Sommaire:

Although entirely unphysical, a generalized understanding of the random packing of spheres in high dimensions could shed light on the low dimensional case. By separating the general effects from the idiosyncratic at each dimension it may be possible to uncover simple principles which are masked in dimensions two and three. Towards this end, I report on a newly developed, highly efficient, GPU-based simulation package with which to effect the packing of polydisperse spheres in arbitrary dimensions. I further report on a newly developed analysis technique with which to efficiently characterize high dimensional packings by the additively weighted Voronoi tesselation. Using these techniques we explore the structure of packings in several arenas: 1) we test the assumption of a flat measure over packing structure for 2D bidisperse packings, 2) we measure the dependence of number of nearest neighbors on global packing fraction, and 3) we observe the evolution of the Granocentric Model with increased dimension.

Pour plus d'informations, merci de contacter Clusel M.