| Recent developments in tiling theory, primarily in the study of anisohedral shapes, have been the product of exhaustive computer searches through various classes of polygons. I present a brief background of tiling theory and past work, with particular emphasis on isohedral numbers, aperiodicity, Heesch numbers, criteria to characterize isohedral tilings, and various details that have arisen in past computer searches.My results include the description of this novel approach to testing tiling properties, a correction to previous descriptions of the criteria for characterizing isohedral tilings, the verification of some previous results on polyforms, and the discovery of two new 4-anisohedral polysnakes.I then develop and implement a new "boundary-based" technique, characterizing shapes as a sequence of characters representing unit length steps taken from a finite language of directions, to replace the "area-based" approaches of past work, which treated the Euclidean plane as a regular lattice of cells manipulated like a bitmap. The new technique allows me to reproduce and verify past results on polyforms (edge-to-edge assemblies of unit squares, regular hexagons, or equilateral triangles) and then generalize to a new class of shapes dubbed polysnakes, which past approaches could not describe. My implementation enumerates polyforms using Redelmeier's recursive generation algorithm, and enumerates polysnakes using a novel approach. The shapes produced by the enumeration are subjected to tests to either determine their isohedral number or prove they are non-tiling. |
