The Definition of Limits

The definition of a limit has both an intuitive and a rigorous form. Understanding both is important for mastering the concept of limits.

The Intuitive Definition

The intuitive definition of a limit

If, as nn increases without bound, the terms ana_n of the sequence {an}\{a_n\} get infinitely close to a certain constant AA, then the sequence {an}\{a_n\} is said to converge to AA, and AA is called the limit of the sequence.

Written as: limnan=A\lim_{n \to \infty} a_n = A or anA(n)a_n \to A \quad (n \to \infty)

符号说明
符号类型读音/说明在本文中的含义
lim\lim运算符号limit(极限)The limit operator
\infty符号infinity(无穷大)Infinity; nn \to \infty means nn tends to infinity

The Rigorous Definition (the ε-N Definition)

The ε-N definition of a limit

Let {an}\{a_n\} be a sequence and AA a constant. If for any given positive number ε\varepsilon (however small), there always exists a positive integer NN such that whenever n>Nn > N,

anA<ε|a_n - A| < \varepsilon

then the sequence {an}\{a_n\} is said to converge to AA, written as limnan=A\lim_{n \to \infty} a_n = A.

符号说明
符号类型读音/说明在本文中的含义
ε\varepsilon希腊字母Epsilon(伊普西隆)An arbitrarily small positive number specifying the degree of closeness
NN变量NA positive integer indicating from which term the condition holds
Why do we need a rigorous definition?

The intuitive definition is easy to understand, but not precise enough:

  • What does “infinitely close” mean?
  • To what extent does “increase without bound” go?

The rigorous ε-N definition resolves these issues:

  • anA<ε|a_n - A| < \varepsilon quantifies “closeness”
  • n>Nn > N quantifies “large enough”

This rigor is the foundation of mathematical analysis, allowing us to carry out rigorous proofs.

Analogy: just as laws need precise provisions rather than merely “the general meaning”.

Convergence and Divergence

  • Convergent: the sequence has a limit
  • Divergent: the sequence has no limit

Examples of Convergence

Example 1: limn1n=0\lim_{n \to \infty} \frac{1}{n} = 0

Proof (using the ε-N definition):

For any ε>0\varepsilon > 0, to make 1n0<ε|\frac{1}{n} - 0| < \varepsilon, i.e., 1n<ε\frac{1}{n} < \varepsilon, it suffices to take n>1εn > \frac{1}{\varepsilon}.

Take N=1εN = \lceil \frac{1}{\varepsilon} \rceil (the smallest integer greater than 1ε\frac{1}{\varepsilon}). Then when n>Nn > N, 1n0<ε|\frac{1}{n} - 0| < \varepsilon.

Therefore limn1n=0\lim_{n \to \infty} \frac{1}{n} = 0.

Example 2: limn2n+1n=2\lim_{n \to \infty} \frac{2n+1}{n} = 2

2n+1n=2+1n2\frac{2n+1}{n} = 2 + \frac{1}{n} \to 2

Examples of Divergence

Example 1: an=na_n = n diverges (tends to infinity)

Example 2: an=(1)na_n = (-1)^n diverges (oscillates)

Example 3: an=sinna_n = \sin n diverges (irregular oscillation)

Practice Problems

Exercise 1

Use the ε-N definition to prove: limn1n2=0\lim_{n \to \infty} \frac{1}{n^2} = 0

Reference Answer(1 个标签)
limit definition of sequences

Proof:

For any ε>0\varepsilon > 0, to make 1n20<ε|\frac{1}{n^2} - 0| < \varepsilon, i.e., 1n2<ε\frac{1}{n^2} < \varepsilon, it suffices to have n2>1εn^2 > \frac{1}{\varepsilon}, i.e., n>1εn > \frac{1}{\sqrt{\varepsilon}}.

Take N=1εN = \lceil \frac{1}{\sqrt{\varepsilon}} \rceil. Then when n>Nn > N, 1n20<ε|\frac{1}{n^2} - 0| < \varepsilon.

Therefore limn1n2=0\lim_{n \to \infty} \frac{1}{n^2} = 0.

Exercise 2

Determine whether the following sequences converge. If they converge, find their limits:

  1. an=3n+2n+1a_n = \frac{3n+2}{n+1}
  2. an=n2n+1a_n = \frac{n^2}{n+1}
  3. an=(1)nna_n = \frac{(-1)^n}{n}
Reference Answer(1 个标签)
limit definition of sequences

Solution:

  1. an=3n+2n+1=3+2n1+1n31=3a_n = \frac{3n+2}{n+1} = \frac{3 + \frac{2}{n}}{1 + \frac{1}{n}} \to \frac{3}{1} = 3, convergent, with limit 3

  2. an=n2n+1=n1+1na_n = \frac{n^2}{n+1} = \frac{n}{1 + \frac{1}{n}} \to \infty, divergent

  3. an=(1)nna_n = \frac{(-1)^n}{n}: since an=1n0|a_n| = \frac{1}{n} \to 0, we have an0a_n \to 0, convergent, with limit 0

Exercise 3

Prove: if limnan=A\lim_{n \to \infty} a_n = A, then limnan=A\lim_{n \to \infty} |a_n| = |A|

Reference Answer(1 个标签)
limit definition of sequences

Proof:

From limnan=A\lim_{n \to \infty} a_n = A, for any ε>0\varepsilon > 0, there exists an NN such that when n>Nn > N, anA<ε|a_n - A| < \varepsilon.

Using the inequality anAanA||a_n| - |A|| \leq |a_n - A| (the triangle inequality for absolute values),

when n>Nn > N, anAanA<ε||a_n| - |A|| \leq |a_n - A| < \varepsilon.

Therefore limnan=A\lim_{n \to \infty} |a_n| = |A|.


Summary

Symbols Used in This Article

符号类型读音/说明在本文中的含义
lim\lim运算符号limitLimit
\infty符号infinityInfinity
ε\varepsilon希腊字母epsilonAn arbitrarily small positive number
NN变量NA certain positive integer

中英对照

中文术语英文术语音标说明
极限limit/ˈlɪmɪt/The eventual trend of a sequence
收敛convergent/kənˈvɜːdʒənt/The sequence has a limit
发散divergent/daɪˈvɜːdʒənt/The sequence has no limit
ε-N定义epsilon-N definition-The rigorous definition of a limit