A Survey on Spherical Designs: Existence, Numerical Constructions, and Applications
Congpei An, Xiaosheng Zhuang
TL;DR
This survey identifies spherical $t$-designs as exact polynomial quadrature rules on the sphere and traces progress from foundational existence results to practical numerical constructions and broad applications. It highlights the key theoretical milestone that $N_*(d,t)\le C_d t^d$, resolving long-standing bounds, and surveys optimization- and manifold-based algorithms (Newton, BB, trust-region) aided by fast spherical harmonic transforms to generate high-degree designs. The work connects rigorous existence proofs, interval verification, and geometric quality measures with powerful applications in hyperinterpolation, PDEs, integral equations, and spherical data processing. Overall, spherical designs provide a unifying framework for stable quadrature, interpolation, and multiscale analysis on the sphere, with significant implications for numerical analysis and computational science.
Abstract
This paper provides a survey of spherical designs and their applications, with a particular emphasis on the perspective of ``numerical analysis''. A set \(X_N\) of \(N\) points on the unit sphere \(\mathbb{S}^d\) is called a \textit{spherical \(t\)-design} if the average value of any polynomial of degree at most \(t\) over \(X_N\) equals its average over the entire sphere. Spherical designs represent one of the most significant topics in the study of point distributions on spheres. They are deeply connected to algebraic combinatorics, discrete geometry, differential geometry, approximation theory, optimization, coding theory, quantum physics, and other fields, which have led to the development of profound and elegant mathematical theories. This article reviews fundamental theoretical results, numerical construction methods, and applied outcomes related to spherical designs. Key topics covered include existence proofs, optimization-based construction techniques, fast computational algorithms, and applications in interpolation, numerical integration, hyperinterpolation, signal and image processing, as well as numerical solutions to partial differential and integral equations.
