Gaussian Quadrature

Gaussian quadrature chooses both the nodes and the weights of a quadrature rule to maximize its polynomial degree of exactness, integrating polynomials of degree 2n-1 exactly with only n function evaluations.

Definition

Gaussian quadrature is a family of quadrature rules whose nodes are the roots of orthogonal polynomials associated with a weight function, chosen together with their weights to achieve the maximum possible degree of exactness for a given number of nodes.

Scope

This topic covers the construction of Gaussian rules from the roots of orthogonal polynomials, the Gauss-Legendre rule and weighted variants (Gauss-Chebyshev, Gauss-Hermite, Gauss-Laguerre), the Golub-Welsch eigenvalue algorithm for computing nodes and weights, and Gauss-Kronrod extensions used for practical error estimation.

Core questions

• How does placing nodes at the roots of orthogonal polynomials double the degree of exactness compared with fixed-node rules?
• How are the nodes and weights computed accurately for a given weight function?
• How do weighted Gaussian rules handle integrals with singular or infinite-domain weight functions?
• How are reliable error estimates obtained, for example through Gauss-Kronrod pairs?

Key theories

Maximal degree of exactness

An n-point quadrature rule can be exact for polynomials up to degree 2n-1, and this maximum is attained precisely when the nodes are the roots of the degree-n orthogonal polynomial for the weight function, with all weights positive.

Golub-Welsch algorithm

The nodes and weights of a Gaussian rule are obtained as the eigenvalues and squared first eigenvector components of the symmetric tridiagonal Jacobi matrix formed from the recurrence coefficients of the orthogonal polynomials, turning quadrature construction into an eigenvalue computation.

Mechanisms

Orthogonal polynomials satisfy a three-term recurrence whose coefficients populate a symmetric tridiagonal Jacobi matrix; the Golub-Welsch algorithm computes its eigenvalues (the quadrature nodes) and uses the first components of the eigenvectors to recover the weights, all stably. Changing the weight function — to one with built-in singularities or supported on a half-line or the whole line — yields Gauss-Chebyshev, Gauss-Laguerre, or Gauss-Hermite rules that absorb difficult behaviour analytically. Gauss-Kronrod rules reuse the Gauss nodes and add interlacing nodes so that a higher-order estimate, and hence an error estimate, is obtained at modest extra cost.

Clinical relevance

Gaussian quadrature is the workhorse for evaluating element and stiffness integrals in finite-element analysis, for computing moments and expectations against probability weight functions in statistics and uncertainty quantification, and for high-accuracy evaluation of smooth integrals throughout physics and engineering, where minimizing the number of expensive integrand evaluations is paramount.

History

Gauss derived his optimal quadrature in 1814; Jacobi connected it to orthogonal polynomials, and the modern computational treatment was established by the 1969 Golub-Welsch algorithm, which made nodes and weights routinely computable and brought Gaussian rules into standard numerical libraries.

Key figures

• Carl Friedrich Gauss
• Carl Gustav Jacob Jacobi
• Gene H. Golub
• Walter Gautschi

Seminal works

• davis1984
• gautschi2004

Frequently asked questions

How can n points integrate a degree 2n-1 polynomial exactly?

Because both the n nodes and the n weights are free parameters, there are 2n degrees of freedom, enough to match the integrals of 2n basis polynomials (degrees 0 through 2n-1). Placing the nodes at orthogonal-polynomial roots achieves exactly this.

How is the accuracy of a Gaussian rule checked in practice?

A common approach is the Gauss-Kronrod pair, which augments a Gauss rule with extra nodes to produce a higher-order estimate; the difference between the two estimates serves as a practical error estimate used by adaptive integrators.