📝 Taylor series convergence (37 MCQs)

📖 Explanation: This question targets a fundamental misconception in higher-order analysis. Infinite differentiability (smoothness) is a necessary but not sufficient condition for a function to be equal to its Taylor series. A classic counterexample involves functions that are smooth everywhere but whose Taylor series at zero converges only to zero, while the function itself is non-zero elsewhere. This distinction between $C^\infty$ functions and real-analytic functions is crucial; the remainder term $R_n(x)$ must approach zero as $n \to \infty$ for equality to hold, which is not guaranteed solely by the existence of derivatives.

📖 Explanation: This application question tests conceptual understanding of the underlying calculus assumptions. The Maclaurin series for trigonometric functions is derived using differentiation rules like $\frac{d}{dx}\sin x = \cos x$, which are valid only when $x$ is in radians. If degrees were used, the chain rule would introduce factors of $\frac{\pi}{180}$ in every derivative, fundamentally altering the coefficients of the series. Therefore, direct substitution of degree values yields incorrect numerical results because the power series representation is intrinsically linked to the radian definition of the angle, making unit conversion a prerequisite for valid application.

📖 Explanation: This error analysis question requires understanding the behavior of the exponential function's derivatives to optimize error bounds. Since $f^{(n+1)}(x) = e^x$ is strictly increasing, its maximum absolute value on $[-2, 0]$ occurs at $x=0$, giving $M=1$. Using a larger bound like $M=3$ (valid for positive intervals) would result in a gross overestimation of the required number of terms for a specific accuracy. Students must recognize that the tightest possible $M$ depends critically on the specific interval of approximation, and applying generic bounds without analyzing monotonicity leads to inefficient computational strategies in numerical analysis.

📖 Explanation: Graph-based interpretation connects visual divergence to theoretical constraints. When partial sums approximate a function well locally but fail globally, it typically indicates the boundary of the interval of convergence. For real functions defined by power series, the radius of convergence is determined by the distance to the nearest singularity in the complex plane. Even if the function appears smooth on the real line, a complex pole at distance $R$ limits convergence to $|x|<R$. This question tests the deep connection between real-variable graph behavior and complex-analytic properties, moving beyond simple computation to structural understanding of power series representations.

📖 Explanation: This mixed-concept question compares numerical integration techniques. While Simpson’s Rule is powerful, it provides error estimates based on the fourth derivative bound, which can be difficult to compute or loose. For $e^{-x^2}$, the Maclaurin series yields an alternating series after integration. The Alternating Series Estimation Theorem provides a simple, rigorous error bound equal to the first omitted term, eliminating guesswork about partition size. This highlights the strategic advantage of series methods for specific integrands where the resulting series has favorable convergence properties, demonstrating higher-order decision-making in selecting appropriate mathematical tools for precision requirements.

📖 Explanation: This scenario-based question links mathematical truncation to physical modeling assumptions. The first-order model corresponds to $\sin \theta \approx \theta$, valid only for small angles. The full series contains only positive correction terms (e.g., $+k^2/4$), meaning the true period always exceeds the linear approximation for nonzero amplitudes. Understanding this requires recognizing that truncating a Taylor series in a physical context isn't just a mathematical simplification but represents ignoring specific physical nonlinearities. Students must connect the sign and magnitude of omitted terms to qualitative physical behavior, bridging abstract convergence theory with tangible engineering consequences of model fidelity.

📖 Explanation: This challenging question probes endpoint behavior nuances. The series $\sum (-1)^{k+1} x^k/k$ has radius 1. At $x=1$, it becomes $\sum (-1)^{k+1}/k$, the alternating harmonic series, which converges by the Alternating Series Test. At $x=-1$, it becomes $-\sum 1/k$, the harmonic series, which diverges. This asymmetry arises because conditional convergence depends on sign alternation, which is lost at the negative endpoint. Moreover, the singularity of $\ln(1+x)$ at $x=-1$ aligns with the divergence. Students must synthesize knowledge of series tests, function domains, and endpoint analysis to explain asymmetric convergence behavior rigorously.

📖 Explanation: This Olympiad-level question addresses the pathological case of non-analytic smooth functions. While $f^{(n)}(0) = 0$ for all $n$, making the Taylor series identically zero, the function itself is nonzero for all $x \neq 0$. Thus, the series converges everywhere but equals the function only at the expansion point. This demonstrates that infinite differentiability does not imply analyticity. The remainder $R_n(x)$ does not tend to zero for $x \neq 0$. Recognizing this counterexample is essential for deep understanding of Taylor series limitations, distinguishing between formal power series construction and actual functional representation.

📖 Explanation: This direct recall/conceptual question verifies understanding of trigonometric bounds. Since all derivatives of $\cos x$ are $\pm \cos x$ or $\pm \sin x$, and both functions are bounded by 1 in absolute value for all real numbers, $M=1$ is always a valid global bound for real arguments. Distractors exploit misconceptions about oscillation or domain restrictions. However, if the problem involved complex arguments or different functions, $M=1$ might fail. Here, the key insight is recognizing the universal boundedness of trig derivatives on $\mathbb{R}$, making the remainder estimate straightforward compared to unbounded functions like $e^x$.

📖 Explanation: This application question tests understanding of radius of convergence constraints. The binomial series $(1+x)^m$ for non-integer $m$ has radius $R=1$. Setting $x=10$ violates $|x|<1$, causing divergence. To compute $\sqrt{11}$, one must rewrite it as $\sqrt{9(1+2/9)} = 3(1+2/9)^{1/2}$, where $x=2/9 < 1$. This algebraic manipulation to shift the argument into the convergence disk is a critical skill. Students must recognize that series representations have strict domains and that practical computation often requires reformulating the problem rather than brute-force substitution, highlighting the interplay between algebra and analysis.

📖 Explanation: This comparative analysis question evaluates efficiency in series selection. While $\ln(1+1)$ converges, it does so as $\sum (-1)^{n+1}/n$, requiring thousands of terms for modest accuracy. The identity $\ln((1+x)/(1-x)) = 2\sum x^{2k+1}/(2k+1)$ with $x=1/3$ produces terms decaying as $(1/3)^{2k+1}$, achieving high precision with very few terms. This illustrates that mathematical equivalence doesn't imply computational equivalence. Higher-order thinking involves selecting representations optimized for convergence speed, not just validity. Students must understand how variable transformations affect term decay rates and practical utility in numerical approximation.

📖 Explanation: This error analysis question targets confusion between Taylor coefficients and derivative values. The Taylor coefficient $c_k = f^{(k)}(a)/k!$. Given $c_3 = -1$, we have f&#039;&#039;&#039;(a) = -1 \times 3! = -6. Students often mistakenly equate the polynomial coefficient directly to the derivative, neglecting the factorial scaling. This is a common pitfall in reverse-engineering derivatives from series. The question reinforces the precise definition of Taylor polynomials and ensures students can correctly extract derivative information, emphasizing attention to detail in the fundamental relationship between series coefficients and function derivatives.

📖 Explanation: This conceptual question tests understanding of endpoint behavior. The interval of convergence is $(2-3, 2+3) = (-1, 5)$. At the endpoint $x=5$, the ratio/root tests are inconclusive ($\rho=1$). Convergence depends entirely on the specific series; it could be absolute (e.g., $\sum (x-2)^k/k^2$), conditional (e.g., $\sum (-1)^k (x-2)^k/k$), or divergent (e.g., $\sum (x-2)^k/k$). Students must recognize that radius determines interior behavior, but endpoints require individual investigation. This prevents overgeneralization and reinforces the three-part nature of interval determination: interior, endpoints, and synthesis.

📖 Explanation: This advanced concept question addresses subtle issues of uniform convergence and endpoint validity. While integration often improves convergence (e.g., turning divergence into conditional convergence), the resulting series at the endpoint represents $\int f$ only under specific conditions (e.g., Abel’s theorem for continuity). If $f$ has a non-integrable singularity or convergence isn't uniform, the integrated series might converge to something meaningless regarding $\int f$. Students must understand that operational validity at boundaries isn't automatic and requires theoretical justification beyond mere series convergence, highlighting the distinction between formal manipulation and analytical rigor.

📖 Explanation: This direct application question tests basic series manipulation skills. To get $x^2 e^{-x}$, start with $e^u = \sum u^k/k!$, substitute $u=-x$ to get $e^{-x} = \sum (-1)^k x^k/k!$, then multiply by $x^2$ to obtain $\sum (-1)^k x^{k+2}/k!$. This is a fundamental technique for generating new series without recomputing derivatives. Distractors represent common errors like incorrect substitution or confusing differentiation/integration effects. Mastery of algebraic operations on known series is essential for efficient problem-solving, avoiding tedious derivative calculations and leveraging existing knowledge structures.

📖 Explanation: This conceptual question addresses the computational reality of error bounds. Unlike polynomial inequalities, expressions involving $n!$ and $n$ in exponents lack closed-form inverses. Students must recognize that finding minimal $n$ is inherently a discrete search problem, often solved by testing successive integers or using Stirling’s approximation for estimates. This contrasts with solving equations like $x^2 < \epsilon$. Understanding this limitation prevents futile algebraic attempts and promotes practical numerical strategies, linking theoretical bounds to algorithmic implementation in computational mathematics.

📖 Explanation: This application question combines alternating series estimation with physical modeling. For small $\theta > 0$, the Maclaurin series for sine is alternating with decreasing terms. By the Alternating Series Estimation Theorem, truncating after a negative term ($-\theta^3/6$) yields an underestimate, and the error magnitude is bounded by the first omitted term ($\theta^5/120$). This provides immediate qualitative and quantitative error insight without complex remainder analysis. Students must connect series structure (sign pattern) to approximation direction and error bounds, demonstrating integrated understanding of analysis and applied mathematics.

📖 Explanation: This challenging question revisits the critical distinction between convergence and representation. Options A and B are entire functions equal to their series everywhere. Option D has finite radius due to complex poles. Option C is the classic smooth non-analytic function: all derivatives at 0 are zero, so its Maclaurin series is identically zero (infinite radius), but $f(x) \neq 0$ for $x \neq 0$. Thus, the series converges everywhere but represents the function only at one point. This profound counterexample is essential for understanding that Taylor series equality requires more than just derivative existence, testing deep conceptual mastery.

📖 Explanation: This comparative analysis question evaluates error estimation strategies. For alternating series with decreasing terms, the error is bounded by the first neglected term, which is exact for the series structure. Lagrange’s form uses a global derivative bound $M$, which for cosine is 1, but the actual derivative at points near 0 may be much smaller. Thus, the alternating series bound exploits specific series properties for tighter estimates. Students must recognize when specialized tests outperform general ones, optimizing accuracy assessments based on series characteristics rather than defaulting to universal but conservative bounds.

📖 Explanation: This conceptual question tests understanding of differentiation effects on power series. Differentiation preserves radius of convergence $R$ but can alter endpoint behavior. For example, $\sum x^k/k^2$ converges at $x=1$, but its derivative $\sum x^{k-1}/k$ diverges at $x=1$. Conversely, divergence at endpoints may become convergence. The open interval $(a-R, a+R)$ is invariant, but endpoints require re-evaluation. Students must distinguish between radius (always preserved) and interval (possibly changed), avoiding the common misconception that differentiation leaves convergence completely unchanged.

📖 Explanation: This challenging question probes asymptotic analysis of binomial coefficients. Using Stirling/Gamma asymptotics, $\binom{m}{k} \sim k^{-m-1}/\Gamma(-m)$. Thus, terms decay as $k^{-m-1}$. Absolute convergence at $x=1$ requires $-m-1 < -1 \Rightarrow m>0$. For $-1<m<0$, terms alternate and decay slowly, yielding conditional convergence. For $m \leq -1$, terms don't approach zero sufficiently fast, causing divergence. This requires connecting parameter $m$ to p-series behavior via asymptotic analysis, going beyond memorized intervals to understand the mechanistic reason for m-dependent convergence.

📖 Explanation: This practical reasoning question highlights circularity avoidance in numerical analysis. When approximating an unknown constant like $e$, using $e$ itself in the error bound defeats the purpose. Since $e < 3$ is easily established independently, substituting 3 provides a computable, rigorous bound without self-reference. This reflects real-world mathematical practice: bounds must be evaluable without prior knowledge of the target. Students must recognize the logical structure of approximation proofs and the necessity of independent constants in error estimation.

📖 Explanation: This application question tests strategic center selection. $\ln x$ is undefined at 0, precluding Maclaurin series; expanding about $a=1$ for $\ln(1.01)$ leverages proximity and function definition. For $\sin(0.001)$, Maclaurin is ideal. For $e^{10}$, Maclaurin converges slowly; centering near 10 would be better. Students must evaluate function domain, evaluation point proximity, and convergence speed to select optimal expansion centers, demonstrating adaptive problem-solving rather than rote Maclaurin application.

📖 Explanation: This conceptual question tests uniqueness of power series representations. If two power series converge to the same function on an open interval containing 0, their coefficients must be identical. This follows from repeated differentiation and evaluation at 0, or from the identity theorem for analytic functions. This uniqueness is foundational: it justifies finding series via substitution/integration/differentiation rather than direct derivative computation, knowing the result must be THE Taylor series. Students must understand that power series representations are unique, validating indirect derivation methods.

📖 Explanation: This challenging question connects real smoothness to complex singularities. Though $1/(1+x^2)$ is $C^\infty$ on $\mathbb{R}$, its power series radius is determined by distance to nearest complex singularity ($\pm i$), which is 1. Real analysis alone cannot explain this; complex analysis reveals the constraint. This illustrates that real Taylor series behavior is governed by complex analytic structure. Students must transcend real-variable intuition to understand global convergence limitations, recognizing that apparent smoothness on $\mathbb{R}$ doesn't guarantee infinite radius.

📖 Explanation: This application question tests efficient series usage in limits. Expanding $\sin x = x - x^3/6 + O(x^5)$, the numerator becomes $x^3/6 + O(x^5)$. Dividing by $x^3$ gives $1/6 + O(x^2) \to 1/6$. Higher terms contribute $O(x^2)$ which vanishes. Students must identify the dominant balance order and avoid unnecessary computation. This demonstrates strategic truncation: expand only to the order needed to resolve the indeterminate form, balancing precision with efficiency in asymptotic analysis.

📖 Explanation: This optimization question addresses convergence acceleration. Direct $x=-0.5$ in $\ln(1+x)$ converges, but slowly. Transforming to $\ln((1+x)/(1-x))$ with $x=1/3$ for $\ln 2$ yields terms decaying as $(1/3)^{2k+1}$, far faster than $(0.5)^k$. Then negate for $\ln(0.5)$. This exploits series structure for efficiency. Students must recognize that equivalent mathematical expressions have vastly different computational properties, and strategic reformulation is key to practical numerical analysis.

📖 Explanation: This conceptual question tests series addition properties. The sum series converges at least where both converge ($|x|<2$), so radius $\geq 2$. Cancellation of singular terms could extend convergence beyond 2 (e.g., $1/(1-x) + (-1/(1-x)) = 0$, radius $\infty$). Thus, radius is not necessarily preserved exactly. Students must understand that addition preserves the minimum convergence disk but may enlarge it through destructive interference of singular parts, avoiding the misconception that operations preserve radius rigidly.

📖 Explanation: This direct recall question tests differentiability prerequisites. Taylor series at $a$ requires $f$ to be infinitely differentiable at $a$. $\sqrt{x}$ has unbounded derivative at 0, so no Maclaurin series exists. At $x=1$, all derivatives exist, enabling Taylor expansion. This reinforces that series representation demands local smoothness. Students must connect function regularity to series existence, avoiding blind application of formulas to non-differentiable points.

📖 Explanation: This mixed-concept question contrasts representation paradigms. Taylor series are local, polynomial-based, and sensitive to pointwise smoothness. Fourier series are global, trigonometric, and suited for periodic phenomena, with convergence tied to integral properties rather than pointwise derivatives. Confusing them is a common misconception. Students must distinguish local vs. global approximation frameworks, understanding that choice depends on problem structure (local analysis vs. periodic signals), not just preference.

📖 Explanation: This conceptual question tests error sign prediction. For $e^x$ with $x>0$, all Taylor terms are positive. Truncating after $x^3$ omits positive terms, so $p_3(x) < e^x$, making error positive. Unlike alternating series, there's no sign-flipping; error sign matches the first omitted term's sign. Students must analyze series term signs to predict approximation direction, avoiding blind reliance on alternating series intuition for non-alternating cases.

📖 Explanation: This error analysis question targets derivative relationship confusion. Students might misremember $d(\sec x)/dx = \tan x$ (it's $\sec x \tan x$). Correctly, $\tan x = \int \sec^2 x \, dx$, so one should integrate the series for $\sec^2 x$ (obtained by squaring $\sec x$ series or differentiating $\tan x$ series). This tests precise knowledge of trig derivatives and appropriate series operations, preventing flawed derivations based on faulty calculus memory.

📖 Explanation: This challenging question probes endpoint validity justification. Ratio/root tests fail at endpoints. Abel's Theorem bridges the gap: if $\sum a_k R^k$ converges and $f$ is continuous at $x=R$, then $\lim_{x\to R^-} f(x) = \sum a_k R^k$. Since $\ln(1+x)$ is continuous at $x=1$ and the series converges there, equality holds. Students must know specialized theorems for endpoint validation, distinguishing interior convergence (ratio test) from endpoint representation (Abel), completing the convergence analysis framework.

📖 Explanation: This conceptual question tests composition effects on radius. Substituting $x^2$ for $x$ transforms convergence condition $|u|<R$ to $|x^2|<R \Rightarrow |x|<\sqrt{R}$. Thus radius becomes $\sqrt{R}$. For example, $1/(1-x)$ has $R=1$; $1/(1-x^2)$ has $R=1$ (since $\sqrt{1}=1$), but $1/(1-x/4)$ has $R=4$; $1/(1-x^2/4)$ has $R=2=\sqrt{4}$. Students must track how variable transformations scale convergence domains, applying functional composition rules to power series.