On Planar Visibility Counting Problem

TL;DR

This paper introduces a new randomized algorithm for the exact planar visibility counting problem, balancing space, preprocessing, and query time with a tunable parameter, improving over previous approximation methods.

Contribution

It presents a novel randomized approach for exact visibility counting with adjustable complexity trade-offs based on a parameter.

Findings

Achieves space complexity of O(n^{4-4α})

Preprocessing time is O(n^{4-2α})

Query time is O(n^{2α})

Abstract

For a set $S$ of $n$ disjoint line segments in $\mathbb{R}^{2}$, the visibility counting problem is to preprocess $S$ such that the number of visible segments in $S$ from any query point $p$ can be computed quickly. There have been approximation algorithms for this problem with trade off between space and query time. We propose a new randomized algorithm to compute the exact answer of the problem. For any $0<\alpha<1$, the space, preprocessing time and query time are $O_{\epsilon}(n^{4-4\alpha})$, $O_{\epsilon}(n^{4-2\alpha})$ and $O_{\epsilon}(n^{2\alpha})$, respectively. Where $O_{\epsilon}(f(n)) = O(f(n)n^{\epsilon})$ and $\epsilon>0$ is an arbitrary constant number.