📝 Alternating series test (37 MCQs)

📖 Explanation: The Alternating Series Test requires two distinct conditions: the limit of the terms must be zero AND the magnitude of the terms must be monotonically non-increasing. A common misconception is assuming the limit condition alone guarantees convergence. However, if the terms do not decrease monotonically (e.g., they oscillate in magnitude while approaching zero), the partial sums may fail to settle toward a specific limit, causing divergence despite the individual terms vanishing.

📖 Explanation: The Alternating Series Test provides sufficient conditions for convergence, not necessary ones. If the monotonicity condition fails, the test cannot be applied to prove convergence, nor does it automatically imply divergence. The series might still converge through other mechanisms or diverge due to the lack of cancellation. Therefore, one must employ alternative tests like the Ratio Test, Root Test, or direct analysis of partial sums to determine the behavior of such series.

📖 Explanation: For a series satisfying the Alternating Series Test, the partial sums oscillate around the true sum $S$. Specifically, $S$ is always trapped between any two consecutive partial sums $s_n$ and $s_{n+1}$. Consequently, the absolute error incurred by stopping at $s_n$ is guaranteed to be less than or equal to the magnitude of the first omitted term, $a_{n+1}$. This property makes alternating series particularly useful for numerical approximation with controlled precision.

📖 Explanation: While the alternating series converges because $\frac{n}{n^2+1}$ decreases to zero, absolute convergence requires $\sum \frac{n}{n^2+1}$ to converge. By Limit Comparison with $\sum \frac{1}{n}$, the series of absolute values diverges since the limit of the ratio is 1. Thus, the original series is only conditionally convergent. Confusing conditional and absolute convergence is a critical error because conditional convergence lacks properties like rearrangement invariance that absolute convergence possesses.

📖 Explanation: In a convergent alternating series, partial sums alternate above and below the limit $S$. If the oscillations cease crossing the axis after a certain point, it implies the limit $S$ itself acts as the boundary preventing further crossings. Since the amplitude decreases, all subsequent partial sums remain on one side of the axis, meaning $S$ must share the sign of those partial sums and lie between them and the axis (or between consecutive partial sums on that side).

📖 Explanation: The standard alternating harmonic series converges very slowly; achieving $k$ decimal places requires roughly $10^k$ terms. In contrast, substituting $x=1/3$ into $2(x + x^3/3 + \dots)$ yields terms involving $(1/9)^k$, resulting in geometric convergence. Each additional term adds approximately two digits of accuracy. This dramatic difference in convergence rate makes the transformed series vastly superior for practical computation, illustrating how algebraic manipulation can optimize numerical methods derived from Taylor expansions.

📖 Explanation: This challenges the misconception that monotonicity is necessary. Consider $b_n = \frac{1}{n} + \frac{(-1)^n}{n^2}$. While $b_n \to 0$, it is not monotone due to the perturbation. However, splitting the series gives $\sum (-1)^n/n$ (convergent) plus $\sum 1/n^2$ (absolutely convergent). Their sum converges despite non-monotonicity. This demonstrates that the Alternating Series Test is sufficient but not necessary, and subtle structural properties can preserve convergence even when simple monotonicity fails.

📖 Explanation: In physical models combining oscillation and damping, convergence often hinges on competing growth rates. Even if coefficients $c_n$ grow polynomially (e.g., $n^p$), the exponential factor $e^{-kn}$ decays much faster. The product $n^p e^{-kn}$ approaches zero monotonically for sufficiently large $n$, satisfying the Alternating Series Test criteria eventually. This illustrates how asymptotic dominance determines convergence in applied contexts, overriding local irregularities or initial growth patterns in the coefficient sequence.

📖 Explanation: The Alternating Series Estimation Theorem provides a worst-case upper bound $|R_n| \leq a_{n+1}$. In practice, if terms decrease rapidly (e.g., factorially or geometrically), the remainder $R_n \approx a_{n+1} - a_{n+2} + \dots$ is significantly smaller than $a_{n+1}$ alone because subsequent terms partially cancel it. Reporting the full bound as uncertainty ignores this internal cancellation within the tail, potentially overstating the imprecision by an order of magnitude or more.

📖 Explanation: Replacing $n$ with $n + \sin(n)$ creates denominators that fluctuate, breaking strict monotonicity since $\sin(n)$ is not monotone. However, the modified term behaves asymptotically like $1/n$ with bounded perturbations. Using Dirichlet's Test or decomposing into $\sum (-1)^{n+1}/n$ plus a correction term reveals convergence persists. Options A and D alter asymptotic behavior or create divergence, while B maintains monotonicity for large n. This tests deep understanding beyond mechanical test application.

📖 Explanation: When discrete monotonicity is hard to establish, extending $f(n)$ to a continuous differentiable function $f(x)$ allows using derivatives. If f&#039;(x) < 0 for $x \geq N$, then $f(x)$ is decreasing on that interval, implying $f(n)$ is decreasing for $n \geq N$. This bridges discrete sequences and continuous calculus, providing a powerful tool when algebraic comparison of successive terms becomes cumbersome. Note that differentiation with respect to discrete n is undefined, making option D invalid.

📖 Explanation: Unlike series with positive terms, alternating series sums depend critically on phase and cancellation patterns, not just term magnitudes. Even if $a_n < b_n$, the interplay of signs means $S_A$ could exceed, equal, or fall below $S_B$. For example, shifting indices or altering initial terms changes sums disproportionately. This highlights that magnitude comparisons of terms do not translate directly to sum comparisons for alternating series, distinguishing them fundamentally from positive-term series theory.

📖 Explanation: Leibniz's series converges extremely slowly (error ~ 1/n). Euler's transform specifically accelerates alternating series by reweighting partial sums based on finite differences, often converting logarithmic/harmonic convergence to geometric convergence without changing the underlying sum. While other transformations exist, Euler's is tailored for this structure. Merely adding terms is inefficient, geometric series don't directly apply, and Integral Test only assesses convergence. This represents advanced numerical analysis knowledge connecting series theory to computational efficiency.

📖 Explanation: Riemann's Rearrangement Theorem states that conditionally convergent series can be rearranged to converge to any real number or diverge. The specific pattern of 2 positives to 1 negative alters the balance of cancellation, shifting the limit to a new value (specifically $\frac{3}{2}\ln 2$ for alternating harmonic). This contrasts sharply with absolutely convergent series where rearrangements preserve sums. Understanding this fragility is crucial for recognizing why absolute convergence is structurally significant beyond mere convergence.

📖 Explanation: The Alternating Series Test requires monotonic decrease only eventually (for $n \geq N$), not from the start. The function $f(x) = \ln x / x$ has derivative $(1-\ln x)/x^2$, which is negative for $x > e$. Thus, terms decrease for $n \geq 3$. Initial non-monotonicity affects only finitely many terms and doesn't impact convergence. This corrects the common misconception that AST demands strict monotonicity from the first term, emphasizing asymptotic behavior over initial transients.

📖 Explanation: At radius boundaries, the ratio/root tests yield ρ=1 (inconclusive). However, substituting $x=R$ often produces an alternating numerical series whose terms decay appropriately. AST leverages this specific structure—alternation plus decay—to establish convergence where magnitude-based tests fail. This synergy explains why endpoint analysis typically shifts from ratio tests to AST or p-series tests. It underscores that convergence at boundaries depends on delicate cancellations invisible to pure magnitude assessments.

📖 Explanation: Halving even terms creates a sawtooth pattern where $b_{2k} = a_{2k}/2 < a_{2k+1} = b_{2k+1}$ potentially, violating monotonicity. Unlike the earlier counterexample where perturbation preserved convergence via decomposition, here the systematic reduction of every other term disrupts the cancellation mechanism essential to AST. The new series may diverge because the imbalance prevents partial sums from stabilizing. This demonstrates that arbitrary modifications to term magnitudes can destroy convergence even when individual terms still vanish, highlighting the delicacy of conditional convergence.

📖 Explanation: Expanding $\sin x/x = 1 - x^2/3! + x^4/5! - \dots$ and integrating yields an alternating series with factorial-denominated terms. Applying AST error bound $|R_n| \leq a_{n+1}$ avoids computing messy derivatives of $\sin x/x$ needed for Lagrange remainder. The factorial decay ensures the first omitted term is an excellent error proxy. This showcases practical advantage: when integration produces alternating series naturally, AST provides computationally simpler and often tighter error control than general Taylor estimates.

📖 Explanation: The raw form $\sqrt{n+1}-\sqrt{n}$ appears problematic for monotonicity assessment. Rationalizing reveals equivalent form $1/(\sqrt{n+1}+\sqrt{n})$, clearly decreasing to 0. The student misidentified term behavior by focusing on components rather than the difference. This exemplifies why algebraic simplification precedes test application: apparent non-monotonicity may be illusory. The series actually converges by AST after proper manipulation, correcting the superficial analysis that mistook component growth for term growth.

📖 Explanation: Mathematical convergence at boundary $r=R$ via AST validates the series representation precisely at that critical point, extending the domain of validity beyond open interval $(-R,R)$. Physically, this means the modeled quantity remains well-defined and finite at the threshold where naive radius tests are inconclusive. This bridges abstract analysis and applied modeling: proving boundary convergence confirms physical predictability at critical parameters, whereas divergence would signal model breakdown or phase transition requiring alternative formulations.

📖 Explanation: AST provides sufficient conditions for convergence, not necessary conditions for divergence. When $\lim a_n \neq 0$, divergence follows from nth Term Test, not AST failure per se. But if monotonicity fails while limit is zero, AST simply doesn't apply—it doesn't declare divergence. Students often mistakenly treat failed AST as divergence proof. Correct approach: use nth Term Test for non-zero limits, other tests for zero-limit non-monotone cases. Understanding test scope prevents logical errors in series classification.

📖 Explanation: Decompose $\sum (-1)^{n+1}/n^2 = \sum 1/n^2 - 2\sum 1/(2k)^2 = \pi^2/6 - 2(1/4)\sum 1/k^2 = \pi^2/6 - \pi^2/12 = \pi^2/12$. This uses relationship between alternating and absolute series: $S_{alt} = S_{abs} - 2S_{even}$. Recognizing this structural link avoids re-summing and leverages known results. It demonstrates how alternating series connect to their absolute counterparts, enabling efficient evaluation through algebraic decomposition rather than brute-force computation or advanced special functions.

📖 Explanation: The student confused monotonicity of sin(x) with monotonicity of composite sin(1/n). Since 1/n is decreasing and sin is increasing on (0,1], the composition sin(1/n) is indeed decreasing. However, the student's reasoning cited sin x increasing as justification, which is backwards logic even though conclusion happens to be correct. Proper reasoning: outer function increasing + inner function decreasing ⇒ composite decreasing. This highlights need for precise chain-rule thinking in sequence monotonicity, not just recalling elementary function behaviors in isolation.

📖 Explanation: Without evaluating the integral, note $f(x)=1/x$ decreasing implies $\int_n^{n+1} f > \int_{n+1}^{n+2} f$, establishing monotonicity directly from integrand properties. Also $0 < a_n < 1/n \to 0$. Thus AST conditions satisfied purely analytically. This demonstrates powerful technique: leveraging integral properties to verify series conditions when closed forms are unavailable or messy. It connects integral calculus and series convergence structurally, avoiding explicit antiderivative computation while rigorously justifying test applicability.

📖 Explanation: Factorial denominators $n!$ grow super-exponentially, far outpacing geometric $r^n$. Thus $1/(n+1)!$ plummets faster than any geometric bound $Cr^n$. For n=10, $1/11! \approx 2.5\times10^{-8}$ vs geometric $(0.5)^{10} \approx 10^{-3}$. This explains why Taylor series for entire functions like $e^x$ achieve machine precision in few terms, while geometric approximations require many more. Understanding decay hierarchy (factorial > exponential > polynomial) is key to selecting efficient approximation methods in numerical analysis.

📖 Explanation: Partial sums oscillate within envelopes defined by cumulative effect of term magnitudes. For well-behaved alternating series, upper/lower bounds of oscillation trace curves related to term sequence $a_n$ and its integral/sum analogs. Specifically, $s_n$ stays between curves reflecting running totals of positive/negative contributions, often approximated by $S \pm a_n$ locally. Visualizing these envelopes helps diagnose convergence rate and error behavior intuitively, connecting discrete partial sum geometry to continuous term decay profiles without heavy computation.

📖 Explanation: Conditional convergence relies on precise cancellation between positive and negative terms. Floating-point rounding introduces tiny errors that don't cancel perfectly, accumulating systematically. For slowly convergent alternating series (e.g., harmonic), millions of terms amplify these errors until they dominate the true sum. Absolutely convergent series are robust because term magnitudes decay fast enough that rounding errors remain bounded. This practical limitation explains why numerically stable algorithms avoid naive summation of conditionally convergent series, preferring accelerated or rearranged forms.

📖 Explanation: Since $a_n$ decreases to 0 and $\ln n$ increases, $a_n/\ln n$ also decreases to 0 (product of decreasing positive and reciprocal of increasing positive). Thus AST applies directly, guaranteeing convergence. Absolute convergence fails since $a_n/\ln n > a_n/n$ potentially and $\sum a_n$ already diverges. So conditional convergence persists. This shows robustness of AST under certain transformations: dividing by slowly growing functions preserves the structural conditions needed for alternating convergence even when absolute convergence is impossible.

📖 Explanation: Student confused numerator growth with term behavior. Actually $n!/n^n \to 0$ super-exponentially (Stirling/Ratio Test: $\rho=1/e$). Terms decrease monotonically for large n. Thus AST readily confirms convergence—in fact absolute convergence holds. The error stems from misjudging asymptotic dominance: denominator $n^n$ overwhelms $n!$. This corrects intuitive but wrong heuristic about factorials, emphasizing rigorous limit/ratio analysis over superficial growth impressions when assessing series term behavior.

📖 Explanation: Although arctan n → π/2, division by n drives term to 0. Monotonicity: derivative of arctan x / x is $\frac{x/(1+x^2) - \arctan x}{x^2} < 0$ for x>0 since arctan x > x/(1+x²). Thus terms decrease to 0, satisfying AST. Key insight: bounded numerator divided by unbounded denominator yields vanishing terms; derivative analysis confirms monotonicity despite non-obvious numerator behavior. This combines limit arithmetic with calculus verification for non-trivial AST application.

📖 Explanation: For standard alternating series starting with positive term ($a_1 > 0$), odd partial sums $s_{2k-1}$ overestimate S while even partial sums $s_{2k}$ underestimate S. Thus $s_n > S$ implies n is odd. This geometric property arises from oscillation pattern: adding positive term overshoots, subtracting undershoots. Recognizing this parity-sum relationship enables error sign prediction and validation of computational results without recomputation, linking visual oscillation behavior to index parity systematically.

📖 Explanation: Even if $a_n \downarrow 0$ and $b_n \to 0$, the product $a_n b_n$ may not be monotone. Example: $a_n = 1/n$, $b_n = (2 + (-1)^n)/n$. Product oscillates in magnitude, violating AST monotonicity. Convergence isn't guaranteed without additional constraints (e.g., $b_n$ monotone or bounded variation). This illustrates that convergence properties aren't multiplicative; structural conditions like monotonicity are fragile under multiplication, requiring careful joint analysis rather than separate verification.

📖 Explanation: With $s_{50} = -0.001$ and error ≤ 0.002, true sum lies in [-0.003, 0.001]. Since interval includes positive values, extinction (S<0) cannot be conclusively determined from this approximation. The uncertainty spans the decision threshold. This demonstrates critical importance of error bounds in applied modeling: numerical proximity to zero isn't sufficient when tolerance exceeds signal magnitude. Practical decisions require either tighter bounds (more terms) or analytical proof of sign, highlighting intersection of numerical analysis and real-world decision-making under uncertainty.

📖 Explanation: Euler transform specifically targets alternating series, recombining terms via binomial weights to accelerate convergence while preserving sum. For alternating harmonic, it dramatically improves rate. Abel/Cesàro/Borel address divergent or summability issues, not acceleration of convergent alternating series. This specialized technique exploits alternating structure uniquely, unlike general summability methods. Knowledge of such transforms represents advanced series manipulation skill beyond standard curriculum, connecting classical analysis to modern computational mathematics for optimizing slowly convergent alternating series encountered in research and high-precision applications.