaining points in Dom(𝑓) are cluster points of Dom (𝑓) which are in Dom (𝑓):
⇒ 𝑓 is continuous on (−∞, −2) ∪ (−2, −1) ∪ (−1,0) ∪ (0,0.5] ∪ [1.5, ∞)
Conclusion: 𝑓 is continuous on (−∞, −2) ∪ (−2, −1) ∪ (−1,0) ∪ (0,0.5] ∪ {1} ∪ [1.5, ∞)

Remark 5.1.15: From now on (ً ‫ )من اآلن فصاعدا‬when we talk about the continuity of a function
𝑓 at a point 𝑐, we are assuming that 𝑐 is a cluster point of Dom(𝑓).
1 ,𝑥 > 0
Example 5.1.16: Show that the signum function sgn(𝑥) = { 0 , 𝑥 = 0} is discontinuous at 𝑥 = 0.
−1 , 𝑥 < 0
Solution: 0 ∈ Dom(sgn(𝑥)) = ℝ and 𝑥 = 0 is a cluster point of Dom(sgn(𝑥)):
lim+ sgn(𝑥) = 1 & lim− sgn(𝑥) = −1 ⇒ lim sgn(𝑥) DNE ⇒ 𝑓 is discontinuous at 𝑥 = 0.
𝑥→0 𝑥→0 𝑥→0

The Sequential Criteria for Continuity 5.1.17:

Let 𝑓: 𝐴 → ℝ be a function. Then the following are equivalent:
(1) 𝑓 is continuous at a point 𝑐 ∈ 𝐴, that is lim 𝑓 = 𝑓(𝑐).
𝑥→𝑐
(2) ∀ sequence (𝑥𝑛 ) in 𝐴 with 𝑥𝑛 → 𝑐 we have lim 𝑓(𝑥𝑛 ) = 𝑓(𝑐).
𝑛→∞
(3) ∀ sequence (𝑥𝑛 ) in 𝐴 with 𝑥𝑛 → 𝑐 we have
lim 𝑓(𝑥2𝑛 ) = 𝑓(𝑐) and lim 𝑓(𝑥2𝑛+1 ) = 𝑓(𝑐).
𝑛→∞
(4) ∀ sequences (𝑟𝑛 ) in ℚ and (𝑟𝑛′ ) in ℚ𝑐 with 𝑟𝑛 → 𝑐 and 𝑟𝑛′ → 𝑐 we have lim 𝑓(𝑟𝑛 ) = 𝑓(𝑐) and
𝑛→∞
lim 𝑓(𝑟𝑛′ ) = 𝑓(𝑐).
𝑛→∞
 98

Sequential Discontinuity Criteria 5.1.18: Let 𝑓: 𝐴 → ℝ be a function and 𝑐 ∈ 𝐴 be a cluster point

of 𝐴. Then the following are equivalent:
(1) 𝑓 is discontinuous at 𝑥 = 𝑐
(2) ∃ a sequence (𝑥𝑛 ) in 𝐴 with 𝑥𝑛 → 𝑐 such that lim 𝑓(𝑥𝑛 ) ≠ 𝑓(𝑐).
𝑛→∞

Remark 5.1.19: Observe that lim 𝑓(𝑥𝑛 ) ≠ 𝑓(𝑐) has two meanings:
𝑛→∞
 It might be lim 𝑓(𝑥𝑛 ) DNE (does not exist) so we have lim 𝑓(𝑥𝑛 ) ≠ 𝑓(𝑐)
𝑛→∞
 It might be lim 𝑓(𝑥𝑛 ) exists but does not equal to 𝑓(𝑐) so we have lim 𝑓(𝑥𝑛 ) ≠ 𝑓(𝑐)
𝑛→∞

1
sin (𝑥) ,𝑥 ≠ 0
Example 5.1.20: Determine whether the function 𝑓(𝑥) = { } is continuous at 𝑥 = 0.
1 ,𝑥 = 0
Solution: 𝑓(0) = 1 and 𝑥 = 0 is a cluster point of Dom(𝑓) = ℝ.
1
Observe that lim 𝑓 = lim sin (𝑥) DNE (by using the Sequential Criterion for discontinuity:
𝑥→0 𝑥→0
1 1 1
Consider the sequence (𝑛) ⇒ 𝑛 ∈ Dom(𝑓) and 𝑛 → 0
1 1
∴ lim 𝑓 (𝑛) = lim sin ( 1 ) = lim sin(𝑛) DNE (by Example 7 of Section 3.4)
𝑛→∞ 𝑛→∞ 𝑛
𝑛→∞
1
⇒ lim 𝑓 (𝑛) ≠ 𝑓(0) ⇒ 𝑓 is discontinuous at 𝑥 = 0.
𝑛→∞

Theorem 5.1.21: Let 𝑓 and 𝑔 be continuous functions on ℝ. Define the function 𝛹 by

𝑓(𝑥) , 𝑥 ∈ ℚ
𝛹(𝑥) = { }.
𝑔(𝑥) , 𝑥 ∈ ℚ𝑐
Then 𝛹 is continuous only at the points 𝑐 ∈ ℝ iff 𝑓(𝑐) = 𝑔(𝑐).

𝑥3 + 𝑥 ,𝑥 ∈ ℚ
Example 5.1.22: Find all points in ℝ at which the function 𝛹(𝑥) = { } is
−2𝑥 2 , 𝑥 ∈ ℚ𝑐
continuous.
Solution: 𝛹 is continuous at the points when 𝑥 3 + 𝑥 = −2𝑥 2 ⇒ 𝑥(𝑥 + 1)2 = 0
⇒ 𝑥 = 0 or 𝑥 = −1 ⇒ 𝛹 is continuous only at 𝑥 = 0 and 𝑥 = −1
1 , 𝑥 is rational
Example 5.1.23: Show that Dirichlet’s 𝛹(𝑥) = { } function is nowhere
0 , 𝑥 is irrational
continuous.
Solution: 𝛹 is continuous at the points when 1 = 0 ⇒ 𝛹 is discontinuous at all points of ℝ ⇒ 𝛹
is nowhere continuous
 99

Solving Problems:

Problem 5.1.25: Let 𝑓: 𝐴 → ℝ be continuous at 𝑐 ∈ 𝐴. Show that ∀𝜀 > 0, ∃𝛿 > 0 such that
if 𝑥, 𝑦 ∈ 𝐴 with |𝑥 − 𝑐| < 𝛿 and |𝑦 − 𝑐| < 𝛿, then |𝑓(𝑥) − 𝑓(𝑦)| < 𝜀
Proof: Let 𝜀 > 0 be given.
𝜀
Since 𝑓 is continuous at 𝑐, we apply the definition of continuity for 2: ⇒ ∃𝛿 > 0 such that
𝜀
If 𝑡 ∈ 𝐴 with |𝑡 − 𝑐| < 𝛿 , then |𝑓(𝑡) − 𝑓(𝑐)| < 2………①
Let 𝑥, 𝑦 ∈ 𝐴 with |𝑥 − 𝑐| < 𝛿 and |𝑦 − 𝑐| < 𝛿
𝜀 𝜀
① ⇒ |𝑓(𝑥) − 𝑓(𝑐)| < 2 and |𝑓(𝑦) − 𝑓(𝑐)| < 2……….②
Now,
|𝑓(𝑥) − 𝑓(𝑦)| = |(𝑓(𝑥) − 𝑓(𝑐)) − (𝑓(𝑦) − 𝑓(𝑐))|
≤ |𝑓(𝑥) − 𝑓(𝑐)| + |𝑓(𝑦) − 𝑓(𝑐)| (by the triangle inequality)
𝜀 𝜀
< + (by ②)
2 2
=𝜀

Problem 5.1.26: Let 𝑓: 𝐴 → ℝ be continuous at 𝑐 ∈ 𝐴.

(1) Show that if 𝑓(𝑐) > 0, then ∃𝛿 > 0 such that 𝑓(𝑥) > 0, ∀𝑥 ∈ 𝐴 with |𝑥 − 𝑐| < 𝛿.
(2) Show that if 𝑓(𝑐) < 0, then ∃𝛿 > 0 such that 𝑓(𝑥) < 0, ∀𝑥 ∈ 𝐴 with |𝑥 − 𝑐| < 𝛿

Proof: We will prove part (1). The proof of part (2) is similar.
(1) suppose that 𝑓(𝑐) > 0.
𝑓(𝑐)
Since 𝑓 is continuous at 𝑐, we apply the definition of continuity for 𝜀 = 2 :
𝑓(𝑐)
⇒ ∃𝛿 > 0 such that if 𝑥 ∈ 𝐴 with |𝑥 − 𝑐| < 𝛿, then |𝑓(𝑥) − 𝑓(𝑐)| < 𝜀 = 2
𝑓(𝑐) 𝑓(𝑐)
⇒ − < 𝑓(𝑥) − 𝑓(𝑐) < , ∀𝑥 ∈ 𝐴 ∩ 𝑉𝛿 (𝑐)
2 2
𝑓(𝑐) 3𝑓(𝑐)
⇒ < 𝑓(𝑥) < , ∀𝑥 ∈ 𝐴 with |𝑥 − 𝑐| < 𝛿
2 2
𝑓(𝑐)
⇒ 𝑓(𝑥) > > 0, ∀𝑥 ∈ 𝐴 ∩ 𝑉𝛿 (𝑐) ⇒ 𝑓(𝑥) > 0, ∀𝑥 ∈ 𝐴 ∩ 𝑉𝛿 (𝑐)
2

Problem 5.1.27: Let 𝑓: ℝ → ℝ be continuous at 𝑐 ∈ ℝ and let 𝛽 ∈ ℝ.

(1) Show that if 𝑓(𝑐) > 𝛽, then ∃𝛿 > 0 such that 𝑓(𝑥) > 𝛽, ∀𝑥 ∈ 𝐴 with |𝑥 − 𝑐| < 𝛿).
(2) Show that if 𝑓(𝑐) < 𝛽, then ∃𝛿 > 0 such that 𝑓(𝑥) < 𝛽, ∀𝑥 ∈ 𝐴 with |𝑥 − 𝑐| < 𝛿.
Proof: Apply Problem 5.1.26 to the function 𝑔(𝑥) = 𝑓(𝑥) − 𝛽, 𝑥 ∈ ℝ.

Exercises 5.1.24 (page 129): 3,5,8,9,10,13

 100

Section 5.2: Combinations of Continuous Functions

In this section all points that we discuss the continuity at are assumed to be cluster points to the
domains of the functions

Theorem 5.2.1: Let 𝐴 ⊆ ℝ, 𝑓: 𝐴 → ℝ, 𝑔: 𝐴 → ℝ, and let 𝑐 ∈ 𝐴 and 𝑏 ∈ ℝ.

(1) If 𝑓 and 𝑔 are continuous at 𝑐, then 𝑏𝑓, 𝑓 + 𝑔, 𝑓 − 𝑔, and 𝑓𝑔 are continuous at 𝑐.
(2) If 𝑓 and 𝑔 are continuous on 𝐴, then 𝑏𝑓, 𝑓 + 𝑔, 𝑓 − 𝑔, and 𝑓𝑔 are continuous on 𝐴.
𝑓
(3) If 𝑔(𝑥) ≠ 0, ∀𝑥 ∈ 𝐴 and 𝑓, 𝑔 are continuous at 𝑐, then 𝑔 is continuous at 𝑐.
𝑓
(4) If 𝑔(𝑥) ≠ 0, ∀𝑥 ∈ 𝐴 and 𝑓, 𝑔 are continuous on 𝐴, then 𝑔 is continuous on 𝐴.
𝑓
Proof: We prove that 𝑓 + 𝑔 and 𝑔 are continuous at 𝑐, the others can be proved similarly:

 lim(𝑓 + 𝑔) = lim 𝑓 + lim 𝑔 = 𝑓(𝑐) + 𝑔(𝑐) = (𝑓 + 𝑔)(𝑐)

𝑥→𝑐 𝑥→𝑐 𝑥→𝑐

⇒ 𝑓 + 𝑔 is continuous at 𝑐
𝑓 lim 𝑓 𝑓(𝑐) 𝑓 𝑓
 lim = 𝑥→𝑐 = = ( ) (𝑐) ⇒ is continuous at 𝑐
𝑥→𝑐 𝑔 lim 𝑔
𝑥→𝑐
𝑔(𝑐) 𝑔 𝑔

Composition of Continuous Functions:

Definition 5.2.2: Let 𝐴, 𝐵 ⊆ ℝ, 𝑓: 𝐴 → ℝ and 𝑔: 𝐵 → ℝ be functions such that 𝑓(𝐴) ⊆ 𝐵. The

composition function 𝑔 ∘ 𝑓: 𝐴 → ℝ is defined by (𝑔 ∘ 𝑓)(𝑥) = 𝑔(𝑓(𝑥)), 𝑥 ∈ 𝐴.
 101

Theorem 5.2.3: Let 𝐴, 𝐵 ⊆ ℝ and let 𝑓: 𝐴 → ℝ and 𝑔: 𝐵 → ℝ be functions such that 𝑓(𝐵) ⊆ 𝐴.
(1) If 𝑓 is continuous at 𝑐 ∈ 𝐴 and 𝑔 is continuous at 𝑏 = 𝑓(𝑐) ∈ 𝐵, then the composition function
𝑔 ∘ 𝑓: 𝐴 → ℝ is continuous at 𝑐.
(2) If 𝑓 is continuous on 𝐴 and 𝑔 is continuous on 𝐵, then the composition function
𝑔 ∘ 𝑓: 𝐴 → ℝ is continuous on 𝐴.

Proof:
(1) Let 𝑐 ∈ 𝐴 and let 𝑏 = 𝑓(𝑐).
𝑓 is continuous at 𝑐 ⇒ lim 𝑓(𝑥) = 𝑓(𝑐)
𝑥→𝑐

𝑔 is continuous at 𝑏 ⇒ lim 𝑔 = 𝑔(𝑏) = 𝑔(𝑓(𝑐)) = (𝑔 ∘ 𝑓)(𝑐)

𝑦→𝑏

Now, 𝑦 = 𝑓(𝑥): When 𝑥 → 𝑐 we have lim 𝑦 = lim 𝑓(𝑥) = 𝑓(𝑐) = 𝑏. So, 𝑥 → 𝑐 ⇒ 𝑦 → 𝑏

𝑥→𝑐 𝑥→𝑐

So, lim 𝑔 ∘ 𝑓 = lim 𝑔(𝑓(𝑥)) = lim 𝑔(𝑦) = 𝑔(𝑏) = (𝑔 ∘ 𝑓)(𝑐) ⇒ 𝑔 ∘ 𝑓 is continuous at 𝑐.

𝑥→𝑐 𝑥→𝑐 𝑦→𝑏

(2) 𝑓 is continuous on 𝐴 and 𝑔 is continuous on 𝐵

⇒ 𝑓 is continuous at 𝑐 ∈ 𝐴 and 𝑔 is continuous 𝑓(𝑐) for all 𝑐 ∈ 𝐴
⇒ 𝑔 ∘ 𝑓 is continuous at 𝑐 for all 𝑐 ∈ 𝐴 (by part (1))
⇒ 𝑔 ∘ 𝑓 is continuous on 𝐴.

Example 5.2.4: Let 𝐴 ⊆ ℝ, 𝑓: 𝐴 → ℝ and 𝑐 ∈ 𝐴.

(1) If 𝑓 is continuous at 𝑐, then |𝑓| is continuous at 𝑐.
(2) If 𝑓 is continuous on 𝐴, then |𝑓| is continuous on 𝐴.
Proof:
(1) Let 𝑔(𝑥) = |𝑥|, 𝑥 ∈ ℝ ⇒ 𝑔 is continuous on ℝ ⇒ 𝑔 is continuous at any point in ℝ
So, we have 𝑓 is continuous at 𝑐 and 𝑔 is continuous at 𝑓(𝑐)

The Composition Theorem ⇒ 𝑔 ∘ 𝑓 is continuous at 𝑐

But 𝑔 ∘ 𝑓 = 𝑔(𝑓) = |𝑓| ⇒ |𝑓| is continuous at 𝑐.

(2) 𝑓 is continuous on 𝐴 ⇒ 𝑓 is continuous at 𝑐 for all 𝑐 ∈ 𝐴

⇒ |𝑓| is continuous at 𝑐 for all 𝑐 ∈ 𝐴 (by part (1)) ⇒ |𝑓| is continuous on 𝐴.
 102

Example 5.2.5: Let 𝐴 ⊆ ℝ, 𝑓: 𝐴 → ℝ such that 𝑓(𝑥) ≥ 0, ∀𝑥 ∈ 𝐴, and let 𝑐 ∈ 𝐴.

(1) If 𝑓 is continuous at 𝑐, then √𝑓 is continuous at 𝑐.

(2) If 𝑓 is continuous on 𝐴, then √𝑓 is continuous on 𝐴.

Proof:

(1) Let 𝑔(𝑥) = √𝑥, 𝑥 ∈ [0, ∞) ⇒ 𝑔 is continuous on [0, ∞)

⇒ 𝑔 is continuous at any point in [0, ∞)
So, we have 𝑓 is continuous at 𝑐 and 𝑔 is continuous at 𝑓(𝑐) (since 𝑓(𝑐) ≥ 0)

The Composition Theorem ⇒ 𝑔 ∘ 𝑓 is continuous at 𝑐

But 𝑔 ∘ 𝑓 = 𝑔(𝑓) = √𝑓 ⇒ √𝑓 is continuous at 𝑐.

(2) 𝑓 is continuous on 𝐴 ⇒ 𝑓 is continuous at 𝑐 for all 𝑐 ∈ 𝐴 ⇒ √𝑓 is continuous at 𝑐 for all 𝑐 ∈

𝐴 (by part (1)) ⇒ √𝑓 is continuous on 𝐴.

Exercises of Section 5.2 (pages 133-134): 2,3,5-14

 103

Section 5.3: Continuous Functions on Intervals

Definition 5.3.1: A function 𝑓: 𝐴 → ℝ is said to be bounded on 𝐴 if ∃𝑀 > 0 such that

|𝑓(𝑥)| ≤ 𝑀, ∀𝑥 ∈ 𝐴. Equivalently, −𝑀 ≤ 𝑓(𝑥) ≤ 𝑀, ∀𝑥 ∈ 𝐴
∴ 𝒇 is bounded on 𝑨 iff 𝑓(𝐴) = Range(𝑓) = {𝑓(𝑥): 𝑥 ∈ 𝐴} is bounded above
(𝑓(𝐴) has an UB) and is bounded below (𝑓(𝐴) has a LB).

1 1
𝑓(𝑥) = 𝑥 is unbounded on (0, ∞) 𝑓(𝑥) = 1+𝑥 2 is bounded on ℝ
since 𝑓((0, ∞)) = (0, ∞) has no UB since 1 is an UB and 0 is a LB of
𝑓(ℝ) = Range(𝑓) = (0,1]

Boundedness Theorem 5.3.2: If 𝑓: [𝑎, 𝑏] → ℝ is continuous on the closed bounded interval

[𝑎, 𝑏], then 𝑓 is bounded on [𝑎, 𝑏].

Remark 5.3.3: If 𝑓 is continuous on an open bounded interval (𝑎, 𝑏), then 𝑓 need not be bounded
on (𝑎, 𝑏). Consider the following example:
𝜋 𝜋
Example 5.3.4: Let 𝑓(𝑥) = tan 𝑥 is continuous on the open bounded interval (− 2 , 2 ). In fact 𝑓
𝜋 𝜋
is unbounded on (− 2 , 2 ) (see the graph of the function 𝑓 given in the figure)

Question 5.3.5: Let 𝑓: 𝐼 → ℝ be a function that is continuous on an interval 𝐼 and 𝐼 is not closed or
unbounded. When 𝑓 is bounded on 𝐼? The answer is given in the following Theorem:
Theorem 5.3.6: Let 𝐼 be an interval with end points 𝑎, 𝑏 with 𝑎 < 𝑏, and let 𝑓: 𝐼 → ℝ be a
function such that:
(1) 𝑓 is continuous on 𝐼.
(2) lim 𝑓 and lim 𝑓 both exist.
𝑥→𝑎 𝑥→𝑏
Then 𝑓 is bounded on 𝐼, where 𝐼 is any of the following intervals:
[𝑎, 𝑏], ⏟
(𝑎, 𝑏], [𝑎, 𝑏), (𝑎, 𝑏) , ⏟
(−∞, 𝑏], (−∞, 𝑏), [𝑎, ∞), (𝑎, ∞), ℝ = (−∞, ∞).
not closed unbounded
 104

Definition 5.3.7: Let 𝐴 ⊆ ℝ and let 𝑓: 𝐴 → ℝ be a function. We say that:

(3) 𝑓 has an absolute maximum on 𝐴 if ∃𝑥 ∗ ∈ 𝐴 such that 𝑓(𝑥 ∗ ) ≥ 𝑓(𝑥), ∀𝑥 ∈ 𝐴, that is 𝑓(𝑥 ∗ ) =
sup 𝑓(𝐴) = sup {𝑓(𝑥): 𝑥 ∈ 𝐴}
(4) 𝑓 has an absolute minimum on 𝐴 if ∃𝑥∗ ∈ 𝐴 such that 𝑓(𝑥∗ ) ≤ 𝑓(𝑥), ∀𝑥 ∈ 𝐴, that is
𝑓(𝑥∗ ) = inf 𝑓(𝐴) = inf {𝑓(𝑥): 𝑥 ∈ 𝐴}.
Recall that 𝑓(𝐴) = Range(𝑓) = {𝑓(𝑥): 𝑥 ∈ 𝐴}.

Maximum-Minimum Theorem 5.3.8: If 𝑓: [𝑎, 𝑏] → ℝ is continuous on the closed bounded

interval [𝑎, 𝑏], then 𝑓 has an absolute maximum and an absolute minimum on [𝑎, 𝑏], that is
∃𝑥 ∗ , 𝑥∗ ∈ [𝑎, 𝑏] such that 𝑓(𝑥 ∗ ) = sup 𝑓([𝑎, 𝑏]) and 𝑓(𝑥∗ ) = inf 𝑓([𝑎, 𝑏]).
Proof: First we prove 𝑓 has an absolute maximum on [𝑎, 𝑏]:

Since 𝑓 is continuous on the closed bounded interval [𝑎, 𝑏] ⇒ 𝑓 is bounded on [𝑎, 𝑏].
⇒ 𝑓(𝐴) = Range(𝑓) = {𝑓(𝑥): 𝑥 ∈ [𝑎, 𝑏]} has an UB and has a LB. So, Let 𝑠 ∗ = sup 𝑓([𝑎, 𝑏])
Claim 1: ∃𝑥 ∗ ∈ [𝑎, 𝑏] such that 𝑓(𝑥 ∗ ) = 𝑠 ∗ :
1
Pf of Claim 1: Since 𝑠 ∗ = sup 𝑓(𝐴) ⇒ 𝑠 ∗ − 𝑛 is not an UB of Range(𝑓), ∀𝑛 ∈ ℕ.
1
⇒ ∃𝑥𝑛 ∈ [𝑎, 𝑏] such that 𝑠 ∗ − 𝑛 < 𝑓(𝑥𝑛 ) ≤ 𝑠 ∗ , ∀𝑛 ∈ ℕ.

⇒ (𝑥𝑛 ) is a sequence in [𝑎, 𝑏] and since [𝑎, 𝑏] is a bounded interval we have: (𝑥𝑛 ) is a
bounded sequence
The Bolzano-Weierstrass Theorem ⇒ (𝑥𝑛 ) has a convergent subsequence (𝑥𝑛𝑘 ).

Let (𝑥𝑛𝑘 ) converges to 𝑢 that is lim 𝑥𝑛𝑘 = 𝑢.

𝑘→∞

Since (𝑥𝑛𝑘 ) is a subsequence of (𝑥𝑛 ) and (𝑥𝑛 ) is in [𝑎, 𝑏] we have:

𝑎 ≤ 𝑥𝑛𝑘 ≤ 𝑏, ∀𝑘 ∈ ℕ ⇒ 𝑎 ≤ lim 𝑥𝑛𝑘 ≤ 𝑏 ⇒ 𝑎 ≤ 𝑢 ≤ 𝑏 ⇒ 𝑢 ∈ [𝑎, 𝑏].

𝑘→∞

But 𝑓 is continuous on [𝑎, 𝑏] and 𝑢 ∈ [𝑎, 𝑏] ⇒ 𝑓 continuous at 𝑢: lim 𝑓 = 𝑓(𝑢)

𝑥→𝑢

Since (𝑥𝑛𝑘 ) is a sequence in [𝑎, 𝑏] with lim 𝑥𝑛𝑘 = 𝑢 and 𝑓 continuous at 𝑢, then Sequential
𝑘→∞
Criterion implies that lim 𝑓(𝑥𝑛𝑘 ) = 𝑓(𝑢)…….①
𝑘→∞
1
Since 𝑠 ∗ − < 𝑓(𝑥𝑛 ) ≤ 𝑠 ∗ , ∀𝑛 ∈ ℕ and (𝑥𝑛𝑘 ) is a subsequence of (𝑥𝑛 ) we have:
𝑛

1 1
𝑠 ∗ − 𝑛 < 𝑓(𝑥𝑛𝑘 ) ≤ 𝑠 ∗ , ∀𝑘 ∈ ℕ ⇒ lim (𝑠 ∗ − 𝑛 ) < lim 𝑓(𝑥𝑛𝑘 ) ≤ lim 𝑠 ∗
𝑘 𝑘→∞ 𝑘 𝑘→∞ 𝑘→∞

⇒ 𝑠 ∗ ≤ lim 𝑓(𝑥𝑛𝑘 ) ≤ 𝑠 ∗ ⇒ lim 𝑓(𝑥𝑛𝑘 ) = 𝑠 ∗ ⇒ 𝑓(𝑢) = 𝑠 ∗ (by ①)

𝑘→∞ 𝑘→∞
 105

∴ 𝑓(𝑢) = sup 𝑓([𝑎, 𝑏]) ≡ The absolute maximum of 𝑓. So take 𝑥 ∗ = 𝑢

∴ ∃𝑥 ∗ ∈ [𝑎, 𝑏] such that 𝑓(𝑥 ∗ ) = 𝑠 ∗ that is 𝑓 has its absolute maximum on [𝑎, 𝑏].
Claim 2: ∃𝑥∗ ∈ [𝑎, 𝑏] such that 𝑓(𝑥∗ ) = inf 𝑓(𝐴):
We apply Claim 1 to the function −𝑓: Since 𝑓 is continuous on the closed bounded interval [𝑎, 𝑏] we
have −𝑓 is continuous on the closed bounded interval [𝑎, 𝑏]
Claim 1 ⇒ −𝑓 has its absolute maximum on [𝑎, 𝑏] that is ∃𝑥∗ ∈ [𝑎, 𝑏] such that
−𝑓(𝑥∗ ) = sup(−𝑓(𝐴)) = − inf 𝑓(𝐴)

⇒ ∃𝑥∗ ∈ [𝑎, 𝑏] such that 𝑓(𝑥∗ ) = inf 𝑓(𝐴) ≡ The absolute min. of 𝑓
∴ 𝑓 has its absolute minimum on [𝑎, 𝑏].

Example 5.3.9: Show that if the function 𝑓 is continuous on the closed bounded interval [𝑎, 𝑏] and
𝑓(𝑥) > 0, ∀𝑥 ∈ [𝑎, 𝑏], then ∃𝛼 > 0 such that 𝑓(𝑥) ≥ 𝛼, ∀𝑥 ∈ [𝑎, 𝑏]
Proof: 𝑓 is continuous on the closed bounded interval [𝑎, 𝑏] ⇒ 𝑓 has its absolute minimum point at
some 𝑥∗ ∈ [𝑎, 𝑏] that is 𝑓(𝑥) ≥ 𝑓(𝑥∗ ), ∀𝑥 ∈ [𝑎, 𝑏]………①
Since 𝑥∗ ∈ [𝑎, 𝑏] and 𝑓(𝑥) > 0, ∀𝑥 ∈ [𝑎, 𝑏] ⇒ 𝑓(𝑥∗ ) > 0……….②
Take 𝛼 = 𝑓(𝑥∗ ) ⇒ 𝛼 > 0 (by ②) and 𝑓(𝑥) ≥ 𝛼, ∀𝑥 ∈ [𝑎, 𝑏] (by ①).

Location of Roots Theorem 5.3.10: Let 𝑓 be continuous on the closed bounded interval [𝑎, 𝑏] such
𝑓(𝑎)𝑓(𝑏) < 0. Then ∃𝑐 ∈ (𝑎, 𝑏) such that 𝑓(𝑐) = 0, that is the equation 𝑓(𝑥) = 0 has a real root
that lies between 𝑎 and 𝑏.

Proof: Let 𝐵 = {𝑥 ∈ [𝑎, 𝑏]: 𝑓(𝑥) < 0} ⇒ 𝐵 ⊂ [𝑎, 𝑏] ⇒ 𝐵 is a bounded set.

Let 𝑐 = sup 𝐵. Since 𝐵 ⊆ [𝑎, 𝑏] ⇒ sup 𝐵 ∈ [𝑎, 𝑏] ⇒ 𝑐 ∈ [𝑎, 𝑏]…….①
Since 𝑓 is continuous on [𝑎, 𝑏] and 𝑐 ∈ [𝑎, 𝑏] ⇒ 𝑓 continuous at 𝑐 ⇒ lim 𝑓 = 𝑓(𝑐)
𝑥→𝑐

Since 𝑐 = sup 𝐵 ⇒ ∃(𝑥𝑛 ) in 𝐵 such that 𝑥𝑛 → 𝑐. But lim 𝑓 = 𝑓(𝑐), so by the Sequential Criterion
𝑥→𝑐
we have: lim 𝑓(𝑥𝑛 ) = 𝑓(𝑐)
𝑛→∞

Since 𝑥𝑛 ∈ 𝐵, ∀𝑛 ∈ ℕ we have 𝑓(𝑥𝑛 ) < 0, ∀𝑛 ∈ ℕ ⇒ lim 𝑓(𝑥𝑛 ) ≤ 0 ⇒ 𝑓(𝑐) ≤ 0…….②

𝑛→∞
 106

Claim: 𝑓(𝑐) = 0
Pf of the Claim: We will prove it by contradiction:
Assume that 𝑓(𝑐) ≠ 0 ⇒ 𝑓(𝑐) < 0 (by ②)
Since 𝑓 continuous at 𝑐 and 𝑓(𝑐) < 0 we have: ∃𝛿 > 0 such that
𝑓(𝑥) < 0, ∀𝑥 ∈ (𝑐 − 𝛿, 𝑐 + 𝛿) ⊂ [𝑎, 𝑏]
𝛿
Since 𝑑 = 𝑐 + 2 ∈ (𝑐 − 𝛿, 𝑐 + 𝛿) ⇒ 𝑓(𝑑) < 0 ⇒ 𝑑 ∈ 𝐵 ⇒ 𝑑 ≤ sup 𝐵 = 𝑐
𝛿 𝛿
⇒ 𝑐 + ≤ 𝑐 ⇒ ≤ 0 ⇒ 𝛿 ≤ 0 contradiction
2 2

∴ Our assumption is false ⇒ 𝑓(𝑐) = 0

Since 𝑓(𝑎)𝑓(𝑏) < 0 ⇒ 𝑓(𝑎) ≠ 0 and 𝑓(𝑏) ≠ 0 ⇒ 𝑐 ≠ 𝑎 and 𝑐 ≠ 𝑏 …….③
⇒ 𝑐 ∈ (𝑎, 𝑏) (by ①&③). So, ∃𝑐 ∈ (𝑎, 𝑏) such that 𝑓(𝑐) = 0

Example 5.3.11: Show that the function 𝑓(𝑥) = 2 ln 𝑥 + √𝑥 − 2 has a zero in [1,2].
Solution:
 𝑓 is continuous on the closed bounded interval [1,2]
 𝑓(1) = −1 < 0 and 𝑓(2) = 2 ln 2 + √2 − 2 > 0 ⇒ 𝑓(1)𝑓(2) < 0
Location of Roots Theorem ⇒ ∃𝑐 ∈ (1,2) such that 𝑓(𝑐) = 0
∴ 𝑓 has a zero in [1,2]

1
Example 5.3.12: Let 𝑓 be continuous on [0,1] such that 𝑓(0) = 𝑓(1). Prove that ∃𝑐 ∈ [0, 2] such
1
that 𝑓(𝑐) = 𝑓 (𝑐 + 2).
Solution:
1 1 1
Case 1: If 𝑓(0) = 𝑓 (2), take 𝑐 = 0 and so ∃𝑐 ∈ [0, 2] such that 𝑓(𝑐) = 𝑓 (𝑐 + 2) which is 𝑐 = 0.
1 1 1
Case 2: If 𝑓(0) ≠ 𝑓 ( ), then let 𝑔(𝑥) = 𝑓(𝑥) − 𝑓 (𝑥 + ) , 𝑥 ∈ [0, ]
2 2 2
1
 𝑔 continuous on the closed bounded interval [0, 2].
1 1 1 1
 𝑔(0) = 𝑓(0) − 𝑓 (2) and 𝑔 (2) = 𝑓 (2) − 𝑓(1) = 𝑓 (2) − 𝑓(0) (since 𝑓(0) = 𝑓(1))
2
1 1 1
⇒ 𝑔(0)𝑔 (2) = − (𝑓(0) − 𝑓 (2)) < 0 (since 𝑓(0) ≠ 𝑓 (2))

1
⇒ 𝑔(0)𝑔 (2) < 0
1
The Location of Roots Theorem: ⇒ ∃𝑐 ∈ [0, 2] such that 𝑔(𝑐) = 0
1 1
⇒ 𝑓(𝑐) − 𝑓 (𝑐 + 2) = 0 ⇒ 𝑓(𝑐) = 𝑓 (𝑐 + 2).
 107

Bolzano’s Intermediate Value Theorem 5.3.13:

Let 𝑓: 𝐼 → ℝ be continuous on the interval 𝐼 and let 𝑎, 𝑏 ∈ 𝐼 such that 𝑓(𝑎) < 𝑓(𝑏), and let
𝑓(𝑎) < 𝑘 < 𝑓(𝑏), then ∃𝑐 between 𝑎 and 𝑏 such that 𝑓(𝑐) = 𝑘.
Proof: Suppose that 𝑎 < 𝑏 and let 𝑓(𝑎) < 𝑘 < 𝑓(𝑏)
Let 𝑔(𝑥) = 𝑓(𝑥) − 𝑘, 𝑥 ∈ [𝑎, 𝑏] ⇒ 𝑔 is continuous on [𝑎, 𝑏] also:
𝑔(𝑎) = 𝑓(𝑎) − 𝑘 < 0 and 𝑔(𝑏) = 𝑓(𝑏) − 𝑘 > 0
The Location of Roots Theorem: ⇒ ∃𝑐 ∈ (𝑎, 𝑏) such that 𝑔(𝑐) = 0
⇒ 𝑓(𝑐) − 𝑘 = 0 ⇒ 𝑓(𝑐) = 𝑘.

Corollary 5.3.14: Let 𝑓 be continuous on the closed bounded interval [𝑎, 𝑏].
(1) If inf 𝑓([𝑎, 𝑏]) ≤ 𝑘 ≤ sup 𝑓([𝑎, 𝑏]), then ∃𝑐 ∈ [𝑎, 𝑏] such that 𝑓(𝑐) = 𝑘.
(2) 𝑓([𝑎, 𝑏]) is a closed bounded interval, that is 𝑓([𝑎, 𝑏]) = [𝑚, 𝑀], where:
𝑚 = inf 𝑓([𝑎, 𝑏]) and 𝑀 = sup 𝑓([𝑎, 𝑏])

Corollary 5.3.15: Let 𝑓 be continuous on an interval 𝐼. Then 𝑓(𝐼) is an interval, where the
intervals 𝐼 and 𝑓(𝐼) need not be closed and bounded.

Exercises 5.3.16 (page 140): 2,4,5,7,14

 108

Section 5.4: Uniform Continuity

In this section we will study a new concept of continuity of functions on sets.

Remark 5.4.1: Recall that for a function 𝑓: 𝐴 → ℝ, the following are equivalent:
(1) 𝑓 is continuous on 𝐴, that is 𝑓 is continuous at every point 𝑦 ∈ 𝐴.
(2) Given 𝜀 > 0, ∃𝛿 > 0 such that 𝑥 ∈ 𝐴 with |𝑥 − 𝑦| < 𝛿 ⇒ |𝑓(𝑥) − 𝑓(𝑦)| < 𝜀

1
Remark 5.4.2: Let 𝑓(𝑥) = 𝑥 , 𝑥 > 0. Let 𝜀 > 0 be fixed:
1
 Taking 𝑦 = 2 ⇒ 𝑓(𝑦) = 𝑓(2) = 2:
⇒ ∃𝛿1 > 0 satisfying the definition
1 1
 Taking 𝑦 = 2 ⇒ 𝑓(𝑦) = 𝑓(2) = 2:
⇒ ∃𝛿2 > 0 satisfying the definition
 Observe that for a fixed 𝜀 > 0, changing the point 𝑦 implies that 𝛿 may be changes.

Definition 5.4.3:
A function 𝑓: 𝐴 → ℝ is said to be uniformly continuous on 𝑨 if ∀𝜀 > 0, ∃𝛿 > 0 such that
∀𝑥, 𝑦 ∈ 𝐴 with |𝑥 − 𝑦| < 𝛿 ⇒ |𝑓(𝑥) − 𝑓(𝑦)| < 𝜀.

Part 1: Geometric Illustrating of Uniform Continuity

‫ فإن هذا اإلقتران يكون اتصاله منتظما ً على المجموعة 𝐴 إذا لم يحدث الوضع‬،𝐴 ‫ ليكن لدينا اقتران 𝑓 متصالً على مجموعة‬:‫مالحظة‬
𝑓 ‫ فإذا حدث هذا الوضع فإن اإلقتران‬.‫ ال يوجد مماسات لالقتران 𝑓 تقترب من الوضع العامودي بشكل مستمر دون توقف‬:‫التالي‬
ً ‫يكون اتصاله على المجموعة 𝐴 غير منتظم وفي هذه الحالة نبحث عن مجموعة جزئية داخل 𝐴 والتي يكون اإلتصال عليها منتظما‬
‫(أي ال يحدث الوضع السابق فيها) ولتحقيق ذلك علينا أن نضع شرطا ً على المجموعة 𝐴 نمنع فيها اإلقتراب من الوضع العامودي‬
.𝑓 ‫لمماسات المنحنى‬
 109

Example 5.4.4: Find the largest set on which the

function 𝑓 given in the figure is uniformly continuous.

Solution: From the graph of 𝑓 we see that 𝑓 is

continuous on ℝ\{−1,1}. ⇒ 𝑓 is not uniformly
continuous on ℝ\{−1,1}
So, prevent 𝑥 → −1+ and 𝑥 → 1− :
⇒ 𝑓 is uniformly continuous on:
(−∞, −1] ∪ [−𝑎, 𝑎] ∪ (1, ∞), ∀0 < 𝑎 < 1

 The function 𝑓 is uniformly continuous on each of the following sets:

(−∞, −2], (−0.5,0.7], (2, ∞), (−∞, −1), (−∞, −2) ∪ (1,3]
 The function 𝑓 is not uniformly continuous on each of the following sets:
(−1,0.5), [0,1), (−1,0] ∪ [2,3]

Example 5.4.5: Find the largest set on which the function is uniformly continuous:
1
(1) 𝑓(𝑥) = 𝑥 (2) 𝑓(𝑥) = 𝑥 2 (3) 𝑓(𝑥) = 𝑒 𝑥
𝑥 3 −𝑥
(4) 𝑓(𝑥) = 𝑥 3 +𝑥 (5) 𝑓(𝑥) = √𝑥 (6) 𝑓(𝑥) = 𝑥
1
(7) 𝑓(𝑥) = ln 𝑥 (8) 𝑓(𝑥) = sin (𝑥)
Solution:
(1) 𝑓 is continuous on ℝ\{0} ⇒ 𝑓 is not uniformly continuous on ℝ\{0}
because approaching 0 leads to a vertical tangent line
So, prevent 𝑥 → 0− and 𝑥 → 0+ : Let 𝑎 > 0. Then
𝑓 is uniformly continuous on (−∞, −𝑎] ∪ [𝑎, ∞),∀𝑎 > 0.

1
 The function 𝑓(𝑥) = 𝑥 is uniformly continuous on each of the following sets:
(−∞, −5], (−3, −1), (0.1,2], (1, ∞), and (−5, −2] ∪ [−1, −0.1) ∪ (1,7)
1
 The function 𝑓(𝑥) = 𝑥 is not uniformly continuous on each of the following sets:
(−∞, 0), (0,1), (0,2], (0, ∞), and (−5, −2] ∪ (0,4] ∪ (1, ∞)

(2) 𝑓(𝑥) = 𝑥 2 is continuous on ℝ ⇒ 𝑓 is not uniformly continuous on ℝ

because approaching ∞ and −∞ leads to vertical tangent lines
So, prevent 𝑥 → −∞ and prevent 𝑥 → ∞.
Let 𝑎 > 0. Then 𝑓 is uniformly continuous on [−𝑎, 𝑎], ∀𝑎 > 0.
 110

 The function 𝑓(𝑥) = 𝑥 2 is uniformly continuous on each of the following sets:

(−7, −5], (−3, −1), (0.1,2], [4,6], (1,5), and (−5, −2] ∪ [−1, −0.1) ∪ (1,7).
 The function 𝑓(𝑥) = 𝑥 2 is not uniformly continuous on each of the following sets:
(−∞, −5], (0, ∞), (−∞, 1) ∪ [4,6]

(3) 𝑓 is continuous on ℝ ⇒ 𝑓 is not uniformly continuous on ℝ

because approaching ∞ leads to a vertical tangent line
So, prevent 𝑥 → ∞: Let 𝑎 > 0. Then 𝑓 is uniformly continuous
on (−∞, 𝑎], ∀𝑎 > 0 ∈ ℝ

 The function 𝑓(𝑥) = 𝑒 𝑥 is uniformly continuous on each of the following sets:

(−∞, −5], (−∞, 7), (0.1,2], [4,6], (1,5), and (−5, −2] ∪ [−1, −0.1) ∪ (1,7).
 The function 𝑓(𝑥) = 𝑒 𝑥 is not uniformly continuous on each of the following sets: (4, ∞),
[1, ∞), and (−5, −2] ∪ [−1, −0.1) ∪ (1, ∞).

𝑥 3 −𝑥
(4) 𝑓(𝑥) = 𝑥 3 +𝑥 is continuous on ℝ\{−1,0}.
⇒ 𝑓 is not uniformly continuous on ℝ\{−1,0} because approaching −1 (only) leads to a vertical
tangent line. So, prevent 𝑥 → −1 (from right and from left: 𝑥 ↛ −1− and 𝑥 ↛ −1+ )
Let 𝑎 > 1, 0 < 𝑏 < 1. Then 𝑓 is uniformly continuous on (−∞, −𝑎] ∪ [−𝑏, 𝑏] ∪ [𝑎, ∞).
(5) 𝑓(𝑥) = √𝑥 is continuous on [0, ∞)
⇒ 𝑓 is uniformly continuous on [0, ∞).

(6) 𝑓(𝑥) = 𝑥 is continuous on ℝ.

⇒ 𝑓 is uniformly continuous on ℝ = (−∞, ∞).
 111

(7) 𝑓(𝑥) = ln 𝑥 is continuous on (0, ∞).

𝑓 is not uniformly continuous on (0, ∞) (why?)
Let 𝑎 > 0. The 𝑓 is uniformly continuous on [𝑎, ∞).

1
(8) 𝑓(𝑥) = sin (𝑥) is continuous on ℝ\{0}.
Its graph is given in the figure.
⇒ 𝑓 is not uniformly continuous on ℝ\{0}
Let 𝑎 > 0. Then 𝑓 is uniformly continuous on
(−∞, −𝑎] ∪ [𝑎, ∞)

Part 2: Uniform Continuity by Using Definition

Remark 5.4.6: Recall that a function 𝑓: 𝐴 → ℝ is said to be uniformly continuous on A iff ∀𝜀 > 0,
∃𝛿 > 0 such that ∀𝑥, 𝑦 ∈ 𝐴 with |𝑥 − 𝑦| < 𝛿 ⇒ |𝑓(𝑥) − 𝑓(𝑦)| < 𝜀.

Example 5.4.7: Using definition, show that 𝑓(𝑥) = 𝑥 4 is uniformly continuous on [−5,4].
Solution: Let 𝜀 > 0 be given and let 𝑥, 𝑦 ∈ [−5,4].
|𝑓(𝑥) − 𝑓(𝑦)| = |𝑥 4 − 𝑦 4 |
= |(𝑥 2 + 𝑦 2 )(𝑥 2 − 𝑦 2 )|
= |(𝑥 2 + 𝑦 2 )(𝑥 + 𝑦)(𝑥 − 𝑦)|
≤ |(𝑥 2 + 𝑦 2 )(𝑥 + 𝑦)||𝑥 − 𝑦|
≤ ((|𝑥 2 | + |𝑦 2 |)(|𝑥| + |𝑦|)|𝑥 − 𝑦|
= ((𝑥 2 + 𝑦 2 )(|𝑥| + |𝑦|)|𝑥 − 𝑦|
≤ ((25 + 25)(5 + 5))|𝑥 − 𝑦|
= 500|𝑥 − 𝑦|
⇒ |𝑓(𝑥) − 𝑓(𝑦)| ≤ 500|𝑥 − 𝑦|.
𝜀
Now, let 500|𝑥 − 𝑦| < 𝜀 ⇒ |𝑥 − 𝑦| < 500
𝜀
Take 0 < 𝛿 ≤ 500. So, ∀𝑥, 𝑦 ∈ [−5,4] with |𝑥 − 𝑦| < 𝛿 ⇒ |𝑓(𝑥) − 𝑓(𝑦)| < 𝜀
⇒ 𝑓(𝑥) = 𝑥 4 is uniformly continuous on [−5,4]

Remark 5.4.8: If 𝑥, 𝑦 ∈ ℝ, then |√𝑥 − √𝑦| ≤ √|𝑥 − 𝑦|.

 112

Example 5.4.9: Using definition, show that 𝑓(𝑥) = √𝑥 is uniformly continuous on [0, ∞).
Solution: Let 𝜀 > 0 be given and let 𝑥, 𝑦 ∈ [0, ∞).
|𝑓(𝑥) − 𝑓(𝑦)| = |√𝑥 − √𝑦|

≤ √|𝑥 − 𝑦| (by Remark 5.4.8)

Now, let √|𝑥 − 𝑦| < 𝜀 ⇒ |𝑥 − 𝑦| < 𝜀 2

Take 0 < 𝛿 ≤ 𝜀 2 . So, ∀𝑥, 𝑦 ∈ ℝ with |𝑥 − 𝑦| < 𝛿 ⇒ |𝑓(𝑥) − 𝑓(𝑦)| < 𝜀
⇒ 𝑓(𝑥) = √𝑥 is uniformly continuous on [0, ∞).

To simplify dealing with uniform continuity we start with the following Theorems:
Uniform Continuity Theorem 5.4.10: If 𝑓: [𝑎, 𝑏] → ℝ is continuous on the closed bounded
interval [𝑎, 𝑏], then 𝑓 is uniformly continuous on [𝑎, 𝑏].

Remark 5.4.11: Recall the following:

(1) From the definition we have: If 𝑓 is uniformly continuous on 𝐴, then 𝑓 is continuous on 𝐴.
(2) If 𝑓 is uniformly continuous on sets 𝐴 and 𝐵, then 𝑓 is uniformly continuous on 𝐴 ∪ 𝐵.
(3) If 𝑓 is uniformly continuous on 𝐴 and we have 𝐵 ⊆ 𝐴, then 𝑓 is uniformly continuous on 𝐵

Example 5.4.12:
(1)
1
 𝑓(𝑥) = 𝑥 is uniformly continuous on [3,10] (since 𝑓 is continuous on the closed
bounded interval [3,10]).
 𝑓 is uniformly continuous on (3,7] ∪ (8,9) since 𝑓 is uniformly continuous on [3,10]
and (3,7] ∪ (8,9) ⊆ [3,10].
(2) If 𝑃(𝑥) is a polynomial, then 𝑃(𝑥) is uniformly continuous on any closed bounded interval
[𝑎, 𝑏] (since 𝑃(𝑥) is continuous on [𝑎, 𝑏] and [𝑎, 𝑏] is a closed bounded interval).
 For examples: 𝑓(𝑥) = 𝑥 3 − 𝜋𝑥 + 1 and 𝑔(𝑥) = −0.1𝑥 5 − √2𝑥 4 are uniformly
continuous on the closed bounded intervals [−5,0] and [1,2.5].
(3) 𝑓(𝑥) = sin 𝑥 and 𝑔(𝑥) = cos 𝑥 are uniformly continuous on any closed bounded interval
[𝑎, 𝑏] (because they are continuous on [𝑎, 𝑏] and [𝑎, 𝑏] is a closed bounded interval).

Nonuniform Continuity Criterion 5.4.13:

Let 𝑓: 𝐴 → ℝ be a function. Then 𝑓 is not uniformly continuous on 𝐴 iff ∃𝜀0 > 0 and ∃(𝑥𝑛 ), (𝑦𝑛 )
in 𝐴 such that lim (𝑥𝑛 − 𝑦𝑛 ) = 0 and |𝑓(𝑥𝑛 ) − 𝑓(𝑦𝑛 )| ≥ 𝜀0 , ∀𝑛 ∈ ℕ
𝑛→∞

Example 5.4.14: Using the Nonuniform Continuity Criterion, show that 𝑓(𝑥) = 𝑥 2 is not uniformly
continuous on ℝ.
Proof:
1
 Let 𝑥𝑛 = 𝑛 + 𝑛 and 𝑦𝑛 = 𝑛, 𝑛 ∈ ℕ ⇒ (𝑥𝑛 ), (𝑦𝑛 ) are sequences in ℝ.
 113

1 1
 lim (𝑥𝑛 − 𝑦𝑛 ) = lim (𝑛 + 𝑛 − 𝑛) = lim =0
𝑛→∞ 𝑛→∞ 𝑛→∞ 𝑛
1 2 1 1 1
 |𝑓(𝑥𝑛 ) − 𝑓(𝑦𝑛 )| = |(𝑛 + 𝑛) − 𝑛2 | = |𝑛2 + 2 + 𝑛2 − 𝑛2 | = |2 + 𝑛2 | = 2 + 𝑛2 ≥ 2, ∀𝑛 ∈ ℕ

Take 𝜀0 = 2. The Nonuniform Continuity criterion ⇒ 𝑓 is not uniformly continuous on ℝ.

1
Example 5.4.15: Using the Nonuniform Continuity Criterion, show that 𝑓(𝑥) = sin (𝑥) is not uniformly
continuous on (0, ∞).

Solution:
2 1
 Let 𝑥𝑛 = (4𝑛+1)𝜋 and 𝑦𝑛 = 𝑛𝜋 ⇒ (𝑥𝑛 ), (𝑦𝑛 ) are in (0, ∞).
2 1
 lim (𝑥𝑛 − 𝑦𝑛 ) = lim ((4𝑛+1)𝜋 − 𝑛𝜋) = 0
𝑛→∞ 𝑛→∞
1 1 (4𝑛+1)𝜋
 |𝑓(𝑥𝑛 ) − 𝑓(𝑦𝑛 )| = |sin ( 2 ) − sin ( 1 ) | = |sin ( ) − sin(𝑛𝜋) | = 1, ∀𝑛 ∈ ℕ
2
(4𝑛+1)𝜋 𝑛𝜋

Take 𝜀0 = 1. The nonuniform Continuity Criterion holds ⇒ 𝑓 is not uniformly continuous on (0, ∞).

Part 3: Uniform Continuity and Lipschitz Functions

Definition 5.4.16: A function 𝑓: 𝐴 → ℝ is called Lipschitz (‫ )ليبشيتز‬on 𝐴 if ∃𝐾 > 0 such that

|𝑓(𝑥) − 𝑓(𝑦)| ≤ 𝐾|𝑥 − 𝑦|, ∀𝑥, 𝑦 ∈ 𝐴. The number 𝐾 is called the Lipschitz
constant.

Theorem 5.4.17: If the function 𝑓: 𝐴 → ℝ is Lipschitz on 𝐴, then 𝑓 is uniformly continuous on 𝐴.

Proof: Let 𝜀 > 0 be given.

𝑓 is Lipschitz ⇒ ∃𝐾 > 0 such that |𝑓(𝑥) − 𝑓(𝑦)| ≤ 𝐾|𝑥 − 𝑦|, ∀𝑥, 𝑦 ∈ 𝐴
𝜀
⇒ |𝑓(𝑥) − 𝑓(𝑦)| ≤ 𝐾|𝑥 − 𝑦| < 𝜀, ∀𝑥, 𝑦 ∈ 𝐴 ⇒ |𝑥 − 𝑦| < 𝐾
𝜀
Take 0 < 𝛿 ≤ 𝐾. So, 𝑥, 𝑦 ∈ 𝐴 with |𝑥 − 𝑦| < 𝛿 ⇒ |𝑓(𝑥) − 𝑓(𝑦)| < 𝜀
⇒ 𝑓 is uniformly continuous on 𝐴.

Remark 5.4.18:
(1) The Mean-Value Theorem says that: If 𝑓: [𝑎, 𝑏] → ℝ is a function continuous on the closed
bounded interval [𝑎, 𝑏] and differentiable on (𝑎, 𝑏), then ∃𝑐 ∈ (𝑎, 𝑏) such that
𝑓(𝑏)−𝑓(𝑎)
𝑓 ′ (𝑐) = . Equivalently, 𝑓(𝑏) − 𝑓(𝑎) = 𝑓 ′ (𝑐)(𝑏 − 𝑎)
𝑏−𝑎

(2) A function 𝑔: 𝐴 → ℝ is bounded on 𝐴 if ∃𝑀 > 0 such that |𝑓(𝑥)| ≤ 𝑀, ∀𝑥 ∈ 𝐴

 114

Theorem 5.4.19: Let 𝐼 be an interval in ℝ and let 𝑓: 𝐼 → ℝ be differentiable on 𝐼. Then

(1) 𝑓 is Lipschitz on 𝐼 iff 𝑓 ′ is bounded on 𝐼.
(2) If 𝑓 ′ is bounded on 𝐼, then 𝑓 is uniformly continuous on 𝐼.

1
Example 5.4.20: Show that the function 𝑓(𝑥) = 1+𝑥 2 , 𝑥 ∈ ℝ is uniformly continuous on ℝ. (Hint:
Show that 𝑓 is Lipschitz on ℝ).
Solution: We want to find 𝑓 ′ and to show it is bounded on ℝ:
2𝑥 |𝑥| 1 |𝑥| |𝑥|
|𝑓 ′ (𝑥)| = |(1+𝑥 2 )2 | = 2 (1+𝑥 2 )2 = 2 ( 2 ) ( 2 ) ≤ 2(1) ( 2 ) ≤ 2, ∀𝑥 ∈ ℝ
1+𝑥 1+𝑥 1+𝑥

⇒ |𝑓 ′ (𝑥)| ≤ 2, ∀𝑥 ∈ ℝ ⇒ 𝑓 ′ is bounded on ℝ ⇒ 𝑓 is Lipschitz on ℝ

⇒ 𝑓 is uniformly continuous on ℝ.

Exercise 5.4.21: Which of the following is a Lipchitz function on its domain:

(1) 𝑓(𝑥) = 𝑥 2 (2) 𝑓(𝑥) = 𝑒 𝑥 (3) 𝑓(𝑥) = ln(𝑥)
1 2)
(4) 𝑓(𝑥) = sin (𝑥) (5) 𝑓(𝑥) = ln(1 + 𝑥 (6) 𝑓(𝑥) = 3

Part 4: Uniform Continuity using Continuity and limits

Theorem 5.4.22: Let 𝐴 ⊆ ℝ and let 𝑐1 , … , 𝑐𝑛 ∈ 𝐴. If 𝑓 is continuous on 𝐴 except at the points

𝑐1 , … , 𝑐𝑛 such that lim 𝑓 = 𝐿𝑗 ∈ ℝ for 𝑗 = 1, … , 𝑛. Then the function
𝑥→𝑐𝑗
𝑓(𝑥) , 𝑥 ∈ 𝐴 − {𝑐1 , 𝑐2 , … }
𝐿 , 𝑥 = 𝑐1
𝐹(𝑥) = { 1 }
⋮ ⋮
𝐿𝑛 , 𝑥 = 𝑐2
is continuous on 𝐴. The function 𝐹 is called the continuous extension of 𝒇.

Example 5.4.23:
sin 𝑥 sin 𝑥
(1) The function 𝑓(𝑥) = 𝑥 is continuous on ℝ\{0} and lim 𝑓 = lim 𝑥 = 1
𝑥→0 𝑥→0
𝑓(𝑥) , 𝑥 ∈ ℝ − {0}
⇒ 𝐹(𝑥) = { }
1 ,𝑥 = 0
sin 𝑥
, 𝑥 ∈ ℝ − {0}
𝐹(𝑥) = { 𝑥 } is the continuous extension of 𝑓.
1 ,𝑥 = 0
𝑥
(2) The function 𝑓(𝑥) = 𝑥 2 −𝑥 is continuous on ℝ\{0,1}:
𝑥 𝑥 1
 lim 𝑓 = lim = lim = lim doesn’t exist
𝑥→1 𝑥→1 𝑥 2 −𝑥 𝑥→1 𝑥(𝑥−1) 𝑥→1 𝑥−1
⇒ The function 𝑓 has no continuous extension.
 115

Continuous Extension Theorem 5.4.24: The function 𝑓: (𝑎, 𝑏) → ℝ is uniformly continuous on

the bounded interval (𝑎, 𝑏) iff the function 𝑓 has a continuous extension function on the closed
bounded interval [𝒂, 𝒃].

Remark 5.4.25: Let (𝑎, 𝑏) be a bounded interval. Then:

A function 𝑓 is uniformly continuous on (𝑎, 𝑏) iff 𝑓 has a continuous extension function on [𝑎, 𝑏]
iff each of the following hold:
(1) 𝑓 is continuous on (𝑎, 𝑏)
(2) lim+ 𝑓 exist and lim− 𝑓 exist
𝑥→𝑎 𝑥→𝑏

Example 5.4.26:
sin 𝑥
(a) Show that 𝑓(𝑥) = 𝑥 is uniformly continuous on (0,1].
1
(b) Show that 𝑓(𝑥) = 𝑥 is uniformly continuous on (0,1).
Solution:
(a)
(1) 𝑓 continuous on the bounded interval (0,1]
sin 𝑥 sin 𝑥
(2) lim+ = 1 exists and lim− = sin 1 exists
𝑥→0 𝑥 𝑥→1 𝑥
sin 𝑥
,0 < 𝑥 ≤ 1
⇒ 𝑓 has the continuous extension on [0,1] which is 𝐹(𝑥) = { 𝑥 }
1 ,𝑥 = 0
So, 𝑓 is uniformly continuous on (0,1].
1
(b) lim+ 𝑥 does not exist ⇒ 𝑓 hasn’t a continuous extension on [0,1].
𝑥→0
⇒ 𝑓 is not uniformly continuous on (0,1).

Part 5: Theory of Uniformly Continuous Functions

Theorem 5.4.27: Let 𝑓 and 𝑔 be uniformly continuous on 𝐴. Then 𝑓 + 𝑔, 𝑓 − 𝑔, and 𝑘𝑓 are

uniformly continuous on 𝐴, where 𝑘 ∈ ℝ is a constant.
Proof: We will prove 𝑓 + 𝑔 is uniformly continuous on 𝐴. The proof of the other functions is
similar. Let 𝜀 > 0 be given.
𝑓 and 𝑔 be uniformly continuous on 𝐴 ⇒ ∃𝛿 > 0 such that
𝜀 𝜀
𝑥, 𝑦 ∈ 𝐴 and |𝑥 − 𝑦| < 𝛿 ⇒ |𝑓(𝑥) − 𝑓(𝑦)| < 2 and |𝑔(𝑥) − 𝑔(𝑦)| < 2…………①
⇒ |(𝑓 + 𝑔)(𝑥) − (𝑓 + 𝑔)(𝑦)| = |(𝑓(𝑥) − 𝑓(𝑦)) + (𝑔(𝑥) − 𝑔(𝑦))|
≤ |𝑓(𝑥) − 𝑓(𝑦)| + |(𝑔(𝑥) − 𝑔(𝑦)|
𝜀 𝜀
< 2 + 2 (by ①)
=𝜀
So, 𝑥, 𝑦 ∈ 𝐴 and |𝑥 − 𝑦| < 𝛿 ⇒ |(𝑓 + 𝑔)(𝑥) − (𝑓 + 𝑔)(𝑦)| < 𝜀
⇒ 𝑓 + 𝑔 is uniformly continuous on 𝐴.

Example 5.4.28: Give examples of two functions 𝑓 and 𝑔 that are uniformly continuous on a set 𝐴
but 𝑓𝑔 is not uniformly continuous on 𝐴.
Solution: Let 𝑓(𝑥) = 𝑔(𝑥) = 𝑥 ⇒ 𝑓 and 𝑔 are uniformly continuous on ℝ
But (𝑓𝑔)(𝑥) = 𝑥 2 is not uniformly continuous on ℝ.
 116

Theorem 5.4.29: Let 𝑓 and 𝑔 be uniformly continuous on 𝐴. If 𝑓 and 𝑔 are bounded on 𝐴, then 𝑓𝑔
is uniformly continuous on 𝐴.
Proof: Let 𝜀 > 0 be given. 𝑓 and 𝑔 are bounded on 𝐴 ⇒ ∃𝑀 > 0 such that
|𝑓(𝑥)| ≤ 𝑀, |𝑔(𝑥)| ≤ 𝑀, ∀𝑥 ∈ 𝐴………..①
𝑓 and 𝑔 be uniformly continuous on 𝐴 ⇒ ∃𝛿 > 0 such that
𝜀 𝜀
𝑥, 𝑦 ∈ 𝐴 and |𝑥 − 𝑦| < 𝛿 ⇒ |𝑓(𝑥) − 𝑓(𝑦)| < 2𝑀 and |𝑔(𝑥) − 𝑔(𝑦)| < 2𝑀…………②
|(𝑓𝑔)(𝑥) − (𝑓𝑔)(𝑦)| = |𝑓(𝑥)𝑔(𝑥) − 𝑓(𝑥)𝑔(𝑦) + 𝑓(𝑥)𝑔(𝑦) − 𝑓(𝑦)𝑔(𝑦)|
≤ |𝑓(𝑥)𝑔(𝑥) − 𝑓(𝑥)𝑔(𝑦)| + |𝑓(𝑥)𝑔(𝑦) − 𝑓(𝑦)𝑔(𝑦)|
= |𝑓(𝑥)| |𝑔(𝑥) − 𝑔(𝑦)| + |𝑔(𝑦)| |𝑓(𝑥) − 𝑓(𝑦)|
≤ 𝑀 |𝑔(𝑥) − 𝑔(𝑦)| + 𝑀 |𝑓(𝑥) − 𝑓(𝑦)| (by①)
𝜀 𝜀
≤ 𝑀 2𝑀 + 𝑀 2𝑀 (by ②)
=𝜀
𝑥, 𝑦 ∈ 𝐴 and |𝑥 − 𝑦| < 𝛿 ⇒ |(𝑓𝑔)(𝑥) − (𝑓𝑔)(𝑦)| < 𝜀
⇒ 𝑓𝑔 is uniformly continuous on 𝐴

Theorem 5.4.30: Let 𝑓: 𝐴 → ℝ and 𝑔: 𝐵 → ℝ such that 𝑔(𝐵) ⊆ 𝐴. If 𝑓 is uniformly continuous on

𝐴 and 𝑓 is uniformly continuous on 𝐵, then 𝑓 ∘ 𝑔 is uniformly continuous on 𝐵.

Example 5.4.31: Show that 𝑓(𝑥) = |√𝑥 − 1| is uniformly continuous on [0, ∞).
Solution: Let 𝑔(𝑥) = √𝑥 and ℎ(𝑥) = |𝑥 − 1|. Then 𝑔 is uniformly continuous on [0, ∞) and ℎ is
uniformly continuous on ℝ ⇒ ℎ ∘ 𝑔 is uniformly continuous on [0, ∞)
But (𝑔 ∘ ℎ)(𝑥) = 𝑔(ℎ(𝑥)) = |ℎ(𝑥) − 1| = |√𝑥 − 1| = 𝑓(𝑥)
∴ 𝑓 is uniformly continuous on [0, ∞).

Exercise 5.4.32: Show that 𝑓(𝑥) = √|𝑥| − 1 is uniformly continuous on ℝ\(−1,1).

 117

Section 5.6: Monotone and Inverse Functions

Definition 5.6.1: Let 𝐴 ⊆ ℝ and 𝑓 ∶ 𝐴 → ℝ be a function. The function 𝑓 is said to be:

(1) increasing on 𝐴 if whenever 𝑥1 , 𝑥2 ∈ 𝐴 and 𝑥1 < 𝑥2 , then 𝑓(𝑥1 ) ≤ 𝑓(𝑥2 )
(2) strictly increasing on 𝐴 if whenever 𝑥1 , 𝑥2 ∈ 𝐴 and 𝑥1 < 𝑥2 , then 𝑓(𝑥1 ) < 𝑓(𝑥2 )
(3) decreasing on 𝐴 if whenever 𝑥1 , 𝑥2 ∈ 𝐴 and 𝑥1 < 𝑥2 , then 𝑓(𝑥1 ) ≥ 𝑓(𝑥2 )
(4) strictly decreasing on 𝐴 if whenever 𝑥1 , 𝑥2 ∈ 𝐴 and 𝑥1 < 𝑥2 , then 𝑓(𝑥1 ) > 𝑓(𝑥2 )
(5) monotone on 𝐴 if it is either increasing or decreasing on 𝐴.
(6) strictly monotone on 𝐴 if it is either strictly increasing or strictly decreasing on 𝐴.
Theorem 5.6.2: Let 𝑓 ∶ [𝑎, 𝑏] → ℝ be conts on [𝑎, 𝑏] and differentiable on (𝑎, 𝑏). Then
(1) 𝑓 is increasing on [𝑎, 𝑏] iff 𝑓 ′ (𝑥) ≥ 0, ∀𝑥 ∈ (𝑎, 𝑏)
(2) 𝑓 is strictly increasing on [𝑎, 𝑏] iff 𝑓 ′ (𝑥) > 0, ∀𝑥 ∈ (𝑎, 𝑏)
(3) 𝑓 is decreasing on [𝑎, 𝑏] iff 𝑓 ′ (𝑥) ≤ 0, ∀𝑥 ∈ (𝑎, 𝑏)
(4) 𝑓 is strictly decreasing on [𝑎, 𝑏] iff 𝑓 ′ (𝑥) < 0, ∀𝑥 ∈ (𝑎, 𝑏)
(5) 𝑓 is constant on [𝑎, 𝑏] iff 𝑓 ′ (𝑥) = 0, ∀𝑥 ∈ (𝑎, 𝑏)

Remark 5.6.3: Let 𝐴 ⊆ ℝ and 𝑓 ∶ 𝐴 → ℝ be a function.

(1) 𝑓 is increasing on 𝐴 iff −𝑓 is decreasing on 𝐴.
(2) 𝑓 is decreasing on 𝐴 iff −𝑓 is increasing on 𝐴.
(3) 𝑓 is strictly increasing on 𝐴 iff −𝑓 is strictly decreasing on 𝐴.
(4) 𝑓 is strictly decreasing on 𝐴 iff −𝑓 is strictly increasing on 𝐴.

Theorem 5.6.4: Let 𝐼 ⊆ ℝ be an interval and let 𝑓 ∶ 𝐼 → ℝ be increasing on 𝐼. Suppose that 𝑐 ∈ 𝐼

is not an endpoint of 𝐼. Then
(1) lim− 𝑓 = sup{𝑓(𝑥): 𝑥 ∈ 𝐼, 𝑥 < 𝑐}
𝑥→𝑐
(2) lim+ 𝑓 = inf{𝑓(𝑥): 𝑥 ∈ 𝐼, 𝑥 > 𝑐}
𝑥→𝑐
Theorem 5.6.5: Let 𝐼 ⊆ ℝ be an interval and let 𝑓 ∶ 𝐼 → ℝ be increasing on 𝐼. Suppose that 𝑐 ∈ 𝐼
is not an endpoint of 𝐼. Then the following are equivalent:
(1) 𝑓 is conts at 𝑐
(2) lim+ = lim− 𝑓 = 𝑓(𝑐)
𝑥→𝑐 𝑥→𝑐
(3) sup{𝑓(𝑥): 𝑥 ∈ 𝐼, 𝑥 < 𝑐} = inf{𝑓(𝑥): 𝑥 ∈ 𝐼, 𝑥 > 𝑐} = 𝑓(𝑐)

Theorem 5.6.6: Let 𝐼 ⊆ ℝ be an interval and let 𝑓 ∶ 𝐼 → ℝ be monotone on 𝐼. Then the set of
discontinuities of 𝑓 on 𝐼 is a countable set.
 118

Example 5.6.7: True or false: There a function 𝑓 ∶ [0,1) → ℝ decreasing on [0,1) and
discontinuous on [0,1) ∩ ℚ𝑐 . Justify.
Solution: The answer is No. Because assume that there is a function 𝑓 on [0,1) discontinuous on
[0,1) ∩ ℚ𝑐 . Theorem 6 ⇒ [0,1) ∩ ℚ𝑐 is a countable set which is a contradiction since [0,1) ∩ ℚ𝑐
is uncountable ⇒ our assumption is false
⇒ there is no such function.

Remark 5.6.8:
(1) A function 𝑓∶𝐴→ℝ is injective (one-to-one) iff whenever
𝑥1 , 𝑥2 ∈ 𝐴 with 𝑥1 ≠ 𝑥2 we have 𝑓(𝑥1 ) ≠ 𝑓(𝑥2 ). Equivalently, 𝑓 is injective iff whenever
𝑥1 , 𝑥2 ∈ 𝐴 with 𝑓(𝑥1 ) = 𝑓(𝑥2 ) we have 𝑥1 = 𝑥2 .
(2) If a function 𝑓 ∶ 𝐴 → ℝ is injective, then its inverse function 𝑓 −1 exists.
(3) If a function 𝑓 ∶ 𝐴 → ℝ is strictly monotone, then 𝑓 is injective and so its inverse function 𝑓 −1 exists.

Continuous Function Theorem 5.6.9: Let 𝐼 ⊆ ℝ be an interval and let 𝑓 ∶ 𝐼 → ℝ be strictly

monotone and conts on 𝐼. Then the inverse function 𝑓 −1 : 𝑓(𝐼) → ℝ is strictly monotone and continuous
on 𝑓(𝐼).
𝑥 , 𝑥 rational
Example 5.6.10: Let 𝑓 ∶ ℝ → ℝ be defined by 𝑓(𝑥) = { }.
1−𝑥 , 𝑥 irrational
(1) Show that 𝑓 is injective on ℝ.
(2) Find the formula 𝑓 −1 (𝑥), if it exists.
Solution: First we have to observe the following for the given function 𝑓:
 𝑥 is rational iff 𝑓(𝑥) = 𝑥 iff 𝑓(𝑥) is rational
 𝑥 is irrational iff 𝑓(𝑥) = 1 − 𝑥 iff 𝑓(𝑥) is irrational
 𝑓(𝑥1 ) = 𝑓(𝑥2 ) ⇒ both 𝑓(𝑥1 ), 𝑓(𝑥2 ) are rational or both 𝑓(𝑥1 ), 𝑓(𝑥2 ) are irrational

(1) Let 𝑥1 , 𝑥2 ∈ [0,1] with 𝑓(𝑥1 ) = 𝑓(𝑥2 ). We have 2 cases:

Case 1: both 𝑓(𝑥1 ), 𝑓(𝑥2 ) are rational: ⇒ 𝑓(𝑥1 ) = 𝑥1 and 𝑓(𝑥2 ) = 𝑥2
But 𝑓(𝑥1 ) = 𝑓(𝑥2 ) ⇒ 𝑥1 = 𝑥2

Case 2: both 𝑓(𝑥1 ), 𝑓(𝑥2 ) are irrational: ⇒ 𝑓(𝑥1 ) = 1 − 𝑥1 and 𝑓(𝑥2 ) = 1 − 𝑥2

But 𝑓(𝑥1 ) = 𝑓(𝑥2 ) ⇒ 1 − 𝑥1 = 1 − 𝑥2 ⇒ 𝑥1 = 𝑥2
∴ In either case we have 𝑓(𝑥1 ) = 𝑓(𝑥2 ) ⇒ 𝑥1 = 𝑥2 ⇒ 𝑓 is injective
(2) Part (1) ⇒ 𝑓 is injective ⇒ 𝑓 −1 exists. To find 𝑓 −1 : 𝑓(ℝ) → ℝ
First observe that 𝑓(ℝ) = Range(𝑓) = ℝ and 𝑓(𝑓 −1 (𝑥)) = 𝑥, ∀𝑥 ∈ ℝ. So, we have to cases:

Case 1: 𝑓 −1 (𝑥) is rational ⇒ 𝑓 (𝑓

⏟−1 (𝑥)) = 𝑓 −1 (𝑥)
rational
−1 (𝑥)) −1 (𝑥)
But 𝑓(𝑓 =𝑥⇒ 𝑓 = 𝑥 ⇒ 𝑥 rational

Case 2: 𝑓 −1 (𝑥) is irrational ⇒ 𝑓 ( 𝑓

⏟−1 (𝑥) ) = 1 − 𝑓 −1 (𝑥)
irrational
But 𝑓(𝑓 −1 (𝑥)) = 𝑥 ⇒ 1 − 𝑓 −1 (𝑥) = 𝑥
⇒ 𝑓 −1 (𝑥) = 1 − 𝑥 ⇒ 𝑥 irrational
𝑥 , 𝑥 rational
From Cases 1&2: ⇒ 𝑓 −1 (𝑥) = { }
1 − 𝑥 , 𝑥 irrational
∴ 𝑓 −1 (𝑥) = 𝑓(𝑥), ∀𝑥 ∈ ℝ
 119

Exercises 5.6.12 (page 160): 10,11,12,13

 120

Chapter 6: Differentiation

Section 6.1: The Derivative

Definition 6.1.1: Let 𝐼 be an interval, let 𝑓: 𝐼 → ℝ and let 𝑐 ∈ 𝐼. We say that a real number 𝐿 is
𝑓(𝑥)−𝑓(𝑐)
the derivative of 𝒇 at 𝒄, if lim 𝑥−𝑐 = 𝐿. In this case we say that 𝑓 is differentiable at 𝑐 and we
𝑥→𝑐
𝑓(𝑥)−𝑓(𝑐)
write 𝑓 ′ (𝑐) = 𝐿, that is: 𝑓 ′ (𝑐) = lim
𝑥→𝑐 𝑥−𝑐

Remark 6.1.2: According to Definition 1, observe that:

𝑓(𝑥)−𝑓(𝑐)
(1) If lim 𝑥−𝑐 does not exist, we say that 𝑓 is not differentiable at 𝑐.
𝑥→𝑐
(2) The function 𝑓 is differentiable at 𝑐 with 𝑓 ′ (𝑐) = 𝐿 iff 𝑓+′ (𝑐) = 𝐿 and 𝑓−′ (𝑐) = 𝐿, where:
𝑓(𝑥) − 𝑓(𝑐) 𝑓(𝑥) − 𝑓(𝑐)
𝑓+′ (𝑐) = lim+ 𝑎𝑛𝑑 𝑓−′ (𝑐) = lim−
𝑥→𝑐 𝑥−𝑐 𝑥→𝑐 𝑥−𝑐
(3) In the case when the interval 𝐼 = [𝑎, 𝑏].
 Taking 𝑐 = 𝑎, Definition 1 becomes as follows:
𝑓(𝑥)−𝑓(𝑎)
If lim+ 𝑥−𝑎 exists, then the derivative of 𝑓 at 𝑎 from right exists and is written as
𝑥→𝑎
′ (𝑎) 𝑓(𝑥)−𝑓(𝑎)
𝑓+ = lim+ (that is 𝑓 is differentiable from right at 𝑎)
𝑥→𝑎 𝑥−𝑎
 Taking 𝑐 = 𝑏, Definition 1 becomes as follows:
𝑓(𝑥)−𝑓(𝑏)
If lim− 𝑥−𝑏 exists, then the derivative of 𝑓 at 𝑏 from left exists and written as
𝑥→𝑏
′ (𝑏) 𝑓(𝑥)−𝑓(𝑏)
𝑓− = lim (that is 𝑓 is differentiable from left at 𝑏)
𝑥→𝑏 − 𝑥−𝑏
(4) We say that the function 𝑓 is differentiable on an interval 𝐼 if it is differentiable at every
point in 𝐼. Also, when 𝐼 = [𝑎, 𝑏], by saying 𝑓 is differentiable on [𝑎, 𝑏] we mean that
𝑓+′ (𝑐) = 𝑓−′ (𝑐), ∀𝑐 ∈ (𝑎, 𝑏) and both 𝑓+′ (𝑎), 𝑓−′ (𝑏) exist.
(5) The derivative of 𝑓 at 𝑐 can be written as:
𝑓(𝑐 + ℎ) − 𝑓(𝑐)
𝑓 ′ (𝑐) = lim .
ℎ→0 ℎ
This can be obtained by taking ℎ = 𝑥 − 𝑐.
Example 6.1.3: Using definition, show that 𝑓(𝑥) = 𝑥 √𝑥 + 1 is differentiable at
𝑥 = 3 and find 𝑓 ′ (3).
Solution:
𝑓(𝑥) − 𝑓(3) 𝑥√𝑥 + 1 − 3√4
𝑓 ′ (3) = lim = lim
𝑥→3 𝑥−3 𝑥→3 𝑥−3
𝑥√𝑥 + 1 − 𝑥√4 + 𝑥√4 − 6
= lim
𝑥→3 𝑥−3
𝑥√𝑥 + 1 − 2𝑥 2𝑥 − 6 𝑥(√𝑥 + 1 − 2) 2(𝑥 − 3)
= lim + lim = lim + lim
𝑥→3 𝑥−3 𝑥→3 𝑥 − 3 𝑥→3 𝑥−3 𝑥→3 𝑥 − 3
√𝑥 + 1 − 2 √𝑥 + 1 + 2 3 𝑥−3 3 11
= 3 lim × + 2 = lim +2= +2=
𝑥→3 𝑥−3 √𝑥 + 1 + 2 4 𝑥→3 𝑥 − 3 4 4
 121

Example 6.1.4: Show that 𝑓 is not differentiable at 𝑥 = 0, where

2
𝑓(𝑥) = {2𝑥 −2 𝑥 𝑥 ≥ 0}
𝑥 𝑥<0
Solution:
′ (0)
𝑓(𝑥) − 𝑓(0) 2𝑥 − 𝑥 2
𝑓+ = lim+ = lim+ = lim+ (2 − 𝑥) = 2
𝑥→0 𝑥−0 𝑥→0 𝑥 𝑥→0
2
𝑓(𝑥) − 𝑓(0) 𝑥
𝑓−′ (0) = lim− = lim− = lim+ 𝑥 = 0
𝑥→0 𝑥−0 𝑥→0 𝑥 𝑥→0
⇒ 𝑓+′ (0) ≠ 𝑓−′ (0) ⇒ 𝑓 is not differentiable at 𝑥 = 0.

Observe that in Example 6.1.4: The function 𝑓 is continuous at 𝑥 = 0 because:

lim+ 𝑓 = lim+(2𝑥 − 𝑥 2 ) = 0 and lim− 𝑓 = lim+ 𝑥 2 = 0 ⇒ lim 𝑓 = 0 = 𝑓(0)
𝑥→0 𝑥→0 𝑥→0 𝑥→0 𝑥→0
⇒ lim 𝑓 = 𝑓(0) which means 𝑓 is continuous at 𝑥 = 0 (but 𝑓 is not differentiable at 𝑥 = 0)
𝑥→0
So, if a function 𝑓 is continuous at a point 𝑥 = 𝑐, then 𝑓 need not be differentiable at 𝑥 = 𝑐.

Theorem 6.1.5: If 𝑓: 𝐼 → ℝ has a derivative at 𝑐 ∈ 𝐼, then 𝑓 is continuous at 𝑐.

Proof: Since 𝑓 has a derivative at 𝑐, we have 𝑓 ′ (𝑐) exists and is a real number.
𝑓(𝑥) − 𝑓(𝑐)
lim(𝑓(𝑥) − 𝑓(𝑐)) = lim × (𝑥 − 𝑐)
𝑥→𝑐 𝑥→𝑐 𝑥−𝑐
𝑓(𝑥) − 𝑓(𝑐)
= lim × lim(𝑥 − 𝑐) = 𝑓 ′ (𝑐) × 0 = 0
𝑥→𝑐 𝑥−𝑐 𝑥→𝑐
⇒ lim(𝑓(𝑥) − 𝑓(𝑐)) = 0 ⇒ lim 𝑓(𝑥) = 𝑓(𝑐) ⇒ 𝑓 is continuous at 𝑐.
𝑥→𝑐 𝑥→𝑐
Remark 6.1.6: The contraposition of Theorem 5 says that: Let 𝐼 be an interval in ℝ
If 𝑓 is discontinuous at 𝑐 ∈ 𝐼, then 𝑓 is not differentiable at 𝑐.

Theorem 6.1.7: Let 𝐼 be an interval in ℝ and let 𝑓: 𝐼 → ℝ, 𝑔: 𝐼 → ℝ be functions that are

differentiable at 𝑐 ∈ 𝐼. Then
(1) If 𝛼 ∈ ℝ, then the function 𝛼𝑓 is differentiable at 𝑐 with (𝛼𝑓)′ (𝑐) = 𝛼𝑓 ′ (𝑐).
(2) The function 𝑓 + 𝑔 is is differentiable at 𝑐 with (𝑓 + 𝑔)′ (𝑐) = 𝑓 ′ (𝑐) + 𝑔′ (𝑐)
(3) The function 𝑓 − 𝑔 is is differentiable at 𝑐 with (𝑓 − 𝑔)′ (𝑐) = 𝑓 ′ (𝑐) − 𝑔′ (𝑐)
(4) (The Product Rule): The function 𝑓𝑔 is is differentiable at 𝑐 with
(𝑓𝑔)′ (𝑐) = 𝑓(𝑐)𝑔′ (𝑐) + 𝑔(𝑐)𝑓 ′ (𝑐)
𝑓 𝑓 ′
(5) (The Quotient Rule): If 𝑔(𝑐) ≠ 0, then the function 𝑔 is differentiable at 𝑐 with (𝑔) (𝑐) =
𝑔(𝑐)𝑓 ′ (𝑐)−𝑓(𝑐)𝑔′ (𝑐)
𝑔2 (𝑐)

The Chain Rule 6.1.8: Let 𝐼, 𝐽 be intervals in ℝ, let 𝑔: 𝐼 → ℝ, 𝑓: 𝐽 → ℝ be functions such that
𝑓(𝐽) ⊆ 𝐼, and let 𝑐 ∈ 𝐽. If 𝑓 is differentiable at 𝑐 and 𝑔 is differentiable at 𝑓(𝑐), then the composition
function 𝑔 ∘ 𝑓 is differentiable at 𝑐 with
(𝑔 ∘ 𝑓 )′ (𝑐) = 𝑔′ (𝑓(𝑐)) ∙ 𝑓 ′ (𝑐)
 122

Theorem 6.1.9: Let 𝑓 and 𝑔 be functions defined on ℝ that are differentiable at 𝑐 ∈ ℝ. Then
𝑓(𝑥) , 𝑥 ∈ ℚ
𝛹(𝑥) = { }
𝑔(𝑥) , 𝑥 ∈ ℚ𝑐
is differentiable at 𝑐 with 𝛹 ′ (𝑐) = 𝐿 iff ⏟
𝑓(𝑐) = 𝑔(𝑐) and 𝑓 ′ (𝑐) = 𝑔′ (𝑐) = 𝐿.
𝛹 is continuous at 𝑐

Example 6.1.10: Show that the Dirichlet’s function is nowhere differentiable.

1 , 𝑥 rational
Solution: The Dirichlet’s function is 𝛹(𝑥) = { }. So it is continuous at the
−1 , 𝑥 irrational
values of 𝑥 only when 1 = −1. But 1 ≠ −1 ⇒ 𝛹 is nowhere continuous
⇒ 𝛹 is nowhere differentiable.

Example 6.1.11: Find all values of 𝑥 at which the function 𝛹 is differentiability, where
𝑥 3 + 𝑥 , 𝑥 rational
𝛹(𝑥) = { }
−2𝑥 2 , 𝑥 irrational
Solution: Let 𝑓(𝑥) = 𝑥 3 + 𝑥 and 𝑔(𝑥) = −2𝑥 2 .
Step 1: We study the continuity of 𝛹:
𝑓(𝑥) = 𝑔(𝑥) ⇒ 𝑥 3 + 𝑥 = −2𝑥 2 ⇒ 𝑥 3 + 2𝑥 2 + 𝑥 = 0 ⇒ 𝑥(𝑥 2 + 2𝑥 + 1) = 0
⇒ 𝑥(𝑥 + 1)2 = 0 ⇒ 𝑥 = 0, −1 ⇒ 𝛹 is continuous only at 𝑥 = 0, −1.
Step 2: We study the differentiability at the points at which 𝛹 is continuous:
Now: 𝑓 ′ (𝑥) = 3𝑥 2 + 1 and 𝑔′ (𝑥) = −4𝑥:
 At 𝑥 = 0: 𝑓 ′ (0) = 1 and 𝑔′ (0) = 0 ⇒ 𝑓 ′ (0) ≠ 𝑔′ (0)
⇒ 𝛹 is not differentiable at 𝑥 = 0.
 At 𝑥 = −1: 𝑓 ′ (−1) = 4 and 𝑔′ (−1) = 4 ⇒ 𝑓 ′ (−1) = 𝑔′ (−1) = 4
⇒ 𝛹 is differentiable at 𝑥 = −1
⇒ 𝛹 is differentiable only at 𝑥 = −1 with 𝛹 ′ (−1) = 4.
 123

Section 6.2: The Mean Value Theorem

Roll’s Theorem 6.2.1: Let 𝑓: [𝑎, 𝑏] → ℝ be satisfying:

(1) 𝑓 is continuous on [𝑎, 𝑏]
(2) 𝑓 is differentiable on (𝑎, 𝑏)
and
(3) 𝑓(𝑎) = 0 and 𝑓(𝑏) = 0.
Then ∃𝑐 ∈ (𝑎, 𝑏) such that 𝑓 ′ (𝑐) = 0

Remark 6.2.2: Recall Bolzano’s Intermediate Value Theorem 5.3.13 which says the following:
If 𝑓: [𝑎, 𝑏] → ℝ is continuous on [𝑎, 𝑏] and 𝑓(𝑎)𝑓(𝑏) < 0, then ∃𝑐 ∈ (𝑎, 𝑏) such that 𝑓(𝑐) = 0

Example 6.2.3: Show that the equation 𝑥 5 + 𝑥 + 1 = 0 has exactly one real root in ℝ.
Solution: Let 𝑓(𝑥) = 𝑥 5 + 𝑥 + 1, 𝑥 ∈ ℝ
Step 1: First we show that the equation 𝑥 5 + 𝑥 + 1 = 0 has a real root in ℝ
⇒ 𝑓 is continuous on ℝ.
𝑓(−1)𝑓(1) = −1(1) < 0
Bolzano’s Intermediate Value Theorem ⇒ ∃𝑐 ∈ (−1,1) such that 𝑓(𝑐) = 0
𝑐 5 + 𝑐 + 1 = 0 ⇒ 𝑐 is a real root of the equation 𝑥 5 + 𝑥 + 1 = 0
⇒ the equation 𝑥 5 + 𝑥 + 1 = 0 has a real root in ℝ
Step 2: Second we show that the equation 𝑥 5 + 𝑥 + 1 = 0 has a unique real root
We use proof by contradiction:
Assume that there is another real root 𝑐1 of the equation 𝑥 5 + 𝑥 + 1 = 0 other than 𝑐.
⇒ 𝑓(𝑐) = 0 and 𝑓(𝑐1 ) = 0. Let 𝐼 be the interval with end points 𝑐 and 𝑐1.
Now,
(1) 𝑓 is continuous
(2) differentiable on 𝐼
(3) 𝑓(𝑐) = 0 and 𝑓(𝑐1 ) = 0
Roll’s Theorem 6.2.1⇒ ∃𝑐2 ∈ 𝐼 such that 𝑓 ′ (𝑐2 ) = 0
⇒ 5(𝑐2 )4 + 1 = 0 which is impossible
⇒ Our assumption is false ⇒ The equation has exactly one real root 𝑐.

The Mean Value Theorem 6.2.4: Let 𝑓: [𝑎, 𝑏] → ℝ be a function satisfying:

(1) 𝑓 is continuous on [𝑎, 𝑏]
and
(2) 𝑓 is differentiable on (𝑎, 𝑏).
𝑓(𝑏)−𝑓(𝑎)
Then ∃𝑐 ∈ (𝑎, 𝑏) such that 𝑓 ′ (𝑐) = 𝑏−𝑎 .
Proof: Define a function 𝑔: [𝑎, 𝑏] → ℝ by
𝑓(𝑏) − 𝑓(𝑎)
𝑔(𝑥) = 𝑓(𝑥) − [𝑓(𝑎) + ( ) (𝑥 − 𝑎)] , 𝑥 ∈ [𝑎, 𝑏].
𝑏−𝑎
Then
(1) 𝑔 is continuous on [𝑎, 𝑏]
 124

(2) 𝑔 is differentiable on (𝑎, 𝑏)

(3)
𝑓(𝑏)−𝑓(𝑎)
 𝑔(𝑎) = 𝑓(𝑎) − [𝑓(𝑎) + ( ) (𝑎 − 𝑎)] = 0
𝑏−𝑎
𝑓(𝑏)−𝑓(𝑎)
 𝑔(𝑏) = 𝑓(𝑏) − [𝑓(𝑎) + ( ) (𝑏 − 𝑎)] = 0
𝑏−𝑎
𝑓(𝑏)−𝑓(𝑎)
Roll’s Theorem 6.2.1 ⇒ ∃𝑐 ∈ (𝑎, 𝑏) such that 𝑔′ (𝑐) = 0 ⇒ 𝑓 ′ (𝑐) − =0
𝑏−𝑎
𝑓(𝑏)−𝑓(𝑎)
⇒ 𝑓 ′ (𝑐) = .
𝑏−𝑎

Corollary 6.2.5: Let 𝑓: [𝑎, 𝑏] → ℝ be continuous on [𝑎, 𝑏] and differentiable on (𝑎, 𝑏).
(1) If 𝑓 ′ (𝑥) = 0, ∀𝑥 ∈ (𝑎, 𝑏), then 𝑓(𝑥) = 𝐶, ∀𝑥 ∈ [𝑎, 𝑏], where 𝐶 is a constant.
(2) If 𝑓 ′ (𝑥) ≥ 0, ∀𝑥 ∈ (𝑎, 𝑏), then 𝑓 is increasing on [𝑎, 𝑏].
(3) If 𝑓 ′ (𝑥) ≤ 0, ∀𝑥 ∈ (𝑎, 𝑏), then 𝑓 is decreasing on [𝑎, 𝑏].
Proof:
(1) Let 𝑥 ∈ (𝑎, 𝑏] and consider the function 𝑓 on the interval [𝑎, 𝑥]. Then
 𝑓 is continuous on [𝑎, 𝑥] (since 𝑓 continuous on [𝑎, 𝑏] and [𝑎, 𝑥] ⊆ [𝑎, 𝑏])
 𝑓 differentiable on (𝑎, 𝑏) (since 𝑓 differentiable on (𝑎, 𝑏) and (𝑎, 𝑥) ⊆ (𝑎, 𝑏))
The Mean Value Theorem ⇒ ∃𝑐 ∈ (𝑎, 𝑥) such that
𝑓(𝑥)−𝑓(𝑎)
𝑓 ′ (𝑐) = 𝑥−𝑎 ………①
Since 𝑓 ′ (𝑥) = 0, ∀𝑥 ∈ (𝑎, 𝑏) and 𝑐 ∈ (𝑎, 𝑏) ⇒ 𝑓 ′ (𝑐) = 0
𝑓(𝑥)−𝑓(𝑎)
The relation ① ⇒ 𝑥−𝑎 = 0 ⇒ 𝑓(𝑥) = 𝑓(𝑎), ∀𝑥 ∈ (𝑎, 𝑏].
Take 𝐶 = 𝑓(𝑎) ⇒ 𝑓(𝑥) = 𝐶, ∀𝑥 ∈ [𝑎, 𝑏].
(2) Let 𝑥1 , 𝑥2 ∈ [𝑎, 𝑏] with 𝑥1 < 𝑥2 . Then
 𝑓 is continuous on [𝑥1 , 𝑥2 ]
 differentiable on (𝑥1 , 𝑥2 )
The Mean Value Theorem ⇒ ∃𝑐 ∈ (𝑥1 , 𝑥2 ) ⊆ (𝑎, 𝑏) such that
𝑓(𝑥 )−𝑓(𝑥1 )
𝑓 ′ (𝑐) = 2 ……..②
𝑥2 −𝑥1
Since 𝑓 ′ (𝑥) ≥ 0, ∀𝑥 ∈ (𝑎, 𝑏) and 𝑐 ∈ (𝑎, 𝑏) ⇒ 𝑓 ′ (𝑐) ≥ 0
𝑓(𝑥 )−𝑓(𝑥 )
The relation ② ⇒ 𝑥2 −𝑥 1 ≥ 0 ⇒ 𝑓(𝑥2 ) ≥ 𝑓(𝑥1 ).
2 1
⇒ 𝑓 is increasing on [𝑎, 𝑏].
 125

Darboux’s Theorem 6.2.6: Let 𝑓: [𝑎, 𝑏] → ℝ be a function such that

(1) differentiable on [𝑎, 𝑏]
and
(2) 𝑘 is a number between 𝑓 ′ (𝑎) and 𝑓 ′ (𝑏).
Then ∃𝑐 ∈ (𝑎, 𝑏) such that 𝑓 ′ (𝑐) = 𝑘.
Example 6.2.7: Show that there is no function 𝑓: ℝ → ℝ such that
2 , 𝑥≥0
𝑓 ′ (𝑥) = {
−3 , 𝑥 < 0
Solution:

Example 6.2.8: Is there a function 𝑓

defined on the interval [−3,1] whose
derivative 𝑓 ′ is given in the figure.

Example 6.2.9: Is there a function 𝑓

defined on the interval [−1,1] whose
derivative 𝑓 ′ is given in the figure.

Relative Extrema Theorem 6.2.10: Let 𝑛 ∈ ℕ and 𝑓: [𝑎, 𝑏] → ℝ be such that 𝑓 ′ , 𝑓 ′′ , … , 𝑓 (𝑛) are
continuous on a neighborhood 𝑉𝛿 (𝑥0 ) = (𝑥0 − 𝛿, 𝑥0 + 𝛿) of some point 𝑥0 ∈ (𝑎, 𝑏) and
𝑓 ′ (𝑥0 ) = 0, 𝑓 ′′ (𝑥0 ) = 0, … , 𝑓 (𝑛−1) (𝑥0 ) = 0 and 𝑓 (𝑛) (𝑥0 ) ≠ 0.
(1) If 𝑛 is even and 𝑓 (𝑛) (𝑥0 ) > 0, then 𝑓 has a relative minimum at 𝑥0 .
(2) If 𝑛 is even and 𝑓 (𝑛) (𝑥0 ) < 0, then 𝑓 has a relative maximum at 𝑥0 .
(3) If 𝑛 is odd, then 𝑓 has neither a relative maximum nor a relative minimum at 𝑥0 .
 126

Example 6.2.11: Determine whether the function has a relative maximum, a relative minimum, or
neither a relative extremum at 𝑥 = 0:
1
(1) 𝑓(𝑥) = sin2 (𝑥) + 3 𝑥 4 − 𝑥 2
𝑥3
(2) 𝑓(𝑥) = 6
− 𝑥 + sin(𝑥)
𝑥2
(3) 𝑓(𝑥) = 1 + 2
+ cos(𝑥)
Solution:
4 4
(1) 𝑓 ′ (𝑥) = 2sin𝑥 cos𝑥 + 3 𝑥 3 − 2𝑥 = sin(2𝑥) + 3 𝑥 3 − 2𝑥 ⇒ 𝑓 ′ (0) = 0
𝑓 ′′ (𝑥) = 2cos(2𝑥) + 4𝑥 2 − 2 ⇒ 𝑓 ′′ (0) = 0
𝑓 ′′′ (𝑥) = −4sin(2𝑥) + 8𝑥 ⇒ 𝑓 ′′′ (0) = 0
𝑓 (4) (𝑥) = −8cos(2𝑥) + 8 ⇒ 𝑓 (4) (0) = 0
𝑓 (5) (𝑥) = 16sin(2𝑥) ⇒ 𝑓 (5) (0) = 0
𝑓 (6) (𝑥) = 32cos(2𝑥) ⇒ 𝑓 (6) (0) = 32 ≠ 0
⇒ 𝑛 = 6 is even and 𝑓 (6) (0) > 0 ⇒ 𝑓 has a relative minimum at 0.

𝑥2
(2) 𝑓 ′ (𝑥) = 2 − 1 + cos(𝑥) ⇒ 𝑓 ′ (0) = 0
𝑓 ′′ (𝑥) = 𝑥 − sin(𝑥) ⇒ 𝑓 ′′ (0) = 0
𝑓 ′′′ (𝑥) = 1 − cos(𝑥) ⇒ 𝑓 ′′′ (0) = 0
𝑓 (4) (𝑥) = sin(𝑥) ⇒ 𝑓 (4) (0) = 0
𝑓 (5) (𝑥) = cos(𝑥) ⇒ 𝑓 (5) (0) = 1 ≠ 0
⇒ 𝑛 = 5 is odd ⇒ 𝑓 has neither a relative max. nor a relative min. at 0
(3) Is an exercise
 127

Section 6.3: L’Hospital’s Rules

𝑓(𝑥) 𝐴
The indeterminate forms of the limit lim 𝑔(𝑥) = 𝐵 are when:
𝑥→𝑐
𝐴 0 ∞
= 0 , ∞, ∞ − ∞, 0 ∙ ∞, 00 , 1∞ , or ∞0
𝐵
L’Hospital’s Rule (1) 6.3.1: Let−∞ ≤ 𝑎 < 𝑏 ≤ ∞ and let 𝑓, 𝑔 be functions satisfying the following:
(1) 𝑓, 𝑔 are differentiable on (𝑎, 𝑏)
(2) 𝑔′ (𝑥) ≠ 0, ∀𝑥 ∈ (𝑎, 𝑏)
(3) lim+ 𝑓(𝑥) = 0 and lim+ 𝑔(𝑥) = 0
𝑥→𝑎 𝑥→𝑎
𝑓 ′ (𝑥)
(4) lim+ 𝑔′ (𝑥) = 𝐿 ∈ ℝ ∪ {−∞, ∞}.
𝑥→𝑎
𝑓(𝑥) 𝑓 ′ (𝑥)
Then lim+ 𝑔(𝑥) = lim+ 𝑔′ (𝑥) = 𝐿
𝑥→𝑎 𝑥→𝑎

𝑓(𝑥)
Example 6.3.2: Find lim 𝑔(𝑥). Justify your answer, where
𝑥→0
2
1
𝑓(𝑥) = 𝑥 sin ( ) for 𝑥 ≠ 0 and 𝑓(0) = 0 and let 𝑔(𝑥) = sin𝑥, 𝑥 ∈ ℝ
𝑥
1
𝑥 2 sin( ) 0
𝑥
Solution: Observe that lim+ =
𝑥→0 sin 𝑥 0
So, we think of using L’Hospital’s Rule (1) to find this limit
Since 𝑥 → 0 ⇒ Take the intervals (−1,0) ∪ (0,1)
1
Let 𝑓(𝑥) = 𝑥 2 sin (𝑥) and 𝑔(𝑥) = sin𝑥 are diff. on (−1,0) ∪ (0,1)
1
(1) 𝑓(𝑥) = 𝑥 2 sin (𝑥) and 𝑔(𝑥) = sin𝑥 are diff. on (−1,0) ∪ (0,1)
𝑑
(2) 𝑔′ (𝑥) = 𝑑𝑥 (sin𝑥) = cos𝑥 ≠ 0, ∀𝑥 ∈ (−1,0) ∪ (0,1)
1
(3) lim 𝑓(𝑥) = lim𝑥 2 sin (𝑥) = 0 (by the squeeze Theorem)
𝑥→0 𝑥→0
lim 𝑔(𝑥) = lim sin𝑥 = 0
𝑥→0 𝑥→0
1 −1 1 1 1
𝑓 ′ (𝑥) 𝑥 2 cos( )( 2 )+2𝑥sin( ) −cos( )+2𝑥sin( )
𝑥 𝑥 𝑥 𝑥 𝑥
(4) lim = lim = lim
𝑥→0 𝑔′ (𝑥) 𝑥→0 cos𝑥 𝑥→0 cos𝑥
1 1
−lim cos( )+2lim 𝑥sin( ) DNE+2(0)
𝑥→0 𝑥 𝑥→0 𝑥
= == = DNE ∉ ℝ ∪ {±∞}
lim cos𝑥 1
𝑥→0
⇒ L’Hospital’s Rule (1) fails to answer this question. We use another method:
1
𝑓(𝑥) 𝑥 2 sin( )
𝑥
lim = lim
𝑥→0 𝑔(𝑥) 𝑥→0 sin𝑥
𝑥 1
= lim (sin𝑥) (𝑥sin (𝑥))
𝑥→0
𝑥 1
= lim (sin𝑥) lim (𝑥sin (𝑥))
𝑥→0 𝑥→0
= 1(0) = 0
𝑓(𝑥)
⇒ lim 𝑔(𝑥) = 0
𝑥→0
 128

ln(cos𝑥)
Example 6.3.3: Find lim+ . Justify your answer.
𝑥→0 𝑥
ln(cos𝑥) 0
Solution: Observe that lim+ =0
𝑥→0 𝑥
So, we think of using L’Hospital’s Rule (1) to find this limit
𝜋
Since 𝑥 → 0+ , take the interval (0, 4 )
𝜋
Let 𝑓(𝑥) = ln(cos(𝑥)), 𝑔(𝑥) = 𝑥, 𝑥 ∈ (0, 4 ):
𝜋
(1) 𝑓(𝑥) = ln(cos𝑥) and 𝑔(𝑥) = 𝑥 are diff. on (0, 4 )
𝜋
(2) 𝑔′ (𝑥) = 1 ≠ 0, ∀𝑥 ∈ (0, 4 )
(3) lim+𝑓(𝑥) = lim+ln(cos𝑥) = ln(1) = 0 and lim+ 𝑔(𝑥) = lim+ x = 0
𝑥→0 𝑥→0 𝑥→0 𝑥→0
sin(𝑥)
𝑓′ (𝑥) (− ) −sin(𝑥) 0
cos(𝑥)
(4) lim = lim+ = lim = 1 = 0 ∈ ℝ ∪ {±∞}
𝑥→0 𝑔′ (𝑥) 𝑥→0 1 𝑥→0+ cos(𝑥)
ln(cos𝑥) 𝑓(𝑥) 𝑓′ (𝑥)
By L’Hospital’s Rule (1) we have lim+ = lim+ 𝑔(𝑥) = lim =0
𝑥→0 𝑥 𝑥→0 𝑥→0 𝑔′ (𝑥)

L’Hospital’s Rule (2) 6.3.4: Let−∞ ≤ 𝑎 < 𝑏 ≤ ∞ and let 𝑓, 𝑔 be functions satisfying the following:
(1) 𝑓, 𝑔 are differentiable on (𝑎, 𝑏)
(2) 𝑔′ (𝑥) ≠ 0, ∀𝑥 ∈ (𝑎, 𝑏)
(3) lim+ 𝑔(𝑥) = ∞ or − ∞
𝑥→𝑎
𝑓 ′ (𝑥)
(4) lim+ 𝑔′ (𝑥) = 𝐿 ∈ ℝ ∪ {−∞, ∞}.
𝑥→𝑎
𝑓(𝑥) 𝑓 ′ (𝑥)
Then lim+ 𝑔(𝑥) = lim+ 𝑔′ (𝑥) = 𝐿
𝑥→𝑎 𝑥→𝑎

𝑥+sin𝑥
Example 6.3.5: Find lim . Justify your answer.
𝑥→∞ 𝑥−sin𝑥
𝑥+sin𝑥 ∗
Solution: Observe that lim =∞.
𝑥→∞ 𝑥−sin𝑥

Since 𝑥 → ∞ ⇒ Take the interval (0, ∞)

Let 𝑓(𝑥) = 𝑥 + sin 𝑥 and 𝑔(𝑥) = 𝑥 − sin 𝑥, 𝑥 ∈ (0, ∞):
 𝑓(𝑥) = 𝑥 + sin 𝑥 and 𝑔(𝑥) = 𝑥 − sin 𝑥 are diff. on (0, ∞)
 𝑔′ (𝑥) = 1 − cos 𝑥 = 0 for all 𝑥 = 2𝑛𝜋, ∀𝑛 ∈ ℕ
⇒ L’Hospital’s Rule (2) fails to answer this question. We use another method:
sin𝑥 sin𝑥
𝑥+sin𝑥 1+ 1+ lim 1+0
𝑥 𝑥→∞ 𝑥
lim ⏟ = lim sin𝑥 = sin𝑥 = =1
𝑥→∞ 𝑥−sin𝑥 𝑥→∞ 1− 1− lim 1−0
𝑥 𝑥→∞ 𝑥
𝒙 ‫نقسم البسط والمقام على‬
 129

ln(sin𝑥)
Example 6.3.6: Find lim+ . Justify your answer.
𝑥→0 ln𝑥
ln(sin𝑥) ∗
Solution: Observe that lim+ = −∞
𝑥→0 ln𝑥

So, we think of using L’Hospital’s Rule (2) to find this limit

Since 𝑥 → 0+ ⇒ Take the interval (0,1)
Let 𝑓(𝑥) = ln(sin𝑥) and 𝑔(𝑥) = ln𝑥, 𝑥 ∈ (0,1)
 𝑓(𝑥) = ln(sin𝑥) and 𝑔(𝑥) = ln𝑥 are differentiable on (0,1)
1
 𝑔′ (𝑥) = 𝑥 ≠ 0, ∀𝑥 ∈ (0,1)

 lim+ 𝑔(𝑥) = lim+ ln𝑥 = −∞

𝑥→0 𝑥→0
cos𝑥
𝑓 ′ (𝑥) 𝑥cos𝑥 𝑥
 lim+ = lim+ sin𝑥
1 = lim+ = lim+ ( ) cos𝑥 = 1(1) = 1 ∈ ℝ ∪ {±∞}
𝑥→0 𝑔′ (𝑥) 𝑥→0 𝑥→0 sin𝑥 𝑥→0 sin𝑥
𝑥

ln(sin𝑥) 𝑓(𝑥) 𝑓 ′ (𝑥)

By L’Hospital’s Rule (2) we have lim+ = lim+ 𝑔(𝑥) = lim =1
𝑥→0 ln𝑥 𝑥→0 𝑥→0 𝑔′ (𝑥)