home · mobile · calendar · colloquia · 1995-1996 · 

Colloquium - Gabow

Arborescence Packing: A Case Study in Algorithm Design
Department of Computer Science

This talk examines a specific problem in algorithmic graph theory and surveys a number of efficient algorithms for it. The emphasis is to show generic techniques for algorithm design and graphs. These techniques, e.g., the greedy method, depth-first search, halving and NP-completeness, have widespread applicability; our packing problem illustrates just one possible use. The talk will be a survey, both of recent work on this specific problem and of fundamental techniques. As such we hope that the survey talk will interest both specialists in the area and those just seeking the flavor of algorithm design. The results we describe are theoretical in nature. However to be more down-to-earth we begin with a related real-life application, a system that schedules doctors' attending rounds.

The specific problem we focus on is finding the greatest possible number of edge-disjoint arborescences in a digraph. It is closely related to computing edge connectivity, computing the so-called "strength" of an undirected graph and related quantities. The relations are based on some interesting "minimax" theorems.

Refreshments will be served prior to the talk.

Department of Computer Science
University of Colorado Boulder
Boulder, CO 80309-0430 USA
May 5, 2012 (14:13)