| We report on a novel, constraint-based heuristic search method for discovering dense packings of spherical and non-spherical particles. The divide and concur scheme uses a deterministic map in a non-physical configuration space as its search dynamic. Among the important features of the application of this scheme to packing problems are its aggressive variation of unit-cell parameters and the mutability of particle shapes. The generality and efficiency of the method, as demonstrated in several applications, and its highly non-physical dynamics make it a strong alternative to more conventional physics-inspired methods when optimal configurations are more of an interest than physical pathways. We present the main features of the divide and concur search method and some previous and recent results, including surprising results in the problems of tetrahedron and pentatope packing. |
|