المفاهيم الأساسية
Online algorithms for packing convex polygons by translation alone do not admit constant competitive ratios, in contrast to the case where rotations are allowed or the offline setting.
الملخص
The paper investigates several online packing problems where convex polygons arrive one by one and must be placed irrevocably into a container, with the goal of minimizing the used space. The key findings are:
Online algorithms for packing convex polygons by translation alone do not admit constant competitive ratios, in contrast to the case where rotations are allowed or the offline setting.
To prove these lower bounds, the authors introduce a novel combinatorial problem called "Online Sorting" and show that it does not admit constant competitive algorithms either. This reveals a deep connection between geometric packing problems and purely combinatorial problems.
On the positive side, the authors present non-trivial algorithms for both online sorting and online strip packing of convex polygons.
The authors show that sorting the pieces by the slope of their "spine segments" is essential for obtaining efficient offline packing algorithms, but this sorting step is inherently difficult in the online setting.
The results highlight the challenges in extending efficient offline packing algorithms to the online setting, especially when rotations are not allowed.