Comprehensive Guide to Series, Partial Sums, and Convergence Tests for Convergence, and Geometric Series Analysis of Series

Introduction to Series and Sigma Notation

  • Transition from Sequences: Previously, the focus was on sequences, which are lists of terms (a1,a2,a3,a_1, a_2, a_3, \dots). In sequences, we assess whether the list itself converges or diverges by taking the limit of the general term (limnan\lim_{n \to \infty} a_n), using tools like the Squeeze Theorem or comparison tests.

  • Conceptual Shift to Series: If we decide to add the terms of a sequence together rather than just list them, we create a series. This is essentially the progression from a list to a sum.

  • Sigma Notation (Σ\Sigma): The Capital Greek letter Sigma represents a sum. It tells us where to start adding and where to stop.

    • A typical series notation is written as n=1an\sum_{n=1}^{\infty} a_n, which instructs us to take the terms of the sequence and add them together forever (to infinity).

    • The starting point (index) is usually indicated at the bottom (e.g., n=1n=1), and the stopping point is indicates at the top (e.g., \infty).

Partial Sums and the Sequence of Partial Sums

  • The Partial Sum (SnS_n): Because we cannot add terms forever (the speaker notes, "I'm not going to live forever"), we look at what is called a partial sum.

  • A partial sum is a finite sum that adds the first nn terms of a series. It is denoted as:     Sn=k=1nakS_n = \sum_{k=1}^{n} a_k

  • Existence of Partial Sums: Since nn is a finite number (e.g., 10, 100, or a million), the sum of these terms for sure exists. For every integers nn, there is a specific sum SnS_n.

  • Sequence of Partial Sums: For every additional term we include, we get a new sum. This creates a new sequence composed of the individual partial sums:

    • S1=a1S_1 = a_1

    • S2=a1+a2S_2 = a_1 + a_2

    • S3=a1+a2+a3S_3 = a_1 + a_2 + a_3

    • Sn=a1+a2++anS_n = a_1 + a_2 + \dots + a_n

  • This resulting sequence (S1,S2,S3,,SnS_1, S_2, S_3, \dots, S_n) is the "sequence of partial sums."

The Definition of Series Convergence

  • Connecting Sequences to Series: The main reason sequences were studied is because the convergence of a series is defined by the convergence of its sequence of partial sums (SnS_n).

  • Convergence Criterion: If the sequence of partial sums SnS_n converges, then the limit of SnS_n as nn approaches infinity exists. Let that limit be ss:     limnSn=s\lim_{n \to \infty} S_n = s

  • If this limit exists, we say the infinite series converges, and its sum is precisely ss.

  • Divergence Criterion: If the sequence of partial sums diverges (goes to infinity, negative infinity, or oscillates), then the series itself is divergent and we cannot find a finite sum.

  • Calculus Jump: We use limits to equate the infinite sum with the limit of the partial sums:     n=1an=limnk=1nak\sum_{n=1}^{\infty} a_n = \lim_{n \to \infty} \sum_{k=1}^{n} a_k

Example Study: The Series of Natural Numbers

  • Problem: Determine the convergence of n=1n\sum_{n=1}^{\infty} n.

  • Step 1: Find the partial sum (SnS_n). This is the sum of integers from 1 to nn:     Sn=1+2+3++nS_n = 1 + 2 + 3 + \dots + n

  • Step 2: Find a formula for the partial sum. The speaker references the story of the mathematician "gaus" (Gaus) who found this formula in elementary school:     Sn=n(n+1)2S_n = \frac{n(n+1)}{2}

  • Step 3: Analyze the limit as nn \to \infty:     limnn(n+1)2=\lim_{n \to \infty} \frac{n(n+1)}{2} = \infty

  • Conclusion: Since the sequence of partial sums is divergent, the series n\sum n is divergent.

Telescoping Series and Partial Sum Formulas

  • Definition: A telescoping series is one where internal terms cancel each other out, leaving only a few beginning and ending terms. This occurs frequently with differences of fractions.

  • Example Problem: Find the sum of n=1(12n+112n+3)\sum_{n=1}^{\infty} (\frac{1}{2n+1} - \frac{1}{2n+3}).

    • Step 1: List the partial sum terms.

      • For n=1n=1: (1315)(\frac{1}{3} - \frac{1}{5})

      • For n=2n=2: (1517)(\frac{1}{5} - \frac{1}{7})

      • For n=3n=3: (1719)(\frac{1}{7} - \frac{1}{9})

      • For n=n1n=n-1: (12n112n+1)(\frac{1}{2n-1} - \frac{1}{2n+1})

      • For n=nn=n: (12n+112n+3)(\frac{1}{2n+1} - \frac{1}{2n+3})

    • Pattern Identification: The first fraction of one term cancels with the last fraction of the previous term. Following the pattern carefully:

      • 1/5-1/5 cancels with +1/5+1/5.

      • 1/7-1/7 cancels with +1/7+1/7.

      • 1/9-1/9 would cancel with the next fraction.

      • 1/(2n+1)1/(2n+1) cancels with the preceding term's end.

    • Remaining Terms: Only the very first fraction and the very last fraction remain in the partial sum formula:         Sn=1312n+3S_n = \frac{1}{3} - \frac{1}{2n+3}

    • Take the Limit:         limn(1312n+3)=130=13\lim_{n \to \infty} (\frac{1}{3} - \frac{1}{2n+3}) = \frac{1}{3} - 0 = \frac{1}{3}

    • Result: The series converges to 13\frac{1}{3}.

Example Study: Decomposing Fractions for Telescoping Series

  • Problem: Evaluate n=144n21\sum_{n=1}^{\infty} \frac{4}{4n^2 - 1}.

  • Analysis: This does not initially look like a telescoping series because it is not two separate fractions. We must create them.

  • Step 1: Factor the denominator using the difference of squares:     4n21=(2n1)(2n+1)4n^2 - 1 = (2n-1)(2n+1)

  • Step 2: Perform Partial Fraction Decomposition (referred to as a "decomposition" or "decomposing the nasty fraction").     4(2n1)(2n+1)=a2n1+b2n+1\frac{4}{(2n-1)(2n+1)} = \frac{a}{2n-1} + \frac{b}{2n+1}

    • Solve for aa and bb: 4=a(2n+1)+b(2n1)4 = a(2n+1) + b(2n-1).

    • Let n=1/2n = 1/2: 4=a(2)    a=24 = a(2) \implies a = 2.

    • Let n=1/2n = -1/2: 4=b(2)    b=24 = b(-2) \implies b = -2.

    • Resulting Series: n=1(22n122n+1)\sum_{n=1}^{\infty} (\frac{2}{2n-1} - \frac{2}{2n+1}).

  • Step 3: List partial sum terms to find the pattern.

    • n=1n=1: 2123\frac{2}{1} - \frac{2}{3}

    • n=2n=2: 2325\frac{2}{3} - \frac{2}{5}

    • n=3n=3: 2527\frac{2}{5} - \frac{2}{7}

    • n=n1n=n-1: 22n322n1\frac{2}{2n-3} - \frac{2}{2n-1}

    • n=nn=n: 22n122n+1\frac{2}{2n-1} - \frac{2}{2n+1}

  • Cancellations: The "insides" of the terms cancel out. Only the first term of the first set (22) and the last term of the nn-th set (22n+1\frac{2}{2n+1}) remain.

  • Final Formula: Sn=222n+1S_n = 2 - \frac{2}{2n+1}.

  • Limit: limnSn=20=2\lim_{n \to \infty} S_n = 2 - 0 = 2.

  • The series converges to 22.

Geometric Series: Definition, Convergence, and Sum Formula

  • Standard Form: A geometric series typically takes the form:     n=1arn1=a+ar+ar2+ar3+\sum_{n=1}^{\infty} a r^{n-1} = a + ar + ar^2 + ar^3 + \dots

    • Alternatively, it can be written starting at zero: n=0arn\sum_{n=0}^{\infty} a r^{n}.

  • Role of Constants:

    • aa is a constant coefficient (the first term). It doesn't affect the fact of convergence/divergence but does affect the final sum.

    • rr is the common ratio (the base).

  • Convergence Criteria:

    • The series converges if and only if r<1|r| < 1 (meaning 1<r<1-1 < r < 1). This happens only with fractions whose absolute value is less than 1. Powers of such fractions decrease toward zero as the exponent increases.

    • If r=1r = 1, the series is a+a+a+a + a + a + \dots, which goes to infinity (with constant aa).

    • The series diverges if r1|r| \ge 1 .

  • The Sum Formula: If a geometric series converges, its sum is given by:     S=a1rS = \frac{a}{1-r}

Proof of the Geometric Series Sum Formula

  • Define the partial sum SnS_n for a geometric series:     Sn=a+ar+ar2++arn1S_n = a + ar + ar^2 + \dots + ar^{n-1}

  • Multiply the entire equation by rr:     rSn=ar+ar2+ar3++arnr S_n = ar + ar^2 + ar^3 + \dots + ar^n

  • Subtract the two equations (SnrSnS_n - r S_n):     Sn(1r)=aarnS_n (1-r) = a - ar^n

  • Solve for SnS_n:     Sn=a(1rn)1rS_n = \frac{a(1 - r^n)}{1 - r}

  • Apply the limit as nn \to \infty:

    • If r<1|r| < 1, then rn0r^n \to 0. This leaves a(10)1r=a1r\frac{a(1-0)}{1-r} = \frac{a}{1-r}.

    • If r1|r| \ge 1, the limit goes to infinity (diverges).

Analyzing Geometric Series: Convergence Criteria and Examples

  • Example 1: n=13(12)n1\sum_{n=1}^{\infty} 3 (\frac{-1}{2})^{n-1}.

    • a=3a = 3.

    • r=1/2r = -1/2.

    • Convergence check: Since 1/2<1|-1/2| < 1, the series converges.

    • Sum: S=31(1/2)=33/2=2S = \frac{3}{1 - (-1/2)} = \frac{3}{3/2} = 2.

  • Warning on Divergence: If r=4/3r = 4/3, the sum formula will still give you a numerical value (a "silly answer"), but it is incorrect because the series is divergent. You must show convergence before calculating the sum.

  • Example 2 (Pattern-based): Sum the alternating series 5/35/9+5/275/81+5/3 - 5/9 + 5/27 - 5/81 + \dots

    • Notice it is starting at n=1n=1, requiring power n1n-1.

    • Factor out the first term to ensure the internal expression starts with 1: 5/3(11/3+1/91/27+)5/3(1 - 1/3 + 1/9 - 1/27 + \dots).

    • Identify common ratio: r=1/3r = -1/3. First term a=5/3a = 5/3.

    • Sum: S=5/31(1/3)=5/34/3=5/4=1.25S = \frac{5/3}{1 - (-1/3)} = \frac{5/3}{4/3} = 5/4 = 1.25.

Strategic Manipulation of Series Indices and Bases

  • Problem: Evaluate n=1en3n+1\sum_{n=1}^{\infty} \frac{e^n}{3^{n+1}}.

  • Manipulation Goal: Match the index (n=1n=1) to the power (n1n-1).

    • Rewrite en=e×en1e^n = e \times e^{n-1}.

    • Rewrite 3n+1=32×3n1=9×3n13^{n+1} = 3^2 \times 3^{n-1} = 9 \times 3^{n-1}.

    • The series becomes: n=1(e9)(e3)n1\sum_{n=1}^{\infty} (\frac{e}{9}) (\frac{e}{3})^{n-1}.

  • Parameters: a=e/9a = e/9 and r=e/3r = e/3.

  • Check Convergence: Since e2.71e \approx 2.71, the ratio 2.71/32.71/3 is less than 1. This is a convergent series.

  • Calculate Sum:     S=e/91e/3=e/9(3e)/3=e9×33e=e3(3e)S = \frac{e/9}{1 - e/3} = \frac{e/9}{(3-e)/3} = \frac{e}{9} \times \frac{3}{3-e} = \frac{e}{3(3-e)}

  • Algebraic Rule: When stripping off fractions to adjust powers, the ratio rr never changes; only the first term aa changes.

The Harmonic Series and Subsequence Properties

  • Definition of Harmonic Series: n=11n=1+12+13+14+\sum_{n=1}^{\infty} \frac{1}{n} = 1 + \frac{1}{2} + \frac{1}{3} + \frac{1}{4} + \dots

  • Convergence status: Despite having terms that decrease toward zero, the harmonic series is divergent.

  • Theorem on Subsequences:

    • If a sequence is convergent, all of its subsequences are convergent.

    • Conversely, if a subsequence is shown to be divergent, the entire sequence must be divergent.

    • The Divergence of the harmonic series is proven using subsequences of its partial sums.

The Divergence Test (n-th Term Test)

  • Conditional Relationship: If a series an\sum a_n converges, then the limit of the individual terms must be zero (limnan=0\lim_{n \to \infty} a_n = 0).

    • Logic: To arrive at a finite sum, eventually you must be adding nothing but zeros at infinity.

  • The Divergence Test: If limnan0\lim_{n \to \infty} a_n \neq 0 or if the limit does not exist, then the series must diverge.

  • Vital Distinction: This is not a bi-conditional statement. Having a limit equal to zero does not necessarily mean the series converges (e.g., the harmonic series terms go to 0, but it diverges).

  • Example Problem: n=12n2+13n21\sum_{n=1}^{\infty} \frac{2n^2 + 1}{3n^2 - 1}.

    • Test: Take the limit of the general term: limn2n2+13n21=2/3\lim_{n \to \infty} \frac{2n^2+1}{3n^2-1} = 2/3.

    • Result: Since 2/302/3 \neq 0, the series diverges by the Divergence Test. At infinity, you are essentially adding 2/3+2/3+2/32/3 + 2/3 + 2/3 \dots which grows without bound.

Algebraic Properties of Convergent Series

  • Scalar Multiple: can=can\sum c a_n = c \sum a_n. You can pull a constant out of the summation.

  • Sum/Difference Rule: (an±bn)=an±bn\sum (a_n \pm b_n) = \sum a_n \pm \sum b_n.

  • Condition: For the sum/difference property to hold for the resulting sum, both individual series parts must be convergent. If one part is divergent, the entire combined series is divergent.

  • Example Problem: n=12n5n3n=(23)n(53)n\sum_{n=1}^{\infty} \frac{2^n - 5^n}{3^n} = \sum (\frac{2}{3})^n - \sum (\frac{5}{3})^n.

    • Right Part: Ratio r=5/3>1r = 5/3 > 1. It is divergent.

    • Conclusion: Since one piece is divergent, the whole series is divergent. There is no need to calculate the sum of the first part.

Advanced Shortcut for Geometric Series Sums: The "First Term" Method

  • Generalizing the sum formula: The sum of a convergent geometric series can always be calculated as:     S=First Term1RatioS = \frac{\text{First Term}}{1 - \text{Ratio}}

  • This bypasses the need to convert every series into the traditional n1n-1 or nn power format.

  • Example Manipulation: n=2en3n+1\sum_{n=2}^{\infty} \frac{e^n}{3^{n+1}}.

    • Rather than shifting indices, identify the very first term by plugging the starting index (n=2n=2) into the expression:

    • First Term (aa) = e232+1=e227\frac{e^2}{3^{2+1}} = \frac{e^2}{27}.

    • The Common Ratio (rr) = e3\frac{e}{3}.

    • Sum = e2/271e/3\frac{e^2/27}{1 - e/3}. This yields the same result with fewer algebraic steps.

  • Crucial Requirement: For this to work, you must be absolutely sure the series is geometric (constant base to an nn power).

Questions & Discussion

  • Question: For the telescoping example, how did the 1/(2n3)1/(2n-3) part get handled?

  • Response: There is a cancellation relationship where the "outsides" or certain specific fractions remain. If you have an nn-th term, and you don't go any further, that term survives if there is no subsequent term to cancel it. We pretend we have a "dot dot dot" representing all internal terms.

  • Question: Why doesn't the limit of terms to zero guarantee convergence?

  • Response: The harmonic series is the primary counter-example. The terms get small enough for the sequence to converge to 0, but the total sum still grows infinitely large. The Divergence Test only identifies divergence; it never confirms convergence.

  • Question: In the index shortcut, do you always just plug in whatever nn equals?

  • Response: Yes. Whatever the starting index is (n=0,n=1,n=2,n=4n=0, n=1, n=2, n=4), you plug that into the expression to find the actual first term used in the sum formula.