Spring 2010 CSE642 Algorithm Seminar
Locations
and Hours:
Friday 11:00 -12:30 @
Computer Science Lounge (CS Building 1211).
Course
Description:
This reading group
provides a meeting place for Stony Brook faculty, postdocs,
and students interested in the analysis of algorithms. We meet once a
week,
with one of three different missions:
You can get one credit for
participating by simply registering for CSE 642. You are also welcome
to come
without registering.
We have a mailing list.
Further announcement will be distributed through the list. Please
subscribe here.
Schedule:
Feb 4th: Joe brought the puzzles on Communications of the ACM. Jie mentioned a problem on structural balance in social networks. Thanks to Roozbeh Ebrahimi Soorchaei, we have scribed notes on the problems.
Feb 11th: presentation by visiting student Stefan Huber. The title and abstract of the talk are as follows:
Theoretical and
Practical Results on Straight Skeletons of Planar Straight-Line Graphs
The
straight skeleton of a simple polygon, or more generally, of a planar
straight-line graph, is a geometric structure similar to Voronoi
diagrams. Straight skeletons have many applications, including the
computation of straight-line oset curves, generating roofs and
terrains, or solving mathematical origami problems.
We start with an extension of the so-called motorcycle graph to planar
straight-line graphs and present new theoretical contributions
regarding the relation of the motorcycle graph and the straight
skeleton of arbitrary planar straight-line graphs. We give a new
characterization of the straight skeleton based on the lower envelope
of partial linear functions with domains that depend on the motorcycle
graph. An immediate application is a straight-skeleton algorithm based
on graphics hardware by rendering the lower envelope of 3D slabs.
We also present a novel wavefront-type algorithm for the computation of
straight skeletons which bridges the gap between theory and practice of
straight skeleton computations. Our algorithm handles arbitrary planar
straight-line graphs, is easy to implement, and is
fast enough to handle complex data. It can be expected to run in O(n
log n) time in practice and its worst-case complexity is O(n^2 log n)
for an n-vertex input graph. Experimental results conrm an average
runtime of 20 n log n
on a standard PC for virtually all of our
13,500 datasets. Compared to the current CGAL implementation, this
constitutes an average gain in performance by a multiplicative factor
of n, or at least one to two orders of magnitude for our datasets.
Feb 18th: open problem session. We revisited the structural balance problem.
Feb 25th: open problem session. We revisited the structural balance problem on a local search algorithm.
March 4th: open problem session. New results on optimal solution to turn a weakly balanced graph to strongly balanced.
March 11th: presentation by
David
Bergman, the CMU student (undergrad and MS from Stony Brook). His talk
title and abstract are as follows:
Title:
Manipulating Multivalued Decision Diagrams for Combinatorial
Optimization Problems.
Abstract: In this talk we study the application of limited-width
multivalued decision diagrams (MDDs) as discrete relaxations for
combinatorial optimization problems. We introduce a new compilation
method to efficiently construct such MDDs and also present a number of
algorithms for MDD manipulation that yield stronger relaxations. Our
methodology is applied to set covering problems, and evaluated against
relaxations based on linear programming. The experimental results
indicate that MDD-based relaxations are particularly effective on
structured problems, outperforming state-of-the-art integer programming
technology by several orders of magnitude.
March 18th: Steve Skiena led the discussion on a new problem.
March 25th: Presentation by Dr. Walter Diumo on elevator scheduling problem.
April 1st: Paper presentation by Chloe Kong on Shortest Non‐Crossing Walks in the Plane, J.Erickson, A. Nayyer, SODA'11.
April 8th: Paper presentation by Jethro Chu on A Stackelberg Strategy for Routing Flow over Time, U. Bhaskar, L.Fleischer, E. Anshelevic, SODA'11.
April 15th: Michael Bender led the discussion on a new problem.
April 22nd: spring break.
April 29th: CS department GRC day, no reading group.
May 6th: Joe brought a game problem of grabbing pizza pieces, brownies, sub sandwiches, etc.