Problem 4.1: Three Dimensional Mazes

Download as pdf or txt
Download as pdf or txt
You are on page 1of 2

6.

849 Problem Set 4 Erik Demaine


Joseph Bergeron March 6, 2017

Problem 4.1

Three Dimensional Mazes

This question builds on the posed question of the efficiency of extruded structures in terms
of area efficiency. While 2D maze folding has been posed as a question, here we’re interested
in three dimensional mazes. Intuitively, you can imagine an order n three dimensional maze
to be a three dimensional lattice consisting of n by n by n unit cubes packed into one single
cube of side length n. This lattice may have any number of faces missing from any arbitrary
number of the original unit cubes. Think of an n by n by n honeycomb of cubes, where cubes
can have faces missing. This definition should make sense as a three dimensional analog of
a traditional two dimensional maze.

There are, of course, a few natural questions that arise. In keeping with the question of area
efficiency, we may ask, for the class off all order n 3D mazes, what is the area efficiency? Are
there any pathological (poor area efficiency) or nice (good area efficiency) examples? What
gadgets would be convenient to develop in order to try to optimize efficiency? How does area
efficiency between “sparse” (many faces removed from the lattice) and “dense” (few faces
removed from the lattice) mazes compare?

Of course, the question of how to generate a crease pattern also arises. Given a specification
of a 3D maze, how can one generate a crease pattern? Is it simply foldable, etc. etc.

Problem 4.3

a) Given any shape (such as this polygon), the straight skeleton will always contain O(n)
nodes and edges. From this straight skeleton, which necessarily includes the angle bisectors
of the given polygon, we can imagine iteratively rabbit-earring each angle bisector such that
we continually make one more edge align with the rest, until they all do. Each rabbit-ear
corresponds to a constant number of folds, so the resultant crease pattern will have a O(n)
folds, lining up all polygon edges, as desired. (Note that the final rabbit ear may folds
through intermediate layers, however since this is a constant number of rabbit ears, we’re
fine.)

b A sketch of an O(n log n) algorithm is as follows. We first compute the straight skeleton
of the given polygon, which can be done in O(n log n) since our polygon is monotone [1].
(There may potentially be a batter result for convex polygons, especially considering the

1
size off our polygon is O(1)). Then, each vertex on the straight skeleton will correspond to
a single rabbit-ear fold on the crease pattern. We can greedily assign rabbit ears to vertices
on the straight skeleton by arbitrarily choosing a first rabbit ear and continuing aroundf
the model, picking rabbit-ears that work each time. This entire algorithm takes a total of
O(n log n)

References
[1] Therese Beidl, Marin Held, Stefan Huber, Dominik Kaaser, Peter Palfrader A sim-
ple algorithm for computing positively weighted straight skeletons of monotone polygons
http://www.sciencedirect.com/science/article/pii/S0020019014001987

You might also like