Online Class Cover Problem
Minati De, Anil Maheshwari, Ratnadip Mandal
TL;DR
This paper investigates the online class cover problem in the plane with ${\mathcal P}_r$ red points and online-arriving blue points, using the family ${\mathcal F}$ of axis-parallel unit squares. It proves a fundamental lower bound of $\Omega(\log m)$ for deterministic online algorithms and furnishes a matching $O(\log m)$-competitive algorithm that, upon each uncovered blue point, emits at most five candidate squares derived from a staircase-based construction. The algorithm’s correctness is established via geometric invariants and a series of lemmas ensuring coverage without red points, and the competitive ratio is refined from $20\log m$ down to $10+10\log m$. The results illuminate the online behavior of geometric set covers and suggest avenues for extensions to other object families and potential randomized improvements, with practical implications for dynamic resource placement and pattern recognition tasks where red points denote forbidden regions.
Abstract
In this paper, we study the online class cover problem where a (finite or infinite) family $\cal F$ of geometric objects and a set ${\cal P}_r$ of red points in $\mathbb{R}^d$ are given a prior, and blue points from $\mathbb{R}^d$ arrives one after another. Upon the arrival of a blue point, the online algorithm must make an irreversible decision to cover it with objects from $\cal F$ that do not cover any points of ${\cal P}_r$. The objective of the problem is to place a minimum number of objects. When $\cal F$ consists of axis-parallel unit squares in $\mathbb{R}^2$, we prove that the competitive ratio of any deterministic online algorithm is $Ω(\log |{\cal P}_r|)$, and also propose an $O(\log |{\cal P}_r|)$-competitive deterministic algorithm for the problem.
