Chapter 15 – Geometric Shortest Paths and Network Optimization

Given a sequence of k polygons in the plane, a start point s, and a target point, t, we seek a shortest path that starts at s, visits in order each of the polygons, and ends at t. If the polygons are disjoint and convex, we give an algorithm running in time O(kn log (n/k)), where n is the total number of vertices specifying the polygons. We also extend our results to a case in which the convex polygons are arbitrarily intersecting and the subpath between any two consecutive polygons is constrained to lie within a simply connected region; the algorithm uses O(nk2 log n) time. Our methods are simple and allow shortest path queries from s to a query point t to be answered in time O(k log n + m), where m is the combinatorial path length. We show that for nonconvex polygons this "touring polygons" problem is NP-hard.The touring polygons problem is a strict generalization of some classic problems in computational geometry, including the safari problem, the zoo-keeper problem, and the watchman route problem in a simple polygon. Our new results give an order of magnitude improvement in the running times of the safari problem and the watchman route problem: We solve the safari problem in O(n2 log n) time and the watchman route problem (through a fixed point s) in time O(n3 log n), compared with the previous time bounds of O(n3) and O(n4), respectively.

/pdf/touring-a-sequence-of-polygons-2904k7apcq.pdf

Touring a sequence of polygons

https://www.cs.bgu.ac.il/~matya/Papers/tspn.pdf

TSP with neighborhoods of varying size

We revisit one of the most fundamental classes of data structure problems in computational geometry: range searching. Back in SoCG'92, Matousek gave a partition tree method for d-dimensional simplex range searching achieving O(n) space and O(n1-1/d) query time. Although this method is generally believed to be optimal, it is complicated and requires O(n1+e) preprocessing time for any fixed e>0. An earlier method by Matousek (SoCG'91) requires O(n log n) preprocessing time but O(n1-1/d logO(1)n) query time. We give a new method that achieves simultaneously O(n log n) preprocessing time, O(n) space, and O(n1-1/d) query time with high probability. Our method has several advantages: It is conceptually simpler than Matousek's SoCG'92 method. Our partition trees satisfy many ideal properties (e.g., constant degree, optimal crossing number at almost all layers, and disjointness of the children's cells at each node). It leads to more efficient multilevel partition trees, which are important in many data structural applications (each level adds at most one logarithmic factor to the space and query bounds, better than in all previous methods). A similar improvement applies to a shallow version of partition trees, yielding O(n log n) time, O(n) space, and O(n1-1=⌊d=2⌋) query time for halfspace range emptiness in even dimensions d ≥ 4. Numerous consequences follow (e.g., improved results for computing spanning trees with low crossing number, ray shooting among line segments, intersection searching, exact nearest neighbor search, linear programming queries, finding extreme points, ... ).

/pdf/optimal-partition-trees-5ca69yghqq.pdf

Optimal partition trees

In TSP with neighborhoods (TSPN) we are given a collection X of k polygonal regions in the plane, called neighborhoods, with totally n vertices, and we seek the shortest tour that visits each neighborhood. In this paper we present a simple and fast algorithm that, given a start point, computes a TSPN tour of length O(log k) times the optimum in time O(n+k log k). When no start point is given we show how to compute a "good" start point in time O(n2 logn), hence we obtain a logarithmic approximation algorithm that runs in time O(n2 log n). We also present an algorithm which performs at least one of the following two tasks (which of these tasks is performed depends on the given input): (1) It outputs in time O(n log n) a TSPN tour of length O(log k) times the optimum. (2) It outputs a TSPN tour of length less than (1 + e) times the optimum in time O(n3), where e is an arbitrary real constant given as an optional parameter.The same approach can be used for the Red-Blue Separation Problem. We show an algorithm with logarithmic approximation ratio that runs in time O(n log n), where n is the total number of points.

A fast approximation algorithm for TSP with neighborhoods

We propose a very simple ray-shooting algorithm, whose only data structure is a triangulation. The query algorithm, after locating the triangle containing the origin of the ray, walks along the ray, advancing from one triangle to a neighboring one until the polygon boundary is reached. The key result of the paper is a Steiner triangulation of a simple polygon with the property that a ray can intersect at most O(log n) triangles before reaching the polygon boundary. We are able to compute such a triangulation in linear sequential time, or in O(log n) parallel time using O(n/log n) processors. This gives a simple, yet optimal, ray-shooting algorithm for a simple polygon. Using a well-known technique, we can extend our triangulation procedure to a multiconnected polygon with k components and n vertices, so that a ray intersects at most O(?k log n) triangles.

J. Hershberger

Papers

A Pedestrian Approach to Ray Shooting