Volume Maximization and Orthoconvex Approximation of Orthogons

Abstract.

Let n axes-parallel hyperparallelepipeds (also called blocks) of the d-dimensional Euclidean space and a positive integer r be given. The volume maximization problem (VMP) selects at most r blocks such that the volume of their union becomes maximum. VMP is shown to be NP-hard in the 2-dimensional case and polynomially solvable for the line via a constrained critical path problem (CCPP) in an acyclic digraph. This CCPP leads to further well solvable special cases of the maximization problem. In particular, the following approximation problem (OAP) becomes polynomially solvable: given an orthogon P (i.e., a simple polygon in the plane which is a union of blocks) and a positive integer q, find an orthoconvex orthogon with at most q vertices and minimum area, which contains P.