Sampling Random Polygons.
Abstract
Every realization of a Poisson line process is a set of lines which subdivides the plane into a population of nonoverlapping convex polygons. To explore the unknown statistical features of this population, an alternative stochastic construction of random polygons is developed. This construction, which is based on an alternating sequence of random angles and side lengths, provides a fast simulation method for obtaining a random sample from the polygon population. For the isotropic case, this construction is used to obtain a random sample of 2,500,000 polygons, providing the most precise estimates to date of some of the unknown distributional characteristics. Keywords: geometric probability.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 29, 1986
- Accession Number
- ADA176465
Entities
People
- Edward I. George
Organizations
- Stanford University