Integrating power series: 38 Practice MCQs with Answers and Explanation

Q1. A student attempts to find the Maclaurin series for $\ln(1+x)$ by integrating the geometric series for $\frac{1}{1+x}$ . They correctly obtain the series terms but claim the interval of convergence remains exactly $(-1, 1)$ because integration preserves the radius of convergence. What is the flaw in this reasoning?

A.Integration changes the radius of convergence, so the new interval must be recalculated using the ratio test.

B.While the radius of convergence

R=1

is preserved, the behavior at endpoints can change; integration may include an endpoint where the original series diverged. ✅

C.The student made no error; integration always preserves both the radius and the exact interval of convergence including endpoints.

D.Integration only preserves convergence on compact subsets strictly inside the interval, so the new series cannot converge at any endpoint.

💡 Difficulty: hard | ✅ Correct: B

📖 Explanation: This question targets a common misconception regarding endpoint behavior. While Theorem 9.10.4 guarantees that term-by-term integration preserves the radius of convergence $R$ , it does not guarantee identical behavior at the boundary points $x = \pm R$ . For example, the geometric series for $\frac{1}{1+x}$ diverges at $x=1$ , but its integral, the series for $\ln(1+x)$ , converges conditionally at $x=1$ by the Alternating Series Test. Students must distinguish between the open interval of absolute convergence and the potential inclusion of endpoints after integration.

Q2. Consider the function defined by the power series $f(x) = \sum_{n=0}^{\infty} \frac{x^n}{n!}$ . If we define $g(x) = \int_0^x f(t) \, dt$ , which of the following best describes the relationship between the coefficients of $g(x)$ and the derivatives of $g(x)$ at zero?

A.The coefficient of

x^n

in

g(x)

is

\frac{g^{(n)}(0)}{n!}

, confirming

g(x)

is its own Taylor series. ✅

B.The coefficient of

x^n

in

g(x)

corresponds to

\frac{f^{(n-1)}(0)}{n!}

, linking integration to lower-order derivatives of the integrand.

C.Since

f(x)=e^x

,

g(x)=e^x-1

, so the coefficients are simply shifted factorials unrelated to derivative formulas.

D.Integration destroys the Taylor series structure, so coefficients of

g(x)

cannot be expressed via derivatives at zero.

💡 Difficulty: medium | ✅ Correct: A

📖 Explanation: This problem tests the fundamental connection between power series coefficients and Taylor series definitions. When integrating $\sum \frac{x^n}{n!}$ term-by-term, we get $\sum \frac{x^{n+1}}{(n+1)!}$ . Re-indexing shows the coefficient of $x^k$ is $\frac{1}{k!}$ . By definition, the Taylor coefficient is $\frac{g^{(k)}(0)}{k!}$ . Since $g(x) = e^x - 1$ , its derivatives at 0 match these coefficients perfectly. This reinforces that valid power series representations obtained through calculus operations remain the unique Taylor series for the resulting function, validating Theorem 9.10.6 without explicit citation.

Q3. You are modeling a physical system where the response function is $R(x) = \int_0^x \frac{\sin t}{t} \, dt$ . Direct integration is impossible with elementary functions. Using series methods, what is the most efficient strategy to approximate $R(0.5)$ to high precision?

A.Substitute

x=0.5

into the Maclaurin series for

\sin x

and integrate numerically using Simpson's Rule.

B.Expand

\sin t

as a power series, divide by

t

term-by-term, integrate the resulting series term-by-term, then evaluate the alternating sum. ✅

C.Use L'Hôpital's rule to find the limit at zero, then apply standard numerical integration techniques on the original integral.

D.Differentiate

R(x)

repeatedly to find Maclaurin coefficients directly, avoiding series manipulation of the integrand.

💡 Difficulty: easy | ✅ Correct: B

📖 Explanation: Modeling non-elementary integrals like the sine integral requires manipulating series before evaluation. The correct workflow involves expanding $\sin t = \sum (-1)^n t^{2n+1}/(2n+1)!$ , dividing by $t$ to get $\sum (-1)^n t^{2n}/(2n+1)!$ , and integrating term-by-term to obtain $\sum (-1)^n x^{2n+1}/[(2n+1)(2n+1)!]$ . This yields a rapidly converging alternating series perfect for approximation. Option A is inefficient; C ignores the singularity handling; D is computationally prohibitive compared to algebraic series manipulation. This demonstrates the practical utility of power series in solving applied problems where closed-form antiderivatives do not exist.

Q4. Given the graph of a function $y=f(x)$ represented by a power series centered at 0 with radius $R=2$ , and knowing $f(x) > 0$ on $(0,2)$ , what can be definitively inferred about the graph of its antiderivative $F(x) = \int_0^x f(t) dt$ regarding its radius of convergence and monotonicity?

A.

F(x)

has radius

R=2

and is strictly increasing on

(-2, 2)

.

B.

F(x)

has radius

R=2

and is strictly increasing on

[0, 2)

but behavior on

(-2, 0)

depends on sign of

f

. ✅

C.

F(x)

has radius

R > 2

because integration smooths out singularities.

D.

F(x)

has radius

R=2

and is concave up everywhere in the interval.

💡 Difficulty: medium | ✅ Correct: B

📖 Explanation: This question links graphical interpretation with series properties. Integration preserves the radius of convergence $R=2$ , eliminating option C. Since $f(x) > 0$ only on $(0,2)$ , $F'(x)=f(x)$ implies $F$ increases only on that positive sub-interval. We cannot assume $f(x)>0$ on $(-2,0)$ without more information, making A incorrect. Concavity depends on $f'(x)$ , not just positivity of $f$ , eliminating D. Thus, B is the only rigorous conclusion combining analytic radius preservation with graphical derivative-sign analysis, testing careful reading of domain constraints versus series convergence domains.

Q5. A student computes $\int_0^x \sum_{n=0}^\infty (-1)^n t^{2n} \, dt$ and obtains $\sum_{n=0}^\infty \frac{(-1)^n x^{2n}}{2n}$ . Identify the specific mathematical errors in this derivation.

A.The exponent after integration should be

2n+1

and the denominator should be

2n+1

; the student failed to increase the power and adjust the divisor correctly.

B.The student incorrectly integrated

t^{2n}

; the result should be

\frac{t^{2n+1}}{2n+1}

, but they also missed that the series starts at n=0 causing division by zero.

C.Both the integration rule application and the index handling are flawed; correct term is

\frac{(-1)^n x^{2n+1}}{2n+1}

. ✅

D.There is no error; the integration of even powers yields even denominators matching the original exponent pattern.

💡 Difficulty: hard | ✅ Correct: C

📖 Explanation: This item diagnoses procedural fluency gaps. Integrating $t^{2n}$ yields $\frac{t^{2n+1}}{2n+1}$ . The student’s answer $\frac{x^{2n}}{2n}$ contains two critical failures: failing to increment the exponent (power rule violation) and creating a division-by-zero at $n=0$ . Recognizing both errors requires understanding term-by-term integration mechanics AND checking boundary cases ( $n=0$ ). This goes beyond simple computation to validate the entire expression structurally. Such multi-error identification is crucial for debugging computational work in series-based modeling where automated tools might obscure foundational mistakes.

Q6. Suppose $\sum_{n=0}^\infty a_n x^n$ converges to $f(x)$ on $(-R, R)$ . Which statement best characterizes the validity of $\int_a^b f(x) dx = \sum_{n=0}^\infty a_n \frac{b^{n+1}-a^{n+1}}{n+1}$ when $a, b \in (-R, R)$ ?

A.It is always valid regardless of endpoint convergence because definite integration over a compact subinterval within the radius of convergence guarantees uniform convergence. ✅

B.It is valid only if the series converges absolutely at both

a

and

b

.

C.It is valid only if

f(x)

is continuous at the endpoints

a

and

b

.

D.It requires the series to be uniformly convergent on the entire interval

[-R, R]

, not just

(-R, R)

.

💡 Difficulty: medium | ✅ Correct: A

📖 Explanation: This probes deep theoretical conditions for term-by-term definite integration. Within the open interval of convergence $(-R, R)$ , a power series converges uniformly on any compact subset $[a,b] \subset (-R,R)$ . Uniform convergence is sufficient to interchange summation and integration. Endpoint behavior or absolute convergence at specific points is irrelevant as long as limits stay strictly inside the radius. Options B, C, and D impose unnecessarily strong or misdirected conditions. Understanding this distinction prevents students from erroneously testing endpoints when evaluating definite integrals well within the convergence disk, streamlining applied calculations.

Q7. To evaluate $\int_0^{0.1} e^{-x^2} dx$ accurately, one substitutes the Maclaurin series for $e^u$ with $u=-x^2$ . Why is this approach superior to substituting $u=-x^2$ into the already-integrated series for $\int e^u du$ ?

A.Because

\int e^u du = e^u + C

returns the original function, offering no simplification; integrating the expanded polynomial form yields a computable numeric series. ✅

B.Because the chain rule prevents direct substitution into the antiderivative series without adjusting for

du/dx

.

C.Because the series for

e^{-x^2}

converges faster than the series for

e^u

.

D.There is no superiority; both methods yield identical computational complexity and accuracy.

💡 Difficulty: hard | ✅ Correct: A

📖 Explanation: This contrasts functional composition with series manipulation. The antiderivative of $e^u$ is just $e^u$ , which brings us back to the non-elementary $e^{-x^2}$ upon substitution—no progress. Expanding first gives $\sum (-1)^n x^{2n}/n!$ , whose term-by-term integral $\sum (-1)^n x^{2n+1}/[(2n+1)n!]$ provides a concrete polynomial approximation. This highlights that series methods unlock tractability precisely when closed-form antiderivatives fail. It integrates knowledge of substitution rules, series expansion, and the definition of non-elementary integrals, emphasizing strategic sequencing in problem-solving rather than rote procedure.

Q8. If $f(x) = \sum_{n=1}^\infty \frac{x^n}{n^2}$ has radius of convergence $R=1$ , and $F(x) = \int_0^x f(t) dt$ , what is the behavior of $F(x)$ at $x=1$ ?

A.

F(1)

diverges because integrating increases the growth rate of coefficients.

B.

F(1)

converges absolutely because the integrated series has terms proportional to

1/n^3

. ✅

C.

F(1)

converges conditionally because the integrated series becomes alternating.

D.The convergence at

x=1

cannot be determined without testing the differentiated series.

💡 Difficulty: hard | ✅ Correct: B

📖 Explanation: Integrating $\sum x^n/n^2$ yields $\sum x^{n+1}/[n^2(n+1)]$ . At $x=1$ , terms behave asymptotically as $1/n^3$ . Since $p=3 > 1$ , this p-series converges absolutely. Original series converged absolutely at $x=1$ ( $p=2$ ); integration improved convergence further. This challenges students to track coefficient transformations through integration and apply comparison/p-series tests to the NEW series, not the old one. Many mistakenly think integration doesn't affect convergence rate or confuse it with differentiation (which worsens convergence). Tracking the exponent shift from $n^2$ to $n^2(n+1) \approx n^3$ is key.

Q9. When finding the series for $\arctan(x)$ by integrating $\frac{1}{1+x^2}$ , why must we explicitly verify the constant of integration $C$ even though term-by-term integration produces a series with zero constant term?

A.Because indefinite integration always introduces an arbitrary constant that could be nonzero depending on the lower limit chosen.

B.Because the geometric series substitution assumes

|x|<1

, potentially introducing discontinuities that affect the constant.

C.Because the derived series represents

\arctan(x) - \arctan(0)

; since

\arctan(0)=0

,

C=0

here, but this verification is mandatory for correctness in general. ✅

D.Because term-by-term integration is only valid up to an additive constant that must be determined by boundary conditions.

💡 Difficulty: medium | ✅ Correct: C

📖 Explanation: While seemingly basic, this addresses a subtle necessity. Term-by-term integration of $\sum (-1)^n x^{2n}$ yields $\sum (-1)^n x^{2n+1}/(2n+1) + C$ . This series equals $\arctan(x)$ only if $C$ matches the initial condition. Since $\arctan(0)=0$ and the series evaluates to 0 at $x=0$ , $C=0$ . However, if integrating to find $\ln(1+x)$ from $1/(1+x)$ , same logic applies. Skipping this step risks representing $f(x)+C$ instead of $f(x)$ . This reinforces that series represent specific functions anchored by initial values, not just formal antiderivatives.

Q10. A physics model requires evaluating $\int_0^x \frac{1-\cos t}{t^2} dt$ . Direct substitution fails at $t=0$ . How does series integration resolve this apparent singularity?

A.The series expansion of

1-\cos t

begins with

t^2/2

, canceling the denominator's

t^2

and yielding a removable singularity with a well-defined integrable series. ✅

B.L'Hôpital's rule must be applied first to remove the singularity before series expansion is possible.

C.The integral diverges at zero regardless of series methods; the model is physically invalid.

D.Series integration automatically ignores singularities, providing a regularized value without mathematical justification.

💡 Difficulty: easy | ✅ Correct: A

📖 Explanation: This showcases series as a tool for handling indeterminate forms in modeling. Expanding $\cos t = 1 - t^2/2! + t^4/4! - \cdots$ gives $1-\cos t = t^2/2! - t^4/4! + \cdots$ . Dividing by $t^2$ yields $1/2! - t^2/4! + \cdots$ , a perfectly analytic function at 0. Integrating term-by-term gives $x/2! - x^3/(3\cdot4!) + \cdots$ . This resolves the 0/0 form naturally through algebraic cancellation in the series domain, demonstrating how power series extend function definitions analytically across removable singularities—a critical skill in applied mathematics and physics where naive evaluation fails.

Q11. Compare two methods for approximating $\int_0^{0.5} \ln(1+x) dx$ : (I) Numerical quadrature on $\ln(1+x)$ ; (II) Integrating the Maclaurin series for $\ln(1+x)$ term-by-term then summing. Under what circumstance is Method II distinctly advantageous?

A.Method II is always superior due to exact arithmetic.

B.Method II excels when symbolic antiderivatives are unavailable or when integrating over domains where the series converges rapidly and high precision is needed without adaptive meshing. ✅

C.Method I is always better because transcendental functions are optimized in hardware.

D.Method II is only useful for theoretical proofs, never computation.

💡 Difficulty: medium | ✅ Correct: B

📖 Explanation: This evaluates strategic method selection. While modern CAS handle both, series integration shines when: (1) no elementary antiderivative exists (not the case here, but generally); (2) extreme precision is required beyond floating-point quadrature limits; (3) the argument is small ensuring rapid convergence. For $x=0.5$ , the alternating series converges quickly with predictable error bounds via AST. Quadrature may struggle with endpoint singularities or require many evaluations. Understanding trade-offs between analytical series truncation error vs. numerical discretization error is higher-order thinking essential for scientific computing, moving beyond 'plug-and-chug' to informed algorithmic choice.

Q12. Given $f(x) = \sum_{n=0}^\infty c_n x^n$ with radius $R$ , and $F(x) = \int_0^x f(t) dt$ , which inequality correctly bounds the error when approximating $F(x)$ by the partial sum $S_N(x) = \sum_{n=0}^N c_n \frac{x^{n+1}}{n+1}$ for $|x| < R$ ?

A.

|F(x) - S_N(x)| \leq \left| \int_0^x R_{N}(t) dt \right|

, where

R_N

is the remainder of

f

's series.

B.

|F(x) - S_N(x)| \leq \max|R_N(t)| \cdot |x|

for

t

between 0 and

x

.

C.Both A and B are valid error bounding strategies derived from integrating the remainder term. ✅

D.Error cannot be bounded without knowing the closed form of

F(x)

.

💡 Difficulty: medium | ✅ Correct: C

📖 Explanation: Approximating integrated series involves bounding $\int_0^x (f(t)-S_N^f(t)) dt = \int_0^x R_N^f(t) dt$ . By triangle inequality, $|\int R| \leq \int |R| \leq \max|R| \cdot |x|$ . Both formulations are mathematically sound: A is exact integral of remainder; B is a practical upper bound using max remainder estimate. This connects series remainder theory with integral estimation, showing multiple valid pathways to error control. Students often memorize single formulas; recognizing equivalence and applicability of different bounds reflects deeper understanding of analysis foundations underlying computational reliability.

Q13. Why can't we find the Maclaurin series for $\tan x$ by directly integrating the series for $\sec^2 x$ , despite $\tan x = \int \sec^2 x dx$ ?

A.Because the Maclaurin series for

\sec^2 x

is not commonly memorized and deriving it requires squaring the secant series, which is algebraically complex. ✅

B.Because

\sec^2 x

has poles within the radius of convergence, preventing term-by-term integration.

C.Because integration of

\sec^2 x

's series yields

\tan x + C

, but determining

C

is impossible.

D.Direct integration works perfectly; there is no obstacle.

💡 Difficulty: medium | ✅ Correct: A

📖 Explanation: While theoretically valid, practically deriving $\sec^2 x$ 's series requires multiplying $\sec x$ series by itself or differentiating $\tan x$ —circular or tedious. Standard approach uses $\tan x = \sin x / \cos x$ via division or implicit differentiation. This question tests meta-cognitive awareness: knowing WHEN a valid method is impractical. Students might select B (incorrect—poles are outside MAclaurin radius) or D (theoretically true but misleading). Recognizing computational feasibility versus theoretical possibility is advanced problem-solving skill distinguishing proficient users from rote learners who apply first available formula without efficiency consideration.

Q14. In modeling heat diffusion, you encounter $\int_0^x e^{-t^2} dt$ . After obtaining the series $\sum_{n=0}^\infty \frac{(-1)^n x^{2n+1}}{n!(2n+1)}$ , you need 5-decimal accuracy at $x=1$ . How do you determine the minimum number of terms WITHOUT computing all partial sums?

A.Use the Alternating Series Estimation Theorem: find smallest

N

where

\frac{1}{N!(2N+1)} < 0.5 \times 10^{-5}

. ✅

B.Use Lagrange Error Bound on the original function

e^{-x^2}

before integration.

C.Compute partial sums until consecutive differences are below tolerance; no shortcut exists.

D.Apply Ratio Test to find convergence speed, then estimate terms needed.

💡 Difficulty: easy | ✅ Correct: A

📖 Explanation: For alternating series satisfying AST hypotheses (decreasing magnitude, limit zero), error after $N$ terms is bounded by magnitude of FIRST OMITTED TERM. Here, terms alternate and decrease for $x=1$ . So solve $\frac{1}{(N+1)!(2(N+1)+1)} < 5\times10^{-6}$ . This avoids brute-force summation. Option B applies to pre-integrated function incorrectly; C is inefficient; D gives radius info, not truncation error. Mastering AST for integrated series dramatically accelerates engineering computations. This tests recognition that integrated alternating series retain AST applicability and leverages structural properties for efficient error control—a hallmark of expert computational practice.

Q15. Consider $\int_0^x \frac{dt}{1+t^4}$ . The geometric series substitution requires $|t^4| < 1$ . What happens if you attempt to use this integrated series to approximate the integral at $x=1.5$ ?

A.The series diverges because

1.5^4 > 1

, violating the convergence condition; alternative methods like analytic continuation or different expansions are needed. ✅

B.The series still converges because integration extends the radius of convergence beyond the original geometric series limit.

C.The series converges conditionally at

x=1.5

due to alternating signs.

D.The approximation will be inaccurate but finite; increasing terms improves accuracy eventually.

💡 Difficulty: medium | ✅ Correct: A

📖 Explanation: Geometric series $\sum (-t^4)^n$ converges ONLY for $|t|<1$ . Integration preserves radius $R=1$ ; it does NOT extend it. At $x=1.5 > R$ , the series diverges fundamentally—no amount of terms helps. This counters the persistent myth that 'integration improves convergence domain'. While endpoint inclusion may occur AT $R$ , beyond $R$ divergence is absolute. Students must recognize domain limitations persist post-integration and know when to switch strategies (e.g., expand about different center, use asymptotic expansions). This prevents catastrophic modeling errors from blind series application outside validity regions.

Q16. You are given graphs of $f(x)$ and four candidate curves labeled A-D. One curve represents $F(x)=\int_0^x f(t)dt$ obtained via series integration. If $f(x)$ 's series has only odd powers with positive coefficients, which graph MUST represent $F(x)$ ?

A.An even function passing through origin with positive curvature near zero. ✅

B.An odd function passing through origin with negative slope at zero.

C.An even function with y-intercept equal to first coefficient of f.

D.A function symmetric about y-axis with maximum at origin.

💡 Difficulty: medium | ✅ Correct: A

📖 Explanation: If $f(x) = \sum a_n x^{2n+1}$ (odd powers only), then $F(x) = \sum a_n \frac{x^{2n+2}}{2n+2}$ contains ONLY EVEN powers. Even-powered series define EVEN functions: $F(-x)=F(x)$ . Also $F(0)=0$ (integral from 0 to 0). Near zero, leading term dominates: $a_0 x^2/2$ . If $a_0>0$ , parabola opens upward (positive curvature). Thus F is even, passes through origin, concave up near 0. This links series parity to graphical symmetry without computation. Students must translate algebraic series structure into geometric function properties—a sophisticated cross-representational reasoning skill essential for validating computational results visually.

Q17. When integrating $\sum_{n=0}^\infty x^n$ term-by-term to get $-\ln(1-x)$ , why is the constant of integration necessarily zero specifically when using the definite integral form $\int_0^x$ ?

A.Because the lower limit 0 makes every integrated term vanish at the boundary, inherently satisfying

F(0)=0

without adding C. ✅

B.Because

\ln(1-0) = 0

forces C=0 regardless of integration limits.

C.Because indefinite integration always sets C=0 for power series centered at 0.

D.The constant is not necessarily zero; it depends on the branch of logarithm chosen.

💡 Difficulty: easy | ✅ Correct: A

📖 Explanation: Definite integral $\int_0^x t^n dt = \frac{x^{n+1}}{n+1} - \frac{0^{n+1}}{n+1} = \frac{x^{n+1}}{n+1}$ . Summing gives series with zero constant term automatically. This matches $-\ln(1-x)$ since $-\ln(1-0)=0$ . Indefinite integration would require solving for C separately. Using definite integral from center eliminates constant ambiguity entirely. This subtle point explains why textbooks prefer definite integrals for series-derived functions: it builds the initial condition directly into the operation. Confusing indefinite/definite approaches leads to erroneous constants. Precision in integral formulation reflects mature understanding of series-function correspondence.

Q18. A student claims: 'Since $\int_0^x f(t)dt$ converges at $x=R$ , the original series $f(x)$ must also converge at $x=R$ .' Provide a counterexample disproving this.

A.

f(x) = \sum (-1)^n x^n

diverges at

x=1

, but

\int_0^1 f(t)dt = \ln 2

converges. ✅

B.

f(x) = \sum x^n/n

diverges at

x=1

, but its integral converges.

C.No counterexample exists; the statement is actually a theorem.

D.

f(x) = \sum (-1)^n x^n/n

converges at

x=1

, but its integral diverges.

💡 Difficulty: hard | ✅ Correct: A

📖 Explanation: Classic counterexample: geometric series $\sum (-x)^n$ diverges at $x=1$ (oscillates), yet its integral $\sum (-1)^n/(n+1)$ is alternating harmonic series converging to $\ln 2$ . Integration can REGULARIZE divergence at endpoints by adding denominator factors that enable conditional convergence. Student’s claim reverses causality incorrectly. Constructing/disproving such claims requires deep grasp of endpoint behavior asymmetry under calculus operations. This HOTS item moves beyond verification to falsification, demanding students access repertoire of pathological examples to test generalizations—a core mathematical reasoning skill distinguishing advanced learners.

Q19. In approximating $\int_0^{0.2} \frac{e^x - 1}{x} dx$ , why is series integration preferred over applying L'Hôpital's rule followed by numerical integration?

A.L'Hôpital resolves the 0/0 form but yields

e^x/x

which STILL has no elementary antiderivative; series handles both singularity removal and integration simultaneously. ✅

B.L'Hôpital's rule cannot be applied to integrands, only limits.

C.Numerical integration fails near singularities regardless of L'Hôpital preprocessing.

D.Series integration is slower but more accurate; L'Hôpital is faster but less precise.

💡 Difficulty: hard | ✅ Correct: A

📖 Explanation: Applying L'Hôpital to $(e^x-1)/x$ as $x\to0$ confirms limit=1, removing singularity POINTWISE. But the FUNCTION $e^x/x$ still lacks elementary antiderivative! Numerical methods could now work, but series expansion $\sum x^n/(n+1)!$ integrates to $\sum x^{n+1}/[(n+1)(n+1)!]$ , giving analytic expression PLUS automatic singularity resolution PLUS easy error bounds. Series unifies three tasks: regularization, antidifferentiation, and approximation. Recognizing when partial fixes (L'Hôpital) are insufficient versus comprehensive solutions (series) exemplifies expert problem decomposition. This transcends technique application to strategic methodology selection based on holistic problem structure analysis.

Q20. Given $F(x) = \int_0^x \sum_{n=0}^\infty a_n t^n dt$ , which condition ensures $F''(x)$ exists and equals $\sum_{n=1}^\infty n a_n x^{n-1}$ on $(-R,R)$ ?

A.Only that the original series has radius

R>0

; power series are infinitely differentiable within their interval of convergence. ✅

B.That the differentiated series

\sum n a_n x^{n-1}

must independently pass the ratio test.

C.That

F(x)

must be continuous at the endpoints

\pm R

.

D.That

a_n

must decay faster than

1/n^2

.

💡 Difficulty: medium | ✅ Correct: A

📖 Explanation: Fundamental theorem: Within open interval of convergence $(-R,R)$ , power series define smooth ( $C^\infty$ ) functions. Term-by-term differentiation/integration is always valid INSIDE the radius. No additional decay conditions or separate convergence tests needed for interior points. Endpoint behavior is separate issue. This question tests whether students understand that radius of convergence defines a DOMAIN OF ANALYTICITY where all calculus operations are freely permitted. Misconceptions about needing extra conditions reflect incomplete internalization of power series' exceptional regularity properties—a cornerstone concept for advanced analysis and differential equations.

Q21. You compute $\int_0^x \arctan(t) dt$ by integrating the arctan series. Your result includes $x^2/2$ as the leading term. A peer argues the leading term should be $x^3/3$ because arctan starts with $x$ . Who is correct and why?

A.You are correct; integrating

t

gives

t^2/2

, so leading term is

x^2/2

. Peer confused integration with multiplication. ✅

B.Peer is correct; arctan series starts with

x

, so integral must start with cubic term.

C.Neither; the constant term dominates.

D.You are correct, but only because arctan(0)=0; otherwise constant term would lead.

💡 Difficulty: easy | ✅ Correct: A

📖 Explanation: Arctan series: $t - t^3/3 + \cdots$ . Integrating term-by-term: $\int_0^x t dt = x^2/2$ , $\int -t^3/3 dt = -x^4/12$ , etc. Leading term IS $x^2/2$ . Peer mistakenly thinks integrating $x$ gives $x^3/3$ (confusing with $\int x^2$ ) or conflates series order. This diagnostic question tests basic integration mechanics within series context AND ability to articulate WHY a peer's intuition failed. Explaining errors reinforces correct mental models more than solitary computation. Social dimension of error analysis mirrors real collaborative STEM environments.

Q22. For the series $\sum_{n=0}^\infty \frac{(-1)^n x^{2n+1}}{(2n+1)!}$ representing $\sin x$ , what is the radius of convergence of its integrated series $\sum_{n=0}^\infty \frac{(-1)^n x^{2n+2}}{(2n+1)!(2n+2)}$ ?

A.Infinite, same as original sine series. ✅

B.Finite, reduced by factor of 2 due to extra factorial in denominator.

C.Infinite, but only for even powers of x.

D.Cannot be determined without ratio test on new series.

💡 Difficulty: medium | ✅ Correct: A

📖 Explanation: Original sin series has $R=\infty$ . Integration PRESERVES radius of convergence. New series also has $R=\infty$ . Extra factorial in denominator actually IMPROVES convergence rate but doesn't change infinite radius. Option B reflects confusion with differentiation (which can reduce R at boundaries, though not for entire functions). Option C misunderstands that radius concerns |x|, not parity. Option D ignores preservation theorem. This reinforces that for entire functions (sin, cos, exp), all derived/integrated series remain entire. Internalizing this prevents unnecessary re-testing and builds confidence in manipulating fundamental series families common in modeling.

Q23. In a biological growth model, population $P(t) = \int_0^t r(s) ds$ where rate $r(s) = \sum a_n s^n$ . If measurements show $P(t)$ saturates at finite value as $t \to \infty$ , what does this imply about the series representation's validity?

A.The power series model is only valid locally; global saturation behavior cannot be captured by a single power series with infinite or large finite radius unless it represents a bounded entire function (constant). ✅

B.The coefficients

a_n

must alternate in sign to produce saturation.

C.The radius of convergence must be exactly equal to the saturation time.

D.Power series cannot model saturation; exponential functions must be used instead.

💡 Difficulty: medium | ✅ Correct: A

📖 Explanation: Power series with infinite radius representing non-constant functions are unbounded as $t\to\infty$ (Liouville-type reasoning). Saturation implies boundedness, incompatible with non-constant entire power series. Thus, either: (1) series has finite radius and model breaks down before saturation; (2) series approximates local behavior only; (3) different global representation needed. This exposes fundamental LIMITATION of power series for global phenomena despite local utility. Students often overextend series beyond validity. Recognizing when mathematical structure contradicts observed behavior is critical modeling competency—knowing when NOT to use a tool is as important as knowing how to use it.

Q24. When integrating $\frac{1}{1+x} = \sum (-1)^n x^n$ to get $\ln(1+x)$ , the series converges at $x=1$ but diverges at $x=-1$ . Why does integration 'fix' convergence at $x=1$ but not $x=-1$ ?

A.At

x=1

, integrated series becomes alternating harmonic (converges); at

x=-1

, it becomes negative harmonic series (diverges). Sign alternation enables conditional convergence only at right endpoint. ✅

B.Integration always fixes right endpoint but never left endpoint due to orientation.

C.The radius increases to include 1 but excludes -1 asymmetrically.

D.Both endpoints actually converge after integration; the premise is false.

💡 Difficulty: medium | ✅ Correct: A

📖 Explanation: Original $\sum (-1)^n x^n$ diverges at BOTH ±1. Integrated $\sum (-1)^n x^{n+1}/(n+1)$ : at x=1 → $\sum (-1)^n/(n+1)$ (alternating harmonic, converges); at x=-1 → $\sum (-1)^n (-1)^{n+1}/(n+1) = \sum -1/(n+1)$ (harmonic, diverges). Alternation at x=1 enables AST; same-sign at x=-1 causes divergence. Integration adds 1/n factor enabling conditional convergence ONLY when signs alternate. This nuanced endpoint analysis reveals why intervals of convergence for integrated series often become half-open [−R,R) or (−R,R]. Understanding sign-dependent convergence mechanisms is essential for precise domain specification in applications.

Q25. You approximate $\int_0^{0.1} \cos(x^2) dx$ using series. After integrating, you get an alternating series. To guarantee error < $10^{-8}$ , you check the first omitted term. What additional verification is REQUIRED before trusting this error bound?

A.Confirm the terms are decreasing in magnitude AND approaching zero; AST error bound only applies if both conditions hold. ✅

B.Verify the series is absolutely convergent.

C.Check that x=0.1 is within radius of convergence.

D.Ensure no rounding errors occurred in coefficient calculation.

💡 Difficulty: hard | ✅ Correct: A

📖 Explanation: AST error bound $|R_N| \leq a_{N+1}$ REQUIRES: (1) alternating signs, (2) $|a_{n+1}| \leq |a_n|$ , (3) $\lim a_n = 0$ . For small x=0.1, terms decrease rapidly, but this MUST be verified, not assumed. If terms initially increased (possible for larger x), AST wouldn't apply yet. Blindly applying first-omitted-term bound without checking monotonicity is dangerous. This question enforces disciplined hypothesis verification before error estimation—a critical numerical analysis habit. Many students memorize 'error ≤ next term' without internalizing preconditions, leading to false confidence in inaccurate results.

Q26. Which transformation converts $\int_0^x \frac{dt}{\sqrt{1-t^2}}$ into a binomial series integration problem?

A.Substitute

u=t^2

and expand

(1-u)^{-1/2}

using binomial theorem with m=-1/2. ✅

B.Recognize integrand as derivative of arcsin and use known series.

C.Expand

\sqrt{1-t^2}

directly as geometric series.

D.Use partial fractions decomposition before series expansion.

💡 Difficulty: easy | ✅ Correct: A

📖 Explanation: Integrand is $(1-t^2)^{-1/2}$ . Binomial series $(1+u)^m = \sum \binom{m}{k} u^k$ with $u=-t^2, m=-1/2$ gives $\sum \binom{-1/2}{k} (-1)^k t^{2k}$ . Integrate term-by-term: $\sum \binom{-1/2}{k} (-1)^k \frac{x^{2k+1}}{2k+1}$ . This yields arcsin series systematically. Option B bypasses integration; C misidentifies form; D inapplicable. Recognizing binomial structure in radical expressions is key technique for generating series for inverse trig/hyperbolic functions. This tests pattern matching between integrand form and generalized binomial template—a transferable skill for diverse non-elementary integrals encountered in physics and engineering.

Q27. If $\sum a_n x^n$ has radius R=3, what is the radius of convergence for $\int_0^x \sum a_n t^n dt$ ?

A.Exactly 3; integration preserves radius of convergence. ✅

B.Greater than 3; integration improves convergence.

C.Less than 3; integration introduces additional constraints.

D.Cannot be determined without knowing coefficients.

💡 Difficulty: easy | ✅ Correct: A

📖 Explanation: Fundamental theorem: Term-by-term integration of power series preserves radius of convergence R. Interval of convergence may differ at endpoints, but radius remains identical. This is non-negotiable property stemming from root/ratio test behavior under 1/n factor multiplication. Students sometimes confuse radius (determined by coefficient growth rate) with interval (endpoint-sensitive). Reinforcing this invariant prevents unnecessary recalculation and builds reliable mental framework for series manipulation. Quick recall of this property enables efficient problem-solving flow without getting bogged down in redundant convergence testing.

Q28. In evaluating $\int_0^1 \frac{\ln(1+x)}{x} dx$ via series, you integrate $\sum (-1)^{n-1} x^{n-1}/n$ . Why is swapping summation and integration justified HERE specifically?

A.The series converges uniformly on [0,1] (by Abel's theorem for power series at endpoint), permitting term-by-term integration over closed interval. ✅

B.Uniform convergence holds only on [0,1), so swap is invalid at x=1.

C.The integrand is continuous on [0,1], which alone justifies the swap.

D.Swapping is never justified for conditionally convergent series.

💡 Difficulty: medium | ✅ Correct: A

📖 Explanation: Series $\sum (-1)^{n-1} x^{n-1}/n$ converges to $\ln(1+x)/x$ on (0,1] and extends continuously to x=0 (limit=1). By Abel's theorem, convergence is uniform on [0,1] despite conditional convergence at x=1. Uniform convergence on COMPACT interval justifies term-by-term integration. Mere continuity (C) isn't sufficient; need uniform convergence of SERIES. Option B incorrectly excludes endpoint. Option D is false—conditional convergence doesn't prohibit integration if uniform. This advanced justification connects real analysis theorems to computational practice, elevating understanding beyond mechanical procedure to rigorous foundation awareness expected in higher mathematics.

Q29. A computational engine returns $\int_0^x e^{-t^2} dt \approx x - x^3/3 + x^5/10$ for small x. Without recomputing, how can you QUICKLY verify the third coefficient 1/10 is plausible?

A.Third term comes from integrating

t^4/2!

in

e^{-t^2}

expansion:

\int t^4/2 dt = t^5/(5\cdot2) = t^5/10

. Matches. ✅

B.Coefficient should be 1/5! = 1/120 based on exponential series.

C.Coefficient should be negative since

e^{-t^2}

decreases.

D.Verification requires full recomputation; no quick check exists.

💡 Difficulty: hard | ✅ Correct: A

📖 Explanation: $e^{-t^2} = 1 - t^2 + t^4/2! - \cdots$ . Integrating: $x - x^3/3 + x^5/(5\cdot2!) - \cdots = x - x^3/3 + x^5/10$ . Quick mental verification: identify source term in integrand series, apply power rule, confirm arithmetic. This sanity-checking skill catches transcription/computation errors efficiently. Option B confuses with $e^x$ series; C ignores sign alternation in integrated series; D defeats purpose of estimation. Developing rapid plausibility checks through structural understanding prevents propagation of errors in multi-step series workflows—a professional computational hygiene practice.

Q30. When modeling signal processing filters, you integrate a Fourier-like series term-by-term. If the original series has jump discontinuities, what artifact appears in the integrated series?

A.Integrated series converges uniformly to a continuous function; discontinuities are smoothed out, but Gibbs phenomenon may persist in derivative. ✅

B.Integrated series inherits the same jump discontinuities.

C.Integration amplifies high-frequency oscillations near discontinuities.

D.Term-by-term integration is invalid for discontinuous functions.

💡 Difficulty: medium | ✅ Correct: A

📖 Explanation: Integration is a smoothing operator. Even if original series has jumps (pointwise convergence issues), its integral is CONTINUOUS and series converges uniformly to it. However, DERIVATIVE of integrated series (=original) still exhibits Gibbs overshoot near jumps. This duality—smooth integral vs. oscillatory derivative—is fundamental in signal analysis. Option B/C mischaracterize integration effects; D is false (integration valid, improves convergence). Understanding how calculus operations transform convergence properties and artifacts is crucial for interpreting series-based models in engineering. This bridges pure analysis with applied signal processing phenomenology.

Q31. You need $\int_0^x \frac{1}{1+t^3} dt$ . Instead of deriving new series, you substitute $u=-t^3$ into $\int \frac{1}{1-u} du$ series. What critical step is often forgotten in this substitution approach?

A.Adjusting the differential:

du = -3t^2 dt

means simple substitution doesn't directly yield integrable series; must expand

1/(1+t^3)

as geometric series in

-t^3

FIRST, then integrate. ✅

B.Forgetting to change integration limits when substituting.

C.Missing the negative sign from

du

.

D.Assuming radius of convergence remains 1 after substitution.

💡 Difficulty: hard | ✅ Correct: A

📖 Explanation: Naive substitution $u=-t^3$ in $\int du/(1-u)$ gives $\int -3t^2 dt/(1+t^3)$ , NOT original integral. Correct approach: treat $1/(1+t^3)$ as geometric series $\sum (-t^3)^n$ directly, THEN integrate term-by-term. This distinguishes VARIABLE SUBSTITUTION in integrals from SERIES SUBSTITUTION in expressions. Confusing these is common error. Option B/C address standard u-sub issues irrelevant here; D is secondary concern. Clarifying this distinction prevents fundamental setup errors when adapting known series to new arguments—a frequent task in applied series work requiring careful algebraic framing.

Q32. For $F(x) = \int_0^x \sum_{n=0}^\infty c_n t^n dt$ , which statement about $F'(x)$ is ALWAYS true on $(-R,R)$ ?

A.

F&#039;(x) = f(x) = \sum c_n x^n

by Fundamental Theorem of Calculus and series continuity. ✅

B.

F&#039;(x) = \sum n c_n x^{n-1}

by term-by-term differentiation.

C.

F&#039;(x)

exists only if original series converges absolutely.

D.

F&#039;(x)

may differ from f(x) at points where series converges conditionally.

💡 Difficulty: medium | ✅ Correct: A

📖 Explanation: By construction, $F(x) = \int_0^x f(t) dt$ . Since power series are continuous on $(-R,R)$ , FTC Part 1 guarantees $F'(x) = f(x)$ EVERYWHERE in interval. Term-by-term differentiation of F's series (B) also yields f(x), but FTC provides more direct justification independent of series manipulation. Absolute convergence (C) unnecessary; conditional convergence (D) doesn't affect FTC validity for continuous integrands. This anchors series integration in foundational calculus, reinforcing that series-defined functions obey standard analysis rules. Connecting series operations to core theorems builds coherent conceptual framework rather than isolated procedural knowledge.

Q33. In approximating $\int_0^{0.5} x e^{-x} dx$ , you could integrate series for $xe^{-x}$ OR integrate by parts then use $e^{-x}$ series. Which is computationally superior for hand calculation and why?

A.Integrating

xe^{-x}

series directly:

\sum (-1)^n x^{n+2}/(n+1)!

integrates to single sum; IBP introduces extra terms and products. ✅

B.IBP is superior because it reduces to elementary

e^{-x}

series without x-multiplier.

C.Both are equivalent; choice is purely aesthetic.

D.Direct series integration fails because x-multiplier breaks geometric structure.

💡 Difficulty: easy | ✅ Correct: A

📖 Explanation: Direct: $xe^{-x} = \sum (-1)^n x^{n+1}/n!$ . Integrate: $\sum (-1)^n x^{n+2}/[(n+2)n!]$ . Single clean series. IBP: $-xe^{-x} - e^{-x} + 1$ , then expand $e^{-x}$ , multiply, combine—multiple series manipulations with cancellation risks. Direct path minimizes algebraic overhead. This efficiency judgment matters in timed/exam settings and hand computations. Recognizing when direct series manipulation beats classical calculus shortcuts demonstrates operational fluency. Option B overstates IBP benefit; C ignores practical differences; D is false. Strategic method selection based on expression structure is expert-level problem-solving.

Q34. A student integrates $\sum_{n=0}^\infty x^{2n}$ and writes $\sum_{n=0}^\infty \frac{x^{2n+1}}{2n}$ . Beyond the division-by-zero at n=0, what CONCEPTUAL error does the denominator reveal?

A.Confusing exponent with index; should divide by NEW exponent

2n+1

, not old exponent

2n

. This reflects misunderstanding of power rule mechanics. ✅

B.Misapplying geometric series sum formula instead of term-by-term integration.

C.Assuming even exponents integrate to even denominators.

D.No conceptual error beyond n=0 issue; rest is correct.

💡 Difficulty: hard | ✅ Correct: A

📖 Explanation: Power rule: $\int x^k dx = x^{k+1}/(k+1)$ . Here k=2n, so denominator MUST be 2n+1. Writing 2n shows student is dividing by ORIGINAL exponent, not incremented one. This isn't just arithmetic slip—it reveals flawed mental model of integration as 'divide by current power' rather than 'increment then divide by new power'. Diagnosing ROOT CAUSE of errors (conceptual vs. careless) enables targeted remediation. Surface-level correction misses learning opportunity. This HOTS analysis promotes metacognitive awareness of procedural understanding gaps essential for lasting mastery.

Q35. When using series to evaluate $\int_0^1 \frac{\sin x}{x} dx$ , why is the resulting series MORE numerically stable than direct numerical integration of $\sin x / x$ near x=0?

A.Series removes the 0/0 indeterminacy analytically via cancellation, yielding well-behaved polynomial terms; numerical methods face catastrophic cancellation/division-by-zero near singularity. ✅

B.Numerical integration adapts step size automatically; series has fixed truncation error.

C.Both are equally stable; modern computers handle singularities robustly.

D.Series is LESS stable due to factorial growth in denominators causing underflow.

💡 Difficulty: hard | ✅ Correct: A

📖 Explanation: Near x=0, $\sin x / x$ involves subtracting nearly equal numbers (sin x ≈ x) then dividing by tiny x—classic catastrophic cancellation in floating-point. Series $\sum (-1)^n x^{2n}/(2n+1)!$ evaluates each term stably; no subtraction of close values, no division by near-zero. Analytic preprocessing via series ELIMINATES numerical pathology. This demonstrates series as NUMERICAL REGULARIZATION tool, not just symbolic technique. Understanding stability advantages informs algorithm selection in scientific computing. Option B/C underestimate floating-point realities; D misattributes instability. Bridging analysis and numerical methods is advanced interdisciplinary competency.

Q36. Given $f(x) = \sum_{n=0}^\infty a_n x^n$ with R=2, and $g(x) = \int_0^x f(t) dt$ , which statement about g's Taylor coefficients $b_n$ is correct?

A.

b_n = a_{n-1}/n

for n≥1, and

b_0 = 0

; coefficients are shifted and scaled versions of f's coefficients. ✅

B.

b_n = a_n / (n+1)

; same index, different scaling.

C.

b_n = a_{n+1} \cdot (n+1)

; integration inverts differentiation scaling.

D.Coefficients are unrelated without closed-form expression.

💡 Difficulty: medium | ✅ Correct: A

📖 Explanation: $g(x) = \sum a_n x^{n+1}/(n+1) = \sum b_k x^k$ . Let k=n+1 → n=k-1. So $b_k = a_{k-1}/k$ for k≥1. Constant term $b_0 = g(0) = 0$ . This precise index-shift relationship is fundamental for converting between function and antiderivative series representations. Option B uses wrong indexing; C describes differentiation; D ignores deterministic relationship. Mastering coefficient translation enables seamless navigation between function spaces—a key skill in differential equations and transform methods where series coefficients encode system dynamics.

Q37. In a quantum mechanics perturbation calculation, you integrate $\sum_{n=0}^\infty c_n x^n$ from 0 to ∞. Despite R=∞, the integral diverges. Why doesn't infinite radius guarantee integrability over [0,∞)?

A.Radius describes LOCAL convergence near center; global integrability requires DECAY conditions at infinity unrelated to radius. E.g.,

e^x

has R=∞ but diverges on [0,∞). ✅

B.Infinite radius implies boundedness, so integral must converge; premise is flawed.

C.Term-by-term integration fails on infinite intervals regardless of radius.

D.Quantum integrals always require regularization; series methods are inapplicable.

💡 Difficulty: medium | ✅ Correct: A

📖 Explanation: Radius of convergence governs behavior NEAR expansion point. Behavior AT INFINITY depends on asymptotic growth, not local analyticity. $e^x = \sum x^n/n!$ has R=∞ but $\int_0^\infty e^x dx = \infty$ . Conversely, $e^{-x}$ has R=∞ AND convergent integral. Students often conflate 'converges everywhere' with 'integrable everywhere'. This distinction is VITAL in physics where series approximate wavefunctions/potentials; valid local expansion doesn't ensure physical normalizability. Recognizing scope limitations of series properties prevents unphysical conclusions in modeling.

Q38. You approximate $\int_0^{0.3} \ln(1+x^2) dx$ using series. After integrating, you have alternating series. To achieve error < ε, you solve $a_{N+1} < ε$ . What if NO integer N satisfies this due to slow convergence?

A.Switch to faster-converging series representation (e.g., transform variable, use different expansion point) or employ acceleration techniques like Euler transform. ✅

B.Increase ε to acceptable tolerance; no alternative exists.

C.Use numerical integration instead; series method has failed.

D.Add more terms indefinitely until convergence occurs.

💡 Difficulty: hard | ✅ Correct: A

📖 Explanation: When series converges too slowly for practical truncation, PERSISTENCE is wrong; ADAPTATION is key. Options: variable transformation to shrink argument, expansion about closer point, convergence acceleration (Shanks, Euler), or hybrid numerical-analytic methods. Recognizing FAILURE MODES of primary method and having fallback strategies distinguishes experts from novices. This metacognitive monitoring—assessing efficiency mid-computation and pivoting—is advanced problem-solving. Option B compromises accuracy; C abandons analytical benefits; D is computationally infeasible. Real-world modeling demands flexible toolkit, not rigid adherence to single approach.

📝 Integrating power series (38 MCQs)

What is Integrating power series?

📝 All Integrating power series MCQs

🔗 Related Topics (MCQs)