Series Convergence Tests — Lecture 3 Detailed Notes

Course Logistics

  • Practice material now online

    • 1 authentic past mid-semester exam + 1 extra practice paper

    • Format, breadth & difficulty mirror upcoming Week-6 mid-semester exam

  • Weekly extra problems

    • Scans of textbook questions + selected solutions

    • Full solutions not posted → ask in office hours, drop-in centre, or demonstrator consultation

  • Assignment 1

    • Released end of Week 3 / start of Week 4

    • 4–5 questions, each with multiple parts

    • Due sometime in Week 7 (after the mid-semester exam)

Lecture Focus

  • Few new theorems; mostly practice & technique

  • Complete leftover examples and add:

    • Alternating Series Test

    • Ratio Test

    • Root Test

    • Absolute vs Conditional Convergence


Integral Test (catch-up example)

  • Series: \sum_{k=1}^{\infty} k e^{-k}

  • Analogous function: f(x)=x e^{-x} (positive, continuous, decreasing for x\ge1)

  • Evaluate improper integral: \int_1^{\infty} x e^{-x}\,dx

    • Integration by parts

    • Let u=x\,,\; dv=e^{-x}dx\Rightarrow v=-e^{-x}

    • Result after limits: \frac{2}{e} (finite)

  • Integral converges ⇒ series converges by Integral Test

Key Technique Reminder
  • For rational functions inside limits: factor highest power of n in numerator & denominator  leftover fractions \to0.


“Continuous Function commutes with limit” Trick

  • If f is continuous and \lim_{n\to\infty}a_n=L, then \lim_{n\to\infty}f(a_n)=f(L).

  • Example series: \sum_{n=1}^{\infty} \sqrt[3]{ \frac{n^2}{n^2+20n+9} }

    • Use cubic-root continuity: pull limit inside root

    • \lim_{n\to\infty} \frac{n^2}{n^2+20n+9}=1 ⇒ outer limit =1

    • Since term limit \neq0 ⇒ series diverges by Divergence Test (a.k.a. n\to\infty term test).


Choosing a Comparison (Example \sum \frac1{1+n^2})

  • Eyeball dominant term \sim\frac1{n^2} (p-series with p=2>1 ⇒ expect convergence).

  1. Direct Comparison: \frac1{1+n^2} \le \frac1{n^2}.

  2. Limit Comparison (slicker):
    \lim_{n\to\infty}\frac{ \frac1{1+n^2} }{ \frac1{n^2} }=1 ⇒ same behaviour ⇒ convergent.

  3. Integral Test also works but more labour.


Alternating Series Test (Leibniz Criterion)

  • Series of form \sum_{n=1}^{\infty} (-1)^{n-1} b_n with

    1. b_n>0

    2. b_{n+1}\le b_n (monotonically decreasing)

    3. \lim_{n\to\infty} b_n =0
      ⇒ Series converges (actually conditionally unless also absolutely conv.).

Classic Example: Alternating Harmonic Series

\sum_{n=1}^{\infty} (-1)^{n-1}\frac1n = \ln 2 (convergent although \sum 1/n diverges).

Counter-Examples / Quick Checks
  • If either monotonicity or b_n\to0 fails → cannot apply test; often diverges by \lim a_n test.

    • Example a_n=(-1)^n\frac{3n}{4n-1}: b_n\not\to0 ⇒ diverges.

    • Example (-1)^n \frac{n^2}{n^2+1}: monotone ↑ and \lim b_n=1 ⇒ diverges.


Absolute vs Conditional Convergence

  • Absolute Convergence: \sum |a_n| converges.

  • Conditional Convergence: Original series converges but \sum |a_n| diverges.

  • Relation: \text{Absolute} \Rightarrow \text{Convergent} by Triangle Inequality
    \Big|\sum_{n=1}^{N} a_n\Big|\le\sum_{n=1}^{N}|a_n|.

Examples
  1. (-1)^{n}\dfrac{3n}{4n-1} – not abs. (limit ≠0) and not cond. ⇒ diverges.

  2. Alternating harmonic (-1)^{n-1}/n – convergent but not absolute ⇒ conditionally convergent.

  3. (-1)^{n-1}/n^2 – p=2 so abs. conv. ⇒ both absolute & conditional.


Ratio Test

  • Define R =\lim_{n\to\infty}\left|\frac{a_{n+1}}{a_n}\right|.

    • If R<1 ⇒ absolutely convergent.

    • If R>1 ⇒ divergent.

    • If R=1 ⇒ inconclusive.

  • Heuristic: compares with geometric series (common ratio R).

Good when series contains factorials, exponentials, large powers.
Example 1

a_n = (-1)^n\frac{n^3}{3^n}

  • \left|\frac{a_{n+1}}{a_n}\right|=\frac{(n+1)^3}{3^{n+1}}\cdot\frac{3^{n}}{n^3}=\frac{(n+1)^3}{3 n^3}

  • Pull n^3: \to\frac13\Big(1+\frac1n\Big)^3 \xrightarrow[n\to\infty]{}\frac13<1 ⇒ absolutely convergent.

Example 2 (factorial)

a_n=\frac{n^n}{n!}

  • \left|\frac{a_{n+1}}{a_n}\right|=\frac{(n+1)^{n+1}}{(n+1)!}\cdot\frac{n!}{n^n}=\frac{(n+1)^{n}}{n^n}\cdot\frac{1}{n+1}

  • Simplify: \Big(1+\frac1n\Big)^n\cdot\frac1{n+1}\to e\cdot0 = e>1

  • Actually final ratio =e>1 ⇒ diverges.


Root Test

  • Define L=\lim_{n\to\infty}\sqrt[n]{|a_n|}=\lim_{n\to\infty}|a_n|^{1/n}.

    • L<1 ⇒ absolutely convergent.

    • L>1 ⇒ divergent.

    • L=1 ⇒ inconclusive.

  • Often simpler when term already has nth power.

Example

a_k=\left( \dfrac{2k-1}{k^2+3} \right)^{!k}, k\ge2

  • Root test: L=\lim_{k\to\infty}\left|a_k\right|^{1/k}=\lim_{k\to\infty}\frac{2k-1}{k^2+3}

  • Simplify: factor k/k^2 ⇒ L=\frac{2}{1}=2>1 ⇒ diverges.


How to Select a Test (Heuristics)

  • Divergence test first: if \lim a_n\neq0 ⇒ stop (diverges).

  • p-series / geometric resemblance ⇒ comparison or limit comparison.

  • Positive, monotone, integrable function ⇒ integral test.

  • Alternating signs ⇒ try Alternating Series Test.

  • Factorials, exponentials, n^n ⇒ Ratio Test.

  • a_n^n pattern ⇒ Root Test.

  • When absolute values easier ⇒ decide absolute vs conditional.


Ethical / Practical Notes

  • Copyright notice: material reproduced under Section 113P, further reproduction prohibited.

  • Use provided past exams & extra problems responsibly; seek help but do not share full solutions online.


Quick Formula & Reference Sheet

  • Triangle Inequality (finite sums): \Big|\sum_{k=1}^N x_k\Big|\le\sum_{k=1}^N |x_k|

  • Geometric series: \sum_{n=0}^{\infty} ar^n = \frac{a}{1-r}\,(|r|<1)

  • p-series: \sum_{n=1}^{\infty} \frac1{n^p} converges \iff p>1.

  • Stirling (handy for factorial limits): n!\sim\sqrt{2\pi n}\,(n/e)^n (not used directly today but useful with Ratio/Root).


Big Take-aways

  • Master pattern recognition: decide test before heavy algebra.

  • Always check the easy tests (divergence, comparison) first.

  • Alternation aids convergence; absolute convergence is a stronger requirement.

  • Ratio & Root tests deliver fast verdicts when cancellation of large growth terms occurs.

  • Show all limit steps clearly; most proofs hinge on correct limit evaluation.