Series (mathematics) - Classification and Convergence Tests

Numerical Series and Convergence Tests Introduction A numerical series is an infinite sum of the form $\sum{n=1}^{\infty} an = a1 + a2 + a3 + \cdots$. The central question in studying series is whether this infinite sum converges to a finite value or diverges. To answer this, mathematicians have developed numerous convergence tests that help us determine the behavior of a series without necessarily computing its exact value. Understanding these tests and the special types of series they apply to is essential for calculus. Types of Numerical Series Alternating Series An alternating series has terms that alternate in sign. The general form is: $$\sum{n=1}^{\infty} (-1)^{n}bn \quad \text{or} \quad \sum{n=1}^{\infty} (-1)^{n+1}bn$$ where $bn > 0$. For example, $1 - \frac{1}{2} + \frac{1}{3} - \frac{1}{4} + \cdots$ is alternating. Alternating series are important because they often converge even when their "absolute value" counterpart diverges. This distinction—between convergence and absolute convergence—is crucial in series theory. Telescoping Series A telescoping series is one where consecutive terms cancel, leaving only a few terms in the partial sum. For example: $$\sum{n=1}^{\infty} \left(\frac{1}{n} - \frac{1}{n+1}\right)$$ When you write out the partial sum: $$SN = \left(1 - \frac{1}{2}\right) + \left(\frac{1}{2} - \frac{1}{3}\right) + \left(\frac{1}{3} - \frac{1}{4}\right) + \cdots + \left(\frac{1}{N} - \frac{1}{N+1}\right)$$ Most terms cancel ("telescope"), leaving $SN = 1 - \frac{1}{N+1}$. As $N \to \infty$, we get $S = 1$. Telescoping series are special because their convergence is easy to determine—just simplify the partial sum and take the limit. Arithmetico-Geometric Series <extrainfo> An arithmetico-geometric series combines arithmetic and geometric progressions. Each term equals the product of a term from an arithmetic sequence and a term from a geometric sequence: $$\sum{n=1}^{\infty} n \cdot r^n$$ is a common example. These require special techniques (like multiplying by the common ratio and subtracting) to evaluate. While interesting, they appear less frequently on exams than other series types. </extrainfo> Convergence Tests When faced with a series, you need to determine whether it converges. The following tests provide systematic ways to make this determination. The Term Test (Vanishing Condition) — Use First as a Quick Check Test: If $\displaystyle \lim{n\to\infty} an \neq 0$, then $\sum an$ diverges. Why it works: If the terms don't approach zero, they're too large for the infinite sum to be finite. Important caveat: This test can only prove divergence. If $\lim{n\to\infty} an = 0$, the series might still diverge (it just passes this test). You must use another test. Example: The series $\sum \frac{n}{n+1}$ diverges because $\lim{n\to\infty} \frac{n}{n+1} = 1 \neq 0$. The Integral Test — For Series Related to Easy-to-Integrate Functions Test: Suppose $f(x)$ is positive, continuous, and decreasing for $x \geq 1$, with $f(n) = an$. Then: $$\sum{n=1}^{\infty} an \text{ converges if and only if } \int1^{\infty} f(x)\,dx \text{ converges}$$ Why it works: The series sum and the improper integral are closely related geometrically—the sum can be bounded above and below by rectangles under the curve. When to use: When the series terms look like a function you can integrate (especially $p$-series and variations). Example: For $\sum \frac{1}{n^p}$, we check $\int1^{\infty} \frac{1}{x^p}\,dx$. This converges when $p > 1$ and diverges when $p \leq 1$. The Comparison Tests — When You Know a Similar Series Direct Comparison Test: If $0 \le an \le bn$ for all sufficiently large $n$: If $\sum bn$ converges, then $\sum an$ converges (smaller series converges) If $\sum an$ diverges, then $\sum bn$ diverges (larger series diverges) Limit Comparison Test: If $an, bn > 0$ and $\displaystyle \lim{n\to\infty}\frac{an}{bn} = c$ where $0 < c < \infty$, then $\sum an$ and $\sum bn$ either both converge or both diverge. When to use: When your series resembles a known series (like a $p$-series or geometric series) but isn't quite identical. The limit comparison test is especially powerful because you only need the limit to exist and be positive—not the inequality. Example: For $\sum \frac{2n+1}{3n^2+2n}$, compare with $\sum \frac{1}{n}$. The limit of the ratio is $\frac{2}{3}$, so since $\sum \frac{1}{n}$ diverges, our series also diverges. The Ratio Test — For Series with Factorials or Powers Test: Let $L = \displaystyle \lim{n\to\infty}\left|\frac{a{n+1}}{an}\right|$ If $L < 1$: series converges absolutely If $L > 1$: series diverges If $L = 1$: test is inconclusive Why it works: If each term is significantly smaller than the previous one (ratio less than 1), the sum is finite. If terms grow (ratio greater than 1), the sum diverges. When to use: When your series contains factorials, exponentials, or powers—anything where the ratio simplifies nicely. Example: For $\sum \frac{n!}{2^n}$, we compute $\frac{a{n+1}}{an} = \frac{(n+1)!}{2^{n+1}} \cdot \frac{2^n}{n!} = \frac{n+1}{2}$. As $n \to \infty$, this grows without bound, so $L = \infty > 1$ and the series diverges. The Root Test — Similar to Ratio Test, Especially for $n$-th Powers Test: Let $L = \displaystyle \lim{n\to\infty}\sqrt[n]{|an|}$ If $L < 1$: series converges absolutely If $L > 1$: series diverges If $L = 1$: test is inconclusive When to use: When series terms involve $n$-th powers, like $\left(\frac{n}{2n+1}\right)^n$. Comparison with ratio test: The root test often works when the ratio test fails (when $L = 1$). Some advanced series require the root test specifically. The Alternating Series Test (Leibniz Test) — Specifically for Alternating Series Test: An alternating series $\sum (-1)^n bn$ (where $bn > 0$) converges if: $bn$ is decreasing: $b1 \geq b2 \geq b3 \geq \cdots$ $\displaystyle \lim{n\to\infty} bn = 0$ Why it works: Alternating signs cause partial sums to oscillate around the limit, getting closer each step. Key insight: This test is powerful because alternating series can converge even when the corresponding non-alternating series diverges. For instance, $\sum \frac{(-1)^n}{n}$ converges, but $\sum \frac{1}{n}$ diverges. Example: The series $\sum \frac{(-1)^n}{n}$ converges because $\frac{1}{n}$ is decreasing and $\lim{n\to\infty} \frac{1}{n} = 0$. The Cauchy Condensation Test — For Series with Non-increasing Terms Test: If $an$ is non-increasing and non-negative, then: $$\sum an \text{ converges if and only if } \sum 2^k a{2^k} \text{ converges}$$ When to use: This is specialized for non-increasing sequences. It's particularly useful for determining convergence of $p$-series and related series. <extrainfo> Example: For $\sum \frac{1}{n(\log n)^p}$, the condensation test gives $\sum 2^k \cdot \frac{1}{2^k(\log 2^k)^p} = \sum \frac{1}{(k \log 2)^p}$, which behaves like a $p$-series. </extrainfo> Advanced Tests: Dirichlet's and Abel's Tests <extrainfo> These tests handle more complex situations where series themselves don't converge but products of series do. Dirichlet's Test: If the partial sums of $\sum an$ are bounded and $bn$ is a monotone sequence decreasing to zero, then $\sum an bn$ converges. Abel's Test: If $\sum an$ converges and $bn$ is a bounded monotone sequence, then $\sum an bn$ converges. These are less commonly tested but useful for theoretical work. </extrainfo> Choosing Which Test to Use When faced with a new series: First: Apply the term test as a quick check. If $\lim an \neq 0$, you're done—the series diverges. For series with clear functions: Use the integral test. For series resembling known series: Use comparison or limit comparison tests. For factorials and exponentials: Use the ratio test. For $n$-th powers: Use the root test. For alternating series: Apply the alternating series test directly. Remember: Different tests apply to different series. Developing intuition about which test to try comes with practice.