Date of Award

2006

Degree Type

Thesis

Degree Name

Master of Science

Program

Computer Science

Supervisor

Dr. Yuri Boykov

Abstract

Maximum flow algorithms must balance two competing factors in performance: flow should reach its destination via short paths, but finding short paths is expensive. Push-relabel algorithms compute these paths via an evolving map, called a labeling, over the vertices. The first push-relabel algorithms (Goldberg & Tarjan [32]), use three well-known relabeling operations: (local) relabel, global relabel, and gap relabel. We contribute a new push-relabel algorithm that uses more flexible relabeling operations. The three standard operations all become special cases of our new and more general relabeling operations. On sparse graphs, our algorithm is easy to parallelise and operates efficiently within limited memory. What makes our contribution of particular interest is that the original push-relabel algorithms still, to this day, have the most robust performance in practice [1, 16, 9. Our preliminary experiments suggest that we have opened a door to improving the state of the art for immense, sparse grid-graphs common in computer vision and biomedical imaging applications.

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.