1/31
Friday integration lecture (playre2.pdf) and completed Quiz 3 review. Conceptual rules, method choices, domains and common errors; complements Differentiation and Week 2 — Approximations, Extrema & Proof Ideas.
What is an indefinite integral, and why must it include + C?
∫f(x) dx is the family F(x) + C with F′(x) = f(x). Differentiation loses additive constants, so the integrand cannot determine the original vertical position.
How is a definite integral evaluated using an antiderivative?
If f is continuous on [a,b] and F′ = f, then ∫ from a to b f(x) dx = F(b) − F(a). The same constant cancels at both endpoints; evaluate upper minus lower.
Is a definite integral always the total geometric area?
No: it is signed area, with contributions below the x-axis negative. For an odd integrable function, the integral over [−a,a] is 0 by cancellation. Total area is ∫|f(x)| dx.
Why is ∫(1/x) dx = ln|x| + C rather than simply ln x + C?
The integrand is defined for positive and negative x, excluding 0. ln|x| differentiates to 1/x on either side of 0; ln x only works for x > 0. Use the rule on intervals that do not cross 0.
What is u-substitution reversing, and how do you recognise a useful choice?
It reverses the chain rule. Look for a composite g(x) together with its derivative, up to a constant. Set u = g(x), calculate du = g′(x) dx, then rewrite the whole integral in u.
What must change when you substitute in a definite integral?
Change the integrand, the differential and both limits. If u = g(x), the new limits are g(a) and g(b). Evaluate in u with those limits, or return entirely to x before using the original limits; never mix the two.
How does u = e^(−x²) evaluate ∫ from 0 to ∞ x e^(−x²) dx?
First use a finite upper bound b. du = −2x e^(−x²) dx, and the limits become 1 and e^(−b²). The integral is −(1/2)∫ from 1 to e^(−b²) du = (1 − e^(−b²))/2. Let b → ∞ to obtain 1/2. The whole exponential is a valid substitution.
How do substitution and partial fractions give the same formula for arctanh a when 0<a<1?
Integrate 1/(1−x²) from 0 to a. With x=tanh y, dx=sech²y dy and 1−x²=sech²y, so the integral equals arctanh a. By partial fractions, the integrand is ½[1/(1−x)+1/(1+x)], giving ½ log((1+a)/(1−a)). Equating the results gives the inverse hyperbolic tangent formula.
How does the tutorial substitution x=√(sin y) handle ∫ from 0 to 1 of x/√(1−x⁴) dx?
Use a finite upper limit b with b between 0 and 1 and choose y in [0,arcsin(b²)]. Since x²=sin y, 2x dx=cos y dy and √(1−x⁴)=cos y on this range. The integral is ½ arcsin(b²). Letting b approach 1 from below gives π/4.
How do you recognise and integrate (log x)/x in this course?
Here log x = ln x, with x > 0. Set u = ln x, so du = dx/x. Then ∫(ln x)/x dx = ∫u du = (ln x)²/2 + C.
How do you decide between substitution, integration by parts and partial fractions?
Try substitution for a composite with its derivative; parts for a product where differentiating one factor simplifies it; partial fractions for a rational function whose denominator can be factored. Simplify first and check the domain.
What is integration by parts, and where does it come from?
Integrating the product rule gives ∫u dv = uv − ∫v du. For limits a,b: ∫ from a to b u v′ dx = [uv] from a to b − ∫ from a to b u′v dx. In an indefinite answer, include + C.
How should you choose u and dv in integration by parts?
Choose u so differentiation simplifies it, and dv so you can find v. For ∫x e^x dx, choose u = x and dv = e^x dx: du = dx and v = e^x, giving x e^x − e^x + C.
Why may integration by parts need to be repeated?
A polynomial factor loses one degree each time it is differentiated. For x²e^x, applying parts twice gives e^x(x² − 2x + 2) + C. Continue until the remaining integral is straightforward.
What terms are required for a repeated linear factor in partial fractions?
For (x − a)^m, include A₁/(x − a) + A₂/(x − a)² + … + Aₘ/(x − a)^m. Every power from 1 through m is needed. For 1/[(x − 1)(x + 2)²], use A/(x − 1) + B/(x + 2) + C/(x + 2)².
How do you find the constants in a partial-fraction decomposition?
Multiply the identity by the full denominator. Substitute roots of factors to eliminate terms, then compare coefficients or use another convenient x for the remaining constants. For 1/[(x − 1)(x + 2)²], A = 1/9, B = −1/9 and C = −1/3.
How do simple and repeated linear factors integrate differently?
∫A/(x − a) dx = A ln|x − a| + C. For m > 1, use the power rule: ∫A(x − a)^(−m) dx = A(x − a)^(1−m)/(1−m) + C. A squared denominator does not integrate to a logarithm.
What must you check before decomposing a rational function into partial fractions?
The numerator degree must be lower than the denominator degree; otherwise divide polynomials first. Factor the denominator fully over the reals. An irreducible quadratic needs a linear numerator, such as (Ax + B)/(x² + c).
What makes an integral improper?
An integration interval is unbounded, or the integrand is unbounded near an endpoint or an interior point. Define the integral using limits of proper integrals, even when the written endpoints are finite.
What does convergence of an improper integral mean?
The defining limit exists as a finite real number. If the limit is infinite or does not exist, the integral diverges. An unbounded integrand can still have a convergent integral; its growth near the singularity matters.
How do you define ∫ from a to ∞ f(x) dx?
Compute ∫ from a to b f(x) dx for finite b, then take b → ∞. For e^(−x) on [0,∞), the result is lim(1 − e^(−b)) = 1 because e^(−b) → 0. Infinity is a limiting process, not a number to substitute.
How do you handle an improper integral with a singularity inside the interval or with both endpoints infinite?
Split the integral at an interior singularity or at a finite point when both ends are infinite. Take the two limits independently and require each to be finite. For example, split ∫ over the real line of 1/(1+x²) at 0. Using arctan x, both halves give π/2, so the integral is π. Divergent pieces cannot cancel in an ordinary improper integral.
When do power integrals converge near 0 and near ∞?
∫ from 0 to 1 x^(−p) dx converges exactly when p < 1. ∫ from 1 to ∞ x^(−p) dx converges exactly when p > 1. At p = 1, logarithmic divergence occurs in both cases. Check which endpoint is problematic.
Why does ∫ from 0 to 1 1/√x dx converge while ∫ from 0 to 1 1/x² dx diverges?
Use a lower endpoint ε > 0 and let ε → 0⁺. The first is 2 − 2√ε → 2. The second is 1/ε − 1 → ∞. Both integrands blow up, but only the first has a finite limiting area.
Which identities guide trigonometric and hyperbolic substitutions?
sin²t + cos²t = 1 and cosh²t − sinh²t = 1. They simplify square roots: x = sin t turns 1 − x² into cos²t; x = cosh t turns x² − 1 into sinh²t. Also 1 + sinh²t = cosh²t.
How does x = sin t simplify ∫ from 0 to 1 √(1 − x²) dx?
Choose t in [0,π/2]. Then dx = cos t dt, √(1 − sin²t) = cos t, and the new limits are 0 and π/2. The integral becomes ∫ from 0 to π/2 cos²t dt = π/4. Both the root and dx contribute a cos t.
Why must you check signs when simplifying √(cos²t) or √(sinh²t)?
The square root is nonnegative: √(z²) = |z|. Thus √(cos²t) = |cos t| and √(sinh²t) = |sinh t|. You may drop the absolute value only on a chosen range where the expression is nonnegative.
How do you integrate cos²t after a trigonometric substitution?
Use cos²t = (1 + cos 2t)/2. Then ∫cos²t dt = t/2 + sin(2t)/4 + C. The factor 1/4 comes from both the half-angle coefficient and reversing the chain rule for 2t.
How does x = cosh t simplify ∫dx/√(x² − 1), and which branch is used?
For x > 1, choose t > 0. Then dx = sinh t dt and √(cosh²t − 1) = sinh t, so the integrand becomes 1. The answer is arcosh x + C = ln(x + √(x² − 1)) + C on this branch. The x < −1 branch requires separate treatment.
What does cosh^(−1)x mean when undoing x = cosh t?
It means the inverse hyperbolic cosine arcosh x, not 1/cosh x. With t ≥ 0, x = cosh t is one-to-one and t = arcosh x. A reciprocal would be written 1/cosh x (or sech x).
What substitution is useful for √(a² + x²), with a > 0?
Try x = a sinh t. Then dx = a cosh t dt and √(a² + x²) = a cosh t, since cosh t > 0 and 1 + sinh²t = cosh²t. The identity is what makes the substitution useful.
How can you check an antiderivative and catch a missing sign or chain-rule factor?
Differentiate your proposed answer and simplify back to the original integrand, on its domain. For a definite integral of a nonnegative function with lower limit below upper limit, the answer must be nonnegative; a negative result signals a sign or limits error.