24 Sample problems and algorithms ETH Z（24个示例问题和算法ETH Z）.pdfVIP

Sample problems and algorithms 1 24 Sample problems and algorithms The nature of geometric comput ation: Three pr oblems and algorithms chosen to illustrate the variety of issues encountered. (1) Convex hull yields to simpl e and eff icient algorithms, straightf orward to impl ement and analyze. (2) Objects with spe cial pr operties, such as convexity, are often much simpl er to pr ocess than are general objects. (3) Visibility pr oblems are surpr isingly compl ex ; even if this compl exity does not show in the design of an algorithm, it sneaks into its analysis. 24.1 Geometry and geometric computation Classical geometry, shaped by the ancient Greeks, is more axiomatic than constructive: It emphasizes axioms, theorems, and proofs, rather than algorithms. The typical statement of Euclidean geometry is an assertion about all geometric configurations with certain properties (e.g., the theorem of Pythagoras: In a right- angled triangle, the square on the hypotenuse c is equal to the sum of the squares on the two catheti a and b: c2 = a2 + b2) or an assertion of existence (e.g., the parallel axiom: Given a line L and a point P ∉ L, there is exactly one line parallel to L passing through P). Constructive solutions to problems do occur, but the theorems about the impos sibility of constructive solutions steal the glory: You cannot trisect an arbitrary angle using ruler and compass only, and the proverbial It is impossible to square the circle. Computational geometry, on the other hand, starts out with problems of construction so simple that, until the 1970s, they were dismissed as trivial: Given n line segments in the plane, are they free of intersections? If not, compute (construct) all intersections. But