Unit 10 · Topic 10.10

The Alternating Series Error Bound

For a series that alternates with sizes decreasing to 0, the sum lies within a_(n + 1) , the first omitted term, of S_n : |S − S_n| ≤ a_(n + 1) .

Practice 1: Implementing Mathematical ProcessesPractice 3: JustificationPractice 4: Communication and Notation

Part 1 · Hook

Why this matters

A partial sum is an estimate of a series' sum. For most series you cannot tell how good the estimate is, but for an alternating series you can, at a glance: the error is never bigger than the first term you left out.

Part 2 · Before you start

What this builds on

Part 3 · Prerequisite check

Quick check before you start

1. Which conditions does the alternating series test need for ∑(−1)nan\sum (-1)^n a_n with an>0a_n > 0?

  1. The ana_n decrease and approach 0
  2. The ana_n approach 0
  3. The ana_n decrease
  4. ∑an\sum a_n converges
Show the answer

Both: decreasing sizes that approach 0.

  • Correct: The ana_n decrease and approach 0:
  • The ana_n approach 0:
  • The ana_n decrease:
  • ∑an\sum a_n converges:

2. What is S3S_3 for ∑n=1∞an\sum_{n=1}^{\infty} a_n?

  1. a1+a2+a3a_1 + a_2 + a_3
  2. a3a_3
  3. a1+a2a_1 + a_2
  4. a3+a4+⋯a_3 + a_4 + \cdots
Show the answer

The third partial sum is the sum of the first three terms.

  • Correct: a1+a2+a3a_1 + a_2 + a_3:
  • a3a_3:
  • a1+a2a_1 + a_2:
  • a3+a4+⋯a_3 + a_4 + \cdots:

Part 4 · See it

See it first

Partial sums of 1 minus 1/4 plus 1/9 minus 1/16 and so on, for N = 1 to 8, zigzagging around a dashed line at pi squared over 12, about 0.822. A shaded band runs from S sub 4 = 115/144, about 0.7986, up to S sub 4 plus the next term, 1/25. The sum lies inside the band: the error of S sub 4 is about 0.0239, less than 1/25.
For ∑(−1)n+1n2\sum \frac{(-1)^{n+1}}{n^2}, the sum lies between S4S_4 and S4+a5S_4 + a_5: the error of S4S_4 is less than a5=125a_5 = \frac{1}{25}. LevlPrep original diagram.

Part 5 · Step by step

How it works, step by step

  1. The partial sums of an alternating series zigzag around the sum with shrinking stepsthe sum always lies between two consecutive partial sums, SnS_n and Sn+1S_{n+1}
  2. Sn+1S_{n+1} and SnS_n differ by the next term, an+1a_{n+1} in sizethe distance from SnS_n to the sum is at most an+1a_{n+1}: |S−Sn|≤an+1|S - S_n| \le a_{n+1}
  3. The sign of the first omitted termtells the direction: if it is negative, SnS_n overestimates the sum; if positive, it underestimates

Part 6 · Key ideas

Key ideas

  • If a series alternates and its sizes decrease to 0, then |S−Sn|≤an+1|S - S_n| \le a_{n+1}: the error is at most the first term left out.
  • State the conditions every time; the bound means nothing without them.
  • Write the result as an inequality: "error ≤1720<0.002\le \frac{1}{720} < 0.002". The error is at most the bound, not equal to it.
  • The sign of the first omitted term says whether SnS_n is too big or too small.

Part 7 · Misconception

A common mistake

The wrong idea: The error of S3S_3 is at most a3a_3, the last term added.

What actually happens: The bound is the first term left out, a4a_4. The last term added is already part of S3S_3.

Part 8 · Check yourself

Check yourself

Exam-style questions. Anything you miss goes into your review queue.

Table

The terms of an alternating series

The series SS=∑n=1∞(−1)n+1n⋅2n= \sum_{n=1}^{\infty} \frac{(-1)^{n+1}}{n \cdot 2^n}=12−18= \frac{1}{2} - \frac{1}{8}+124−⋯+ \frac{1}{24} - \cdots and the sizes of its terms
nn112233445566
an=1n⋅2na_n = \frac{1}{n \cdot 2^n}12\frac{1}{2}18\frac{1}{8}124\frac{1}{24}164\frac{1}{64}1160\frac{1}{160}1384\frac{1}{384}

1. Find S3S_3, the sum of the first three terms.

No calculator. Type the exact value (a fraction, π, e or a root is fine).

Show the answer

S3S_3=12−18= \frac{1}{2} - \frac{1}{8}+124+ \frac{1}{24}=12−3+124= \frac{12 - 3 + 1}{24}=1024= \frac{10}{24}=512= \frac{5}{12}.

  • Answer: 512\frac{5}{12} ≈ 0.417

2. Use the alternating series error bound to find a number EE with |S−S3|≤E|S - S_3| \le E.

No calculator. Type the exact value (a fraction, π, e or a root is fine).

Show the answer

The series alternates and its sizes 1n⋅2n\frac{1}{n \cdot 2^n} decrease to 0, so the alternating series error bound applies: |S−S3|≤a4=164|S - S_3| \le a_4 = \frac{1}{64}.

  • Answer: 164\frac{1}{64} ≈ 0.016

3. Is S3S_3 an overestimate or an underestimate of SS?

  1. An underestimate: the terms keep getting smaller in size
  2. An underestimate: S3S_3 ends with a positive term, 124\frac{1}{24}
  3. Neither: S3S_3 equals SS to two decimal places
  4. An overestimate: the next term, −164-\frac{1}{64}, is negative
Show the answer

The partial sums zigzag around SS, and the next step, −164-\frac{1}{64}, goes down. So S<S3S < S_3: S3S_3 is an overestimate. (Indeed S3≈0.4167S_3 \approx 0.4167 and S≈0.4055S \approx 0.4055.)

  • An underestimate: the terms keep getting smaller in size: Shrinking terms do not decide the direction; the sign of the next term does.
  • An underestimate: S3S_3 ends with a positive term, 124\frac{1}{24}: Ending with a positive term means S3S_3 overshot; the sum is below it.
  • Neither: S3S_3 equals SS to two decimal places: SS lies strictly between S3S_3 and S4S_4, so S3S_3 overshoots it.
  • Correct: An overestimate: the next term, −164-\frac{1}{64}, is negative: Right: the next step would go down, so SS is below S3S_3.

4. Which statement correctly justifies that S4S_4 is within 0.01 of SS?

  1. |S−S4|=1160|S - S_4| = \frac{1}{160}, which is less than 0.01
  2. The series alternates with sizes decreasing to 0, so |S−S4|≤1160<0.01|S - S_4| \le \frac{1}{160} < 0.01
  3. The terms approach 0, so S4S_4 is close to SS
  4. |S−S4|≤164<0.01|S - S_4| \le \frac{1}{64} < 0.01, since 164\frac{1}{64} is the fourth term
Show the answer

The alternating series error bound applies because the series alternates and 1n⋅2n\frac{1}{n \cdot 2^n} decreases to 0. It gives |S−S4|≤a5|S - S_4| \le a_5=1160= \frac{1}{160}=0.00625<0.01= 0.00625 < 0.01.

  • |S−S4|=1160|S - S_4| = \frac{1}{160}, which is less than 0.01: The bound gives "at most", not "equals"; the actual error is smaller than 1160\frac{1}{160}.
  • Correct: The series alternates with sizes decreasing to 0, so |S−S4|≤1160<0.01|S - S_4| \le \frac{1}{160} < 0.01: Right: conditions, the bound, and the comparison with 0.01.
  • The terms approach 0, so S4S_4 is close to SS: Terms approaching 0 do not give a size; the bound a5a_5 with its conditions does.
  • |S−S4|≤164<0.01|S - S_4| \le \frac{1}{64} < 0.01, since 164\frac{1}{64} is the fourth term: The fourth term is used in S4S_4; the bound is the first term left out, 1160\frac{1}{160}.

5. A student says the error in S10S_{10} for ∑n=1∞1n2\sum_{n=1}^{\infty} \frac{1}{n^2} is at most 1121\frac{1}{121}, the first term left out. What is wrong?

  1. The first term left out is 1100\frac{1}{100}, not 1121\frac{1}{121}
  2. The bound should be 1121+1144\frac{1}{121} + \frac{1}{144}, the next two terms
  3. Nothing: the argument is correct as written
  4. The series does not alternate, so this bound does not apply
Show the answer

The alternating series error bound needs alternating signs. Here every omitted term is positive, so the error is 1121+1144+⋯\frac{1}{121} + \frac{1}{144} + \cdots, which is more than 1121\frac{1}{121} (about 0.0950.095).

  • The first term left out is 1100\frac{1}{100}, not 1121\frac{1}{121}: S10S_{10} uses n=1n = 1 to 1010, so the first omitted term is 1121\frac{1}{121}.
  • The bound should be 1121+1144\frac{1}{121} + \frac{1}{144}, the next two terms: Adding two terms is still too small: the error is the sum of the whole tail.
  • Nothing: the argument is correct as written: The bound needs alternating signs; here every omitted term is positive.
  • Correct: The series does not alternate, so this bound does not apply: Right: with positive terms the error is the whole positive tail, larger than one term.

6. The first omitted term of an alternating series is 0.0030.003, and the alternating series error bound applies. Which sentence is correct?

  1. The error is exactly 0.0030.003
  2. The error is at most 0.0030.003
  3. The error is at least 0.0030.003
  4. The error is about 0.00150.0015
Show the answer

The alternating series error bound says |S−Sn|≤0.003|S - S_n| \le 0.003: the error is at most 0.0030.003. Saying it "equals" 0.0030.003 loses the point.

  • The error is exactly 0.0030.003: The actual error is less than or equal to the bound, usually less.
  • Correct: The error is at most 0.0030.003: Right: the bound is a maximum.
  • The error is at least 0.0030.003: The bound is an upper limit, not a lower one.
  • The error is about 0.00150.0015: The bound does not estimate the error; it caps it.

Part 9 · Summary

Summary

For a series that alternates with sizes decreasing to 0, the sum lies within an+1a_{n+1}, the first omitted term, of SnS_n: |S−Sn|≤an+1|S - S_n| \le a_{n+1}. State the conditions, compute the bound, and compare it with the required accuracy using "≤…<\le \ldots <".

Part 10 · Up next

What comes next

Part 11 · Connections

Connections