This book is the modern first treatment of experimental designs,
providing a comprehensive introduction to the interrelationship between
the theory of optimal designs and the theory of cubature formulas in
numerical analysis. It also offers original new ideas for constructing
optimal designs.
The book opens with some basics on reproducing kernels, and builds up to
more advanced topics, including bounds for the number of cubature
formula points, equivalence theorems for statistical optimalities, and
the Sobolev Theorem for the cubature formula. It concludes with a
functional analytic generalization of the above classical results.
Although it is intended for readers who are interested in recent
advances in the construction theory of optimal experimental designs, the
book is also useful for researchers seeking rich interactions between
optimal experimental designs and various mathematical subjects such as
spherical designs in combinatorics and cubature formulas in numerical
analysis, both closely related to embeddings of classical
finite-dimensional Banach spaces in functional analysis and Hilbert
identities in elementary number theory. Moreover, it provides a novel
communication platform for "design theorists" in a wide variety of
research fields.