While the Fundamental Theorem of Calculus provides a powerful way to evaluate definite integrals using antiderivatives, there are many situations where finding an analytical antiderivative is difficult or impossible. This occurs for:

• Functions with no elementary antiderivative: For example, functions like $e^{-x^2}$ (crucial in statistics), $\sin(x^2)$, or $\frac{\sin x}{x}$ do not have antiderivatives that can be expressed in terms of elementary functions (polynomials, exponentials, logarithms, trigonometric functions, etc.).
• Functions defined by data: In many scientific and engineering applications, a function is known only through a set of discrete data points obtained from measurements or experiments. In such cases, there's no explicit formula to integrate.
• Highly complex integrands: Even if an antiderivative exists, it might be extremely complicated to find or evaluate.

Numerical integration (also known as numerical quadrature) provides methods to approximate the value of a definite integral $\int_a^b f(x) \, dx$. These methods typically involve dividing the interval $[a,b]$ into smaller subintervals and approximating the area under the curve on each subinterval using simple geometric shapes (like rectangles, trapezoids, or parabolas).

The accuracy of the approximation generally improves as the number of subintervals increases (i.e., as the width of each subinterval decreases).

The Trapezoidal Rule approximates the area under the curve $y=f(x)$ from $x=a$ to $x=b$ by dividing the interval into $n$ subintervals of equal width $h = \frac{b-a}{n}$, and then approximating the area over each subinterval with a trapezoid.

The area of a single trapezoid formed by the points $(x_i, 0)$, $(x_{i+1}, 0)$, $(x_i, y_i)$, and $(x_{i+1}, y_{i+1})$ is $\frac{1}{2}(y_i + y_{i+1})h$.

Summing the areas of all $n$ trapezoids gives the formula for the Trapezoidal Rule:

$$ \int_a^b f(x) \, dx \approx \frac{h}{2} [y_0 + 2y_1 + 2y_2 + \dots + 2y_{n-1} + y_n] $$

Or, more compactly:

$$ \int_a^b f(x) \, dx \approx \frac{h}{2} \left[ f(x_0) + 2 \sum_{i=1}^{n-1} f(x_i) + f(x_n) \right] $$

Where $x_i = a + ih$ and $y_i = f(x_i)$. The rule essentially averages the function values at the endpoints of each subinterval and multiplies by the width.

The Mid-ordinate Rule (also known as the Midpoint Rule) approximates the definite integral by summing the areas of rectangles. For each subinterval, the height of the rectangle is taken as the value of the function at the midpoint of that subinterval.

Again, divide the interval $[a,b]$ into $n$ subintervals of equal width $h = \frac{b-a}{n}$.

The midpoint of the $i$-th subinterval $[x_i, x_{i+1}]$ is $m_i = \frac{x_i + x_{i+1}}{2} = a + (i + \frac{1}{2})h$.

The area of the rectangle for this subinterval is $h \cdot f(m_i)$.

The formula for the Mid-ordinate Rule is:

$$ \int_a^b f(x) \, dx \approx h [f(m_0) + f(m_1) + \dots + f(m_{n-1})] $$

Or, more compactly:

$$ \int_a^b f(x) \, dx \approx h \sum_{i=0}^{n-1} f\left(a + \left(i + \frac{1}{2}\right)h\right) $$

The Mid-ordinate Rule is often more accurate than the Trapezoidal Rule for the same number of subintervals, as the errors from approximating the curve above and below the rectangle tend to cancel out more effectively.

Simpson's Rule provides a more accurate approximation by fitting parabolic arcs to pairs of subintervals, instead of straight lines (Trapezoidal Rule) or constant values (Mid-ordinate Rule).

A key requirement for Simpson's 1/3 Rule is that the number of subintervals, $n$, must be even. This means there will be an odd number of ordinates ($y_0, y_1, \dots, y_n$).

The interval $[a,b]$ is divided into $n$ (even) subintervals of width $h = \frac{b-a}{n}$.

The formula for Simpson's 1/3 Rule is:

$$ \int_a^b f(x) \, dx \approx \frac{h}{3} [y_0 + 4y_1 + 2y_2 + 4y_3 + \dots + 2y_{n-2} + 4y_{n-1} + y_n] $$

More compactly:

$$ \int_a^b f(x) \, dx \approx \frac{h}{3} \left[ f(x_0) + 4\sum_{i=1, i \text{ odd}}^{n-1} f(x_i) + 2\sum_{i=2, i \text{ even}}^{n-2} f(x_i) + f(x_n) \right] $$

Simpson's Rule is generally much more accurate than the Trapezoidal and Mid-ordinate rules for smooth functions, as it uses a quadratic approximation (parabola) which can better fit the curvature of $f(x)$. It is exact for polynomials of degree 3 or less.

Numerical integration finds wide application across various fields of science, engineering, and finance:

• Engineering:
  • Calculating areas of irregular shapes (e.g., cross-sections of beams, land surveying).
  • Determining volumes of solids (e.g., earthwork calculations, fluid tank capacities).
  • Finding centroids, moments of inertia, and other geometric properties.
  • Solving problems in fluid dynamics, heat transfer, and structural analysis where quantities are expressed as integrals.
• Physics:
  • Calculating work done by a variable force: $W = \int F(x) \, dx$.
  • Finding the distance traveled from a variable velocity: $s = \int v(t) \, dt$.
  • Problems in electromagnetism, quantum mechanics, and thermodynamics often involve integrals that lack simple analytical solutions.
• Statistics and Probability:
  • Calculating probabilities from probability density functions (e.g., the normal distribution function $e^{-x^2}$).
  • Determining expected values and variances.
• Computer Graphics:
  • Rendering realistic lighting and shadows often involves solving integral equations.
• Economics and Finance:
  • Calculating consumer surplus or producer surplus.
  • Valuing financial derivatives where payoffs are path-dependent.
• When dealing with experimental data: If a function is only known at discrete data points, numerical integration is the only way to estimate its integral.

The choice of method (Trapezoidal, Simpson's, etc.) often depends on the desired accuracy, the nature of the function (smoothness), and computational cost.