Summary#
This chapter introduces numerical integration (quadrature) as a powerful alternative to analytical integration when closed-form antiderivatives are difficult or impossible to obtain. It builds upon ideas from polynomial interpolation to construct interpolatory quadrature rules, and explores both theoretical and practical aspects.
🔹 Section 3.1 – Introduction and Classical Rules
Motivation for numerical quadrature using examples of complicated integrands.
Classical rules reviewed: Left endpoint, Right endpoint, Midpoint, Trapezoidal, and Simpson’s rule.
All rules are cast in the general quadrature form \(Q[f] = \sum w_i f(x_i)\).
🔹 Section 3.2 – Quadrature from Polynomial Interpolation
Derivation of quadrature rules by integrating Lagrange interpolants.
Demonstrates how weights are computed as integrals of cardinal basis functions.
Examples include revisiting Simpson’s rule and introducing Gauß-Legendre quadrature.
A first look at the accuracy of quadrature rules
🔹 Section 3.3 – Degree of Exactness
Defines degree of exactness as the highest degree of polynomial a rule integrates exactly.
Empirical testing and code examples to compute this degree for various quadrature rules.
Notable results: Simpson and Gauss-Legendre have higher precision per point than trapezoidal or midpoint.
🔹 Section 3.4 – Error Estimates
Derives general error bounds for quadrature based on interpolation theory.
Presents sharper estimates for specific rules:
Trapezoidal error: \(\frac{(b-a)^3}{12} f{\prime}{\prime}(\xi)\)
Simpson error: \(-\frac{(b-a)^5}{2880} f^{(4)}(\xi)\)
🔹 Section 3.5 – Newton-Cotes Formulas
Discusses rules with equispaced nodes.
Highlights issues with negative weights for higher-order Newton-Cotes rules, limiting practical use to \(n \leq 6\).
Differentiates between closed (includes endpoints) and open (excludes endpoints) Newton-Cotes rules.
🔹 Section 3.6 – Composite Quadrature Rules
General construction of quadrature rules and composite quadrature rules starting from a reference interval.
Improves accuracy by dividing the domain into subintervals and applying quadrature on each.
Derives and implements composite trapezoidal and midpoint rules.
🔹 Section 3.7 – Composite Trapezoidal Rule & Convergence
Experiments show error decreases as \(\mathcal{O}(h^2)\).
Includes full implementation and analysis of error convergence plots.
Defines experimental order of convergence (EOC) and demonstrates how to estimate it numerically.
🎯 Learning Outcomes for Chapter 3
By the end of this chapter, students will be able to:
💡 Conceptual Understanding
Explain the motivation for numerical integration and its relevance when antiderivatives are unknown or intractable.
Describe classical quadrature rules and identify their strengths and weaknesses.
🔢 Quadrature Design and Implementation
Derive numerical quadrature rules from polynomial interpolation using Lagrange polynomials.
Compute weights in quadrature rules from cardinal functions.
Implement rules such as:
Left/right endpoint
Midpoint and trapezoidal
Simpson’s rule
Gauß-Legendre quadrature
Implement composite quadrature rules for improved accuracy.
📈 Error Analysis
Define and determine the degree of exactness of a quadrature rule.
Use interpolation error estimates to bound integration error based on function regularity and number of nodes.
Apply error estimates for single interval quadrature rules to derive bounds for composite rules.
🧠 Theory and Practice
Discuss Newton-Cotes formulas, including the issues with negative weights and their practical implications.
Implement composite quadrature rules for improved accuracy on general intervals.
Explain and apply the concept of experimental order of convergence (EOC) using log-log plots.
Use composite quadrature rules (e.g., composite trapezoidal) to efficiently and accurately approximate integrals.