Online Packing to Minimize Area or Perimeter

TL;DR

This paper studies online algorithms for packing axis-parallel rectangles to minimize bounding box area or perimeter, providing bounds, algorithms, and analysis for various cases including rotations, aspect ratios, and minimum edge lengths.

Contribution

It introduces new bounds and algorithms for online rectangle packing, especially for minimizing area, and improves analysis for special cases like squares and bounded aspect ratios.

Findings

Perimeter minimization algorithms with competitive ratio just under 4.

Lower bounds showing asymptotic ratios of at least 8(1) for area minimization.

Algorithms matching these bounds for various rectangle constraints.

Abstract

We consider online packing problems where we get a stream of axis-parallel rectangles. The rectangles have to be placed in the plane without overlapping, and each rectangle must be placed without knowing the subsequent rectangles. The goal is to minimize the perimeter or the area of the axis-parallel bounding box of the rectangles. We either allow rotations by 90 degrees or translations only. For the perimeter version we give algorithms with an absolute competitive ratio slightly less than 4 when only translations are allowed and when rotations are also allowed. We then turn our attention to minimizing the area and show that the asymptotic competitive ratio of any algorithm is at least $\Omega(\sqrt{n})$, where $n$ is the number of rectangles in the stream, and this holds with and without rotations. We then present algorithms that match this bound in both cases. We also show that the competitive ratio cannot be bounded as a function of OPT. We then consider two special cases. The first is when all the given rectangles have aspect ratios bounded by some constant. The particular variant where all the rectangles are squares and we want to minimize the area of the bounding square has been studied before and an algorithm with a competitive ratio of 8 has been given [Fekete and Hoffmann, Algorithmica, 2017]. We improve the analysis of the algorithm and show that the ratio is at most 6, which is tight. The second special case is when all edges have length at least 1. Here, the $\Omega(\sqrt n)$ lower bound still holds, and we turn our attention to lower bounds depending on OPT. We show that any algorithm has an asymptotic competitive ratio of at least $\Omega(\sqrt{OPT})$ for the translational case and $\Omega(\sqrt[4]{OPT})$ when rotations are allowed. For both versions, we give algorithms that match the respective lower bounds.