Polytopes that fill \(\mathbb{R}^ n\) and scissors congruence (Q1892432): Difference between revisions

A nice connection between two problems of D. Hilbert (namely, the eighteenth and the third one) is presented. First, Hilbert asked for polyhedra that can fill the space with congruent copies and do not appear as fundamental regions of groups of motions. (An answer was given by K. Reinhardt in 1928, who described such a non-convex polyhedron in \(\mathbb{R}^ 3\).) Second, Hilbert referred to Euclidean scissors congruence of three-dimensional polyhedra of equal volume. (Two polyhedra \(P\) and \(Q\) are Euclidean scissors congruent if \(P\) can be cut up into finitely many pieces by plane cuts such that these pieces can be reassembled, by Euclidean motions, to form \(Q\). It was shown by M. Dehn in 1900 that the regular tetrahedron is not Euclidean scissors congruent to the 3-cube.) The authors prove that any convex or non-convex polyhedron \(P\), that can fill the 3-space with congruent copies, is Euclidean scissors congruent to a 3-cube. In view of his result mentioned above, M. Dehn derived invariants \(\Delta_{n,j} (P)\), \(1\leq j\leq n\), which gave necessary conditions for two (sets of) polytopes to be Euclidean scissors congruent in \(\mathbb{R}^ n\). Generalizing their necessary criterion stated above, the authors also show (by different methods) that any convex or non-convex polytope that can fill the \(n\)-space with congruent copies of itself satisfies \(\Delta_{n,j} (P) =0\) for \(1\leq j\leq n-1\). Moreover, they prove that any convex or non-convex polytope that fills \(\mathbb{R}^ n\) by translations is translation scissors congruent to an \(n\)-cube. [Editor's remarks: In an `Acknowledgment of priority' (pp. 359-360) the authors have given additional comments concerning prior work on these problems dealt with].