Parallel algorithms for polygonal and rectilinear geometry.
Sumanta Guha · Deep Blue (University of Michigan) · 1991
Parallel algorithms are presented for geometric problems of two types: problems about simple polygons and problems dealing with the rectilinear metric. Of the first type, an important result presented is an efficient parallel algorithm to compute shortest paths and shortest path trees in simple polygons. Applications of this shown include an efficient parallel algorithm to compute farthest neighbors of vertices of a simple polygon (both with distances measured by shortest paths internal and external to the polygon), an efficient parallel algorithm to determine if two simple polygons may be separated by a single translation, and an NC algorithm to compute the Voronoi diagram of a set of points inside a simple polygon. Another result presented is an optimal parallel algorithm to decide if a simple polygon is monotone. Of the second type, the important result obtained is an optimal parallel algorithm to compute the Voronoi diagram of a set of points on the plane under the rectilinear metric. Another result is an NC algorithm to find the shortest rectilinear path between two points on the plane avoiding rectangular barriers. The third result presented is an optimal parallel algorithm to bipartition a planar point set into two subsets of prescribed rectilinear diameter.