Convergence of Series — Mathematics Knowledge System

Overview

An infinite series Σaₙ is said to converge if its sequence of partial sums Sₙ = a₁ + a₂ + ... + aₙ converges to a finite limit as n → ∞; otherwise the series diverges. Because a series is defined entirely in terms of a sequence (its partial sums), the rigorous ε-N machinery of sequential convergence transfers directly. The central question of this topic is not computing sums in closed form (only a few special series, like geometric series, allow that) but determining WHETHER a given series converges at all — a question answered by a toolkit of convergence tests (comparison, ratio, root, integral, alternating series) developed throughout the 18th and 19th centuries.

Intuition

Think of walking along a number line, taking a step of size aₙ at the n-th move. The partial sum Sₙ is your position after n steps. The series converges if your position settles down and stops meaningfully changing — the steps get so small, so fast, that you approach a fixed final resting point rather than wandering off to infinity or oscillating forever. Crucially, it's not enough for the STEPS themselves to shrink to zero (aₙ → 0) — the harmonic series 1 + 1/2 + 1/3 + ... has shrinking steps yet still walks off to infinity, because the steps don't shrink fast enough for the total distance to stay finite.

Interactive Graph

Formal Definition

Definition

Given a sequence (aₙ), the associated series Σaₙ converges to S if its sequence of partial sums converges to S in the ε-N sense:

S_n = \sum_{k=1}^{n} a_k

Partial sum

\sum_{k=1}^{\infty} a_k = S \iff \forall \varepsilon>0\ \exists N: \forall n \ge N,\ |S_n - S| < \varepsilon

The series converges to S exactly when its partial sums converge to S as a sequence

Convergence of a series

\sum_{k=1}^{\infty} |a_k| < \infty \implies \sum_{k=1}^{\infty} a_k \text{ converges}

A stronger convergence notion (absolute convergence) always implies ordinary convergence, but not conversely

Absolute convergence implies convergence

Notation

Notation	Meaning
$\sum_{n=1}^{\infty} a_n$	An infinite series formed by summing the terms of the sequence (aₙ)
$S_n$	The n-th partial sum, Σ from k=1 to n of aₖ
$\sum a_n \text{ converges but } \sum \|a_n\| \text{ diverges}$	The series converges only due to sign cancellation, not because the terms are absolutely small
$\sum \|a_n\| < \infty$	The series of absolute values converges — a strictly stronger property than plain convergence

Derivation

Proof that the harmonic series diverges, by grouping terms into blocks that each exceed 1/2 — the classical argument attributed to Nicole Oresme (14th century).

S_{2^k} = 1 + \frac{1}{2} + \underbrace{\left(\frac{1}{3}+\frac{1}{4}\right)}_{>\, 1/2} + \underbrace{\left(\frac{1}{5}+\cdots+\frac{1}{8}\right)}_{>\, 1/2} + \cdots

Group terms in blocks of doubling length; each block of the form (1/(2^j+1)+\cdots+1/2^{j+1}) contains 2^j terms each ≥ 1/2^{j+1}, so the block sum exceeds 1/2

S_{2^k} > 1 + \frac{k}{2}

After k such blocks (plus the initial 1), the partial sum exceeds 1 + k/2

\lim_{k\to\infty}\left(1+\frac{k}{2}\right) = \infty \implies \lim_{k\to\infty} S_{2^k} = \infty

Since the partial sums have an unbounded subsequence, the sequence of partial sums cannot converge to any finite limit

Properties

n-th term test (necessary condition)

\sum a_n \text{ converges} \implies a_n \to 0

Condition: Contrapositive form is most useful for proving divergence: if aₙ ↛ 0, the series diverges immediately. The converse is FALSE — aₙ→0 does not imply convergence (harmonic series).

Geometric series test

\sum_{n=0}^{\infty} ar^n = \frac{a}{1-r} \text{ converges} \iff |r| < 1

Ratio test

L = \lim_{n\to\infty}\left|\frac{a_{n+1}}{a_n}\right|;\ L<1 \Rightarrow \text{converges (absolutely)},\ L>1 \Rightarrow \text{diverges},\ L=1 \Rightarrow \text{inconclusive}

Alternating Series Test (Leibniz)

\text{If } b_n \text{ is decreasing, } b_n \to 0, \text{ then } \sum (-1)^n b_n \text{ converges}

Condition: Applies to alternating series with monotonically shrinking terms; only guarantees convergence, not absolute convergence.

p-series test

\sum_{n=1}^{\infty} \frac{1}{n^p} \text{ converges} \iff p > 1

Example: The harmonic series (p=1) diverges; Σ1/n² (p=2) converges.

Applications

Analyzing the running time or error of iterative and recursive algorithms often reduces to determining whether an associated series converges, e.g. bounding the total work done across all levels of a recursive divide-and-conquer algorithm.

Worked Examples

This is a geometric series with ratio r = 1/2, and |r|<1, so it converges.
$\left|\frac{1}{2}\right| < 1$
Apply the geometric series sum formula a/(1-r) with a=1, r=1/2.
$\sum_{n=0}^{\infty}\left(\frac{1}{2}\right)^n = \frac{1}{1-\frac{1}{2}} = 2$

Answer: Converges to 2

Practice Problems

Difficulty 5/10

Does Σ 1/n² (n=1 to ∞) converge? Justify using the p-series test.

Difficulty 6/10

If lim(n→∞) aₙ = 5 ≠ 0, what can you conclude about Σaₙ?

It converges to 5It diverges (by the n-th term test)It is inconclusive without more informationIt converges conditionally

Common Mistakes

Common Mistake

Believing aₙ → 0 is sufficient to guarantee Σaₙ converges.

aₙ→0 is only NECESSARY, not sufficient. The harmonic series Σ1/n has terms shrinking to 0 yet diverges to infinity — the terms must shrink fast enough, which aₙ→0 alone does not guarantee.

Common Mistake

Applying the ratio test and concluding divergence (or convergence) when the limit L equals exactly 1.

L=1 is the inconclusive case — the ratio test gives no information. Both Σ1/n (diverges) and Σ1/n² (converges) have ratio-test limit L=1, so a different test (e.g. p-series test) is required.

Quiz

The n-th term test says that if the terms aₙ do NOT approach 0, the series:

convergesdivergesconverges conditionallycannot be determined

The harmonic series Σ1/n famously:

converges to 1diverges, even though its terms 1/n → 0converges to eoscillates

Summary

A series Σaₙ converges iff its sequence of partial sums Sₙ converges (in the ordinary ε-N sequence sense) to a finite value.
The n-th term test gives a necessary (not sufficient) condition: aₙ→0 is required for convergence, but not enough on its own (harmonic series diverges despite aₙ→0).
Key convergence tests: geometric series (|r|<1), ratio test, p-series (p>1), and the Alternating Series Test for series with alternating signs.
Absolute convergence (Σ|aₙ| converges) is strictly stronger than convergence — a series can be conditionally convergent, converging only due to sign cancellation, as with Σ(-1)ⁿ/n.
The harmonic series Σ1/n is the canonical example of a divergent series with terms tending to zero, provable by grouping terms into blocks each exceeding 1/2.

References

WebsiteWikipedia — Convergent series
BookRudin, W. Principles of Mathematical Analysis, 3rd ed. Ch. 3.