Radii of regular polytopes (Q1764177): Difference between revisions

Given an \(n\)-dimensional convex body \(P\), it is possible to define the radius of the body in many different ways. For example, one could fix some value of \(j\) with \(1\leq j\leq n\), and a point \(o\) within \(P\), and examine all possible ``slices'' of \(P\) by \(j\)-dimensional affine spaces passing through \(o\). One could then find, for each slice, the radius of the inscribed or circumscribed \(j\)-sphere. This radius would depend not just on \(j\) and \(o\), but also the orientation of the affine space. Minimising or maximising the inscribed or circumscribed radii over all affine spaces for a given \(j\) and \(o\) gives some measure of the radius of the convex body itself. In fact, this process gives four different radius measures for each \(j\). The article under review lists the four different radii for all \(j\), for all \(n\), for the regular simplex, cube and cross-polytope (with \(o\) taken as the centre of the polytope in each case). In the course of preparing the article, the author discovered that certain cases thought unsolved had in fact been previously solved and published in the Russian mathematical literature. Conversely, he also discovered that in other cases, the published solutions had serious flaws in the proof, which he was able to correct. The article gives a brief but thorough literature review of the subject matter, followed by a table of the results. The remainder (and bulk) of the article is devoted to proving the cases whose solutions had not previously been published.