Abstract
dc:description.abstractSampling of combinatorial structures is an important statistical tool used in applications in a number of areas ranging from statistical physics, data mining, to biological sciences. Of comparable importance is the computation of the cor- responding partition function, which, in the case of the uniform distribution, is equivalent to the problem of counting all such structures. For self-reducible combinatorial structures, once we can produce an almost uniform sample from them, then we can approximately count them. Using a Markov chain Monte Carlo method, this thesis presents polynomial-time algorithms to approximately count and almost uniformly sample crossing-free matchings for certain input classes of graphs. Since the problem in its generality appears to be difficult, we made natural restrictions on the in- put graphs. Namely, we consider vertices arranged in a grid in the plane, where edges are line segments connecting the vertices and a matching is crossing-free if no two matching edges intersect. For appropriate bounds on the dimensions of the grid and the edge lengths, we show that a natural Markov chain is rapidly mixing and that the problem is self-reducible.
Degree
thesis:*- Name thesis:degree_name
- Computer Science (MS)
- Level thesis:degree_level
- Thesis
- Discipline thesis:degree_discipline
- Computer Science (GCCIS)
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Sun, Wenbo
- Contributors dc:contributor
-
- Ivona Bezakova
Subjects
dc:subject × 2Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.rit.edu/theses/9057