Mathematics / Analysis Integral Calculus & Series 100% Free Open Access
Chapter 1 • Theory & Derivations

Advanced Techniques of Integration & Reduction Formulas

Integration by Parts, Partial Fractions, Weierstrass Substitution & Wallis Products

§1.1Integration by Parts: Differential Foundations, Cyclic Reductions & Tabular Integration

1. Product Rule & Differential Derivation

Integration by parts is the integral calculus analog of the differential product rule. For two continuously differentiable functions $u, v \in C^1([a, b])$, the differential of their product is given by:

$$d(uv) = u \, dv + v \, du$$

Integrating both sides over the real interval $[a, b]$ yields the fundamental integration by parts identity:

$$\int u \, dv = uv - \int v \, du \quad \Longleftrightarrow \quad \int_a^b u(x) v'(x) \, dx = \Big[ u(x)v(x) \Big]_a^b - \int_a^b v(x) u'(x) \, dx$$

2. Optimal Choice of Partitions: The LIATE Heuristic

To ensure that the residual integral $\int v \, du$ is strictly simpler than the original integral $\int u \, dv$, the choice of $u(x)$ typically follows the LIATE priority hierarchy:

  1. Logarithmic functions: $\ln(x), \log_b(x)$ (differentiate to simple rational functions)
  2. Inverse trigonometric functions: $\arcsin(x), \arctan(x), \operatorname{arcsec}(x)$
  3. Algebraic polynomials: $x^n, ax + b$ (differentiate to lower degrees)
  4. Trigonometric functions: $\sin(x), \cos(x), \sec(x)$ (stable under differentiation)
  5. Exponential functions: $e^{ax}, b^x$ (trivially integrable as $dv$)

3. Cyclic (Self-Referential) Integrals

When integrating products of exponential and trigonometric functions, integration by parts reproduces the original integrand after two successive applications. Consider the archetypal cyclic integral:

$$I = \int e^{ax} \cos(bx) \, dx$$

Step 1: Let $u = e^{ax}$ and $dv = \cos(bx) \, dx$. Then $du = a e^{ax} \, dx$ and $v = \frac{1}{b}\sin(bx)$:

$$I = \frac{1}{b} e^{ax}\sin(bx) - \frac{a}{b} \int e^{ax} \sin(bx) \, dx$$

Step 2: Apply integration by parts to the new integral with $u_1 = e^{ax}$ and $dv_1 = \sin(bx) \, dx$, giving $du_1 = a e^{ax} \, dx$ and $v_1 = -\frac{1}{b}\cos(bx)$:

$$\int e^{ax} \sin(bx) \, dx = -\frac{1}{b} e^{ax}\cos(bx) + \frac{a}{b} \int e^{ax}\cos(bx) \, dx = -\frac{1}{b} e^{ax}\cos(bx) + \frac{a}{b} I$$

Step 3: Substituting back into the primary equation:

$$I = \frac{1}{b} e^{ax}\sin(bx) - \frac{a}{b} \left( -\frac{1}{b} e^{ax}\cos(bx) + \frac{a}{b} I \right) = \frac{e^{ax}}{b}\sin(bx) + \frac{a e^{ax}}{b^2}\cos(bx) - \frac{a^2}{b^2} I$$

Collecting like terms in $I$:

$$\left( 1 + \frac{a^2}{b^2} \right) I = \frac{a^2 + b^2}{b^2} I = \frac{e^{ax}}{b^2} \Big( b \sin(bx) + a \cos(bx) \Big)$$
$$I = \int e^{ax}\cos(bx) \, dx = \frac{e^{ax}}{a^2 + b^2} \Big( a \cos(bx) + b \sin(bx) \Big) + C$$

4. The Stand-Alone Logarithmic & Inverse Trigonometric Technique

Integrals of single functions such as $\int \ln(x) \, dx$ or $\int \arctan(x) \, dx$ are evaluated by setting the algebraic factor $dv = dx \implies v = x$:

$$\int \ln(x) \, dx = x \ln(x) - \int x \cdot \frac{1}{x} \, dx = x \ln(x) - \int 1 \, dx = x \ln(x) - x + C = x(\ln x - 1) + C$$
$$\int \arctan(x) \, dx = x \arctan(x) - \int x \cdot \frac{1}{1 + x^2} \, dx = x \arctan(x) - \frac{1}{2} \ln(1 + x^2) + C$$

§1.2Rational Fractions: Heaviside Cover-Up, Repeated Factors & Irreducible Quadratics

1. Algebraic Theory of Partial Fraction Decomposition

Let $R(x) = \frac{P(x)}{Q(x)}$ be a rational function where $P(x), Q(x) \in \mathbb{R}[x]$ are polynomials with real coefficients and $\gcd(P, Q) = 1$.

  • Proper Rational Fractions: If $\deg(P) < \deg(Q)$, the fraction is proper. If $\deg(P) \ge \deg(Q)$, polynomial long division must first be executed: $R(x) = S(x) + \frac{P_1(x)}{Q(x)}$ where $\deg(P_1) < \deg(Q)$.
  • By the Fundamental Theorem of Algebra, any monic polynomial $Q(x) \in \mathbb{R}[x]$ factors uniquely over $\mathbb{R}$ into products of distinct and repeated linear factors $(x - r)^m$ and irreducible quadratic factors $(x^2 + px + q)^k$ with negative discriminant $p^2 - 4q < 0$.

2. Distinct Linear Factors & Heaviside's Cover-Up Method

When $Q(x) = (x - r_1)(x - r_2)\cdots(x - r_n)$ has distinct real roots $r_k$:

$$\frac{P(x)}{Q(x)} = \frac{A_1}{x - r_1} + \frac{A_2}{x - r_2} + \dots + \frac{A_n}{x - r_n}$$

Multiplying both sides by $(x - r_k)$ and evaluating the limit as $x \to r_k$ eliminates all terms except $A_k$, establishing Heaviside's Cover-Up Formula:

$$A_k = \lim_{x \to r_k} (x - r_k) \frac{P(x)}{Q(x)} = \frac{P(r_k)}{Q'(r_k)}$$

3. Repeated Linear Factors & Irreducible Quadratics

For higher multiplicities and quadratic factors:

  1. Repeated Linear: A factor $(x - r)^m$ contributes $m$ partial fractions: $$\frac{A_1}{x - r} + \frac{A_2}{(x - r)^2} + \dots + \frac{A_m}{(x - r)^m}$$
  2. Irreducible Quadratic: A factor $(x^2 + px + q)$ ($p^2 - 4q < 0$) requires a linear numerator: $$\frac{Bx + C}{x^2 + px + q}$$ Completing the square $x^2 + px + q = \left(x + \frac{p}{2}\right)^2 + \left(q - \frac{p^2}{4}\right) = u^2 + a^2$ decomposes the integral into a logarithmic part and an arctangent part: $$\int \frac{Bx + C}{x^2 + px + q} \, dx = \frac{B}{2}\ln(x^2 + px + q) + \frac{C - \frac{Bp}{2}}{\sqrt{q - p^2/4}} \arctan\left(\frac{x + p/2}{\sqrt{q - p^2/4}}\right) + C_0$$

§1.3Trigonometric Integrals & The Universal Weierstrass Substitution

1. Powers and Products of Trigonometric Functions

Integrals of the form $\int \sin^m(x) \cos^n(x) \, dx$ are categorized by parity of the exponents:

  • Case 1 ($m$ or $n$ is Odd): If $n = 2k + 1$ is odd, isolate $\cos(x)\,dx = d(\sin x)$ and convert remaining cosines using $\cos^{2k}(x) = (1 - \sin^2 x)^k$. The substitution $u = \sin(x)$ yields a purely polynomial integral: $$\int \sin^m(x) \cos^{2k+1}(x) \, dx = \int u^m (1 - u^2)^k \, du$$
  • Case 2 (Both $m$ and $n$ are Even): Apply the half-angle power reduction identities: $$\sin^2(x) = \frac{1 - \cos(2x)}{2}, \quad \cos^2(x) = \frac{1 + \cos(2x)}{2}, \quad \sin(x)\cos(x) = \frac{\sin(2x)}{2}$$

2. The Universal Weierstrass Half-Angle Substitution

For any rational function of trigonometric terms $R(\sin x, \cos x)$, the transformation $t = \tan\left(\frac{x}{2}\right)$ maps the trigonometric integral onto a purely rational algebraic integral over $t \in \mathbb{R}$.

Geometric Derivation: From the double-angle identities:

$$\cos(x) = \cos^2(x/2) - \sin^2(x/2) = \frac{\cos^2(x/2) - \sin^2(x/2)}{\cos^2(x/2) + \sin^2(x/2)} = \frac{1 - \tan^2(x/2)}{1 + \tan^2(x/2)} = \frac{1 - t^2}{1 + t^2}$$
$$\sin(x) = 2\sin(x/2)\cos(x/2) = \frac{2\tan(x/2)}{1 + \tan^2(x/2)} = \frac{2t}{1 + t^2}$$

Differentiating $x = 2\arctan(t)$ yields the differential element:

$$dx = \frac{2}{1 + t^2} \, dt$$
$$\mathbf{\int R(\sin x, \cos x) \, dx = \int R\left( \frac{2t}{1 + t^2}, \, \frac{1 - t^2}{1 + t^2} \right) \frac{2}{1 + t^2} \, dt}$$

This universal substitution transforms every rational trigonometric expression into a standard partial fraction problem.

§1.4Successive Reduction Formulas: Recurrence Relations for Higher Powers

1. General Philosophy of Reduction Formulas

When an integral depends on an integer parameter $n \in \mathbb{N}$, a reduction formula expresses the integral $I_n$ in terms of $I_{n-1}$ or $I_{n-2}$, allowing recursive reduction to elementary base cases ($I_0$ or $I_1$).

2. Reduction Formula for $I_n = \int \sin^n(x) \, dx$

Split the integrand as $\sin^n(x) = \sin^{n-1}(x) \cdot \sin(x)$ and set up integration by parts:

$$u = \sin^{n-1}(x) \implies du = (n-1)\sin^{n-2}(x)\cos(x) \, dx$$
$$dv = \sin(x) \, dx \implies v = -\cos(x)$$

Applying the formula $\int u \, dv = uv - \int v \, du$:

$$I_n = -\sin^{n-1}(x)\cos(x) + (n-1) \int \sin^{n-2}(x)\cos^2(x) \, dx$$

Using the Pythagorean identity $\cos^2(x) = 1 - \sin^2(x)$:

$$\begin{aligned} I_n &= -\sin^{n-1}(x)\cos(x) + (n-1) \int \sin^{n-2}(x)(1 - \sin^2 x) \, dx \\ &= -\sin^{n-1}(x)\cos(x) + (n-1) \int \sin^{n-2}(x) \, dx - (n-1) \int \sin^n(x) \, dx \\ &= -\sin^{n-1}(x)\cos(x) + (n-1) I_{n-2} - (n-1) I_n \end{aligned}$$

Collecting terms in $I_n$ on the left-hand side:

$$I_n + (n-1)I_n = n I_n = -\sin^{n-1}(x)\cos(x) + (n-1)I_{n-2}$$
$$\mathbf{I_n = \int \sin^n(x) \, dx = -\frac{\sin^{n-1}(x)\cos(x)}{n} + \frac{n-1}{n} I_{n-2}}$$

3. Reduction Formula for $K_n = \int \sec^n(x) \, dx$

Decomposing $\sec^n(x) = \sec^{n-2}(x) \cdot \sec^2(x)$ with $u = \sec^{n-2}(x)$ and $dv = \sec^2(x) \, dx$ gives $v = \tan(x)$ and $du = (n-2)\sec^{n-2}(x)\tan(x) \, dx$:

$$K_n = \sec^{n-2}(x)\tan(x) - (n-2) \int \sec^{n-2}(x)\tan^2(x) \, dx$$

Since $\tan^2(x) = \sec^2(x) - 1$:

$$K_n = \sec^{n-2}(x)\tan(x) - (n-2) \left[ K_n - K_{n-2} \right]$$
$$\mathbf{K_n = \int \sec^n(x) \, dx = \frac{\sec^{n-2}(x)\tan(x)}{n-1} + \frac{n-2}{n-1} K_{n-2}}$$

§1.5Wallis Formulas & The Infinite Product for $\pi/2$

1. Definite Integrals on $[0, \pi/2]$

Applying the reduction formula for $\sin^n(x)$ to the definite integral $W_n = \int_0^{\pi/2} \sin^n(x) \, dx$:

$$W_n = \left[ -\frac{\sin^{n-1}(x)\cos(x)}{n} \right]_0^{\pi/2} + \frac{n-1}{n} W_{n-2} = 0 + \frac{n-1}{n} W_{n-2}$$

Base cases:

$$W_0 = \int_0^{\pi/2} 1 \, dx = \frac{\pi}{2}, \qquad W_1 = \int_0^{\pi/2} \sin(x) \, dx = \Big[ -\cos(x) \Big]_0^{\pi/2} = 1$$

2. Closed-Form Wallis Formulas

Unwinding the recurrence relation $W_n = \frac{n-1}{n} W_{n-2}$ yields distinct expressions based on parity:

  • Even Index ($n = 2m$): $$W_{2m} = \frac{2m-1}{2m} \cdot \frac{2m-3}{2m-2} \cdots \frac{1}{2} \cdot W_0 = \frac{(2m-1)!!}{(2m)!!} \cdot \frac{\pi}{2}$$
  • Odd Index ($n = 2m + 1$): $$W_{2m+1} = \frac{2m}{2m+1} \cdot \frac{2m-2}{2m-1} \cdots \frac{2}{3} \cdot W_1 = \frac{(2m)!!}{(2m+1)!!}$$

3. The Wallis Ratio & Infinite Product for $\pi/2$

Since $0 \le \sin(x) \le 1$ for all $x \in [0, \pi/2]$, powers are monotonically decreasing:

$$\sin^{2m+1}(x) \le \sin^{2m}(x) \le \sin^{2m-1}(x) \implies W_{2m+1} \le W_{2m} \le W_{2m-1}$$

Dividing by $W_{2m+1}$:

$$1 \le \frac{W_{2m}}{W_{2m+1}} \le \frac{W_{2m-1}}{W_{2m+1}} = \frac{2m+1}{2m} = 1 + \frac{1}{2m}$$

By the Squeeze Theorem, as $m \to \infty$:

$$\lim_{m \to \infty} \frac{W_{2m}}{W_{2m+1}} = 1 \implies \lim_{m \to \infty} \left( \frac{(2m-1)!! (2m+1)!!}{[(2m)!!]^2} \cdot \frac{\pi}{2} \right) = 1$$

Inverting the ratio establishes Wallis' Celebrated Infinite Product (1655):

$$\mathbf{\frac{\pi}{2} = \lim_{m \to \infty} \prod_{k=1}^m \frac{(2k)(2k)}{(2k-1)(2k+1)} = \frac{2}{1} \cdot \frac{2}{3} \cdot \frac{4}{3} \cdot \frac{4}{5} \cdot \frac{6}{5} \cdot \frac{6}{7} \cdots}$$
TIERED UNIVERSITY HONORS PROBLEMS

Step-by-Step Solved Examination Problems

Comprehensive analytical derivations, multi-tier solutions (Foundational, Intermediate Exam, and Honors/Proof Challenge) with complete line-by-line verification.

Tier 1: Foundational Repeated Integration by Parts for $x^3 e^{2x}$
Evaluate the indefinite integral using repeated integration by parts and verify via the tabular method: $$I = \int x^3 e^{2x} \, dx$$
Tier 2: Intermediate Exam Partial Fractions with Irreducible Quadratics
Evaluate the definite integral using partial fraction decomposition: $$I = \int_0^1 \frac{2x^2 + 3}{(x+1)(x^2 + 1)} \, dx$$
Tier 3: Honors / Proof Challenge Wallis Asymptotics & Gaussian Integral Connection
Consider the integral $I_n = \int_0^1 (1 - x^2)^n \, dx$ for $n \in \mathbb{N}$. (a) Establish a reduction formula connecting $I_n$ to $I_{n-1}$. (b) Evaluate $I_n$ in terms of the double factorial. (c) Prove that $\lim_{n \to \infty} \sqrt{n} \, I_n = \frac{\sqrt{\pi}}{2}$, establishing the link to the Gaussian integral $\int_{-\infty}^\infty e^{-t^2} dt = \sqrt{\pi}$.