Restricted-orientation convexity. (Q1428352): Difference between revisions

Convexity is generalized if the set of all lines is replaced by a smaller set of lines. The authors first treat the two-dimensional case. An orientation set \(O\) is a set of given lines through the origin, an \(O\)-line is a translate of a line in \(O\), and a set is \(O\)-convex if its intersection with every \(O\)-line is connected. Strongly \(O\)-convex sets are defined in the following manner. The \(O\)-block of two points is the intersection of all halfplanes whose boundaries are \(O\)-lines and which contain both points. A set is strongly \(O\)-convex if, for every two of its points, their \(O\)-block is contained in the set. In order to have this possibility also in higher dimensions, orientation sets are then defined in the following way: An orientation set \(O\) is a set of hyperplanes through the origin, an \(O\)-hyperplane is a translate of a hyperplane in \(O\), and an \(O\)-line is a line which is an intersection of \(O\)-hyperplanes. Many properties of \(O\)-convex and strongly \(O\)-convex sets are similar to properties of standard convex sets. For instance, a closed set with nonempty interior is strongly \(O\)-convex if and only if for every boundary point there is a strongly \(O\)-convex hyperplane through it supporting the set. But \(O\)-convex sets may be disconnected. This gives rise to the definition of \(O\)-connected sets. The presentation is elementary, except the chapter on computability where some knowledge on running time is assumed. There are references to applications, but those are not described in this booklet. It is the reviewer's impression that applications of restricted orientation convexity in three dimensions have still to be found.