Tuesday, June 17

MS10
Combinatorial Scientific Computing: Where Discrete Mathematics Meets Computational Science

3:00 PM - 5:30 PM
Room: Silver Maple Room - 4th Floor

Combinatorial Scientific Computing (CSC) is an inter-disciplinary field where discrete mathematics and algorithmics are used to identify and solve combinatorial subproblems from computational science and engineering. While these problems occur in areas as disparate as astrophysics and nanoscience, a unifying factor is that they can be posed and solved as problems on graphs and hypergraphs. Practical efficiency and ease of implementation are paramount considerations for CSC problems since they enable large scale simulations in computational science. To continue to foster the CSC community, three international workshops have been organized since 2004, and more are planned. The speakers in this minisymposium will talk about recent developments in some CSC problems: hypergraph partitioning, graph sparsification, combinatorial preconditioning, and approximations for vertex-weighted matchings.

Organizer: Alex Pothen
Purdue University
Bruce Hendrickson
Sandia National Laboratories

3:00-3:25 Approximation Algorithms for Vertex Weighted Matching

Alex Pothen, Purdue University; Florin Dobrian, Columbia University; Mahantesh Halappanavar, Old Dominion University

3:30-3:55 Open Problems in Combinatorial Preconditioning

Sivan A. Toledo, Massachusetts Institute of Technology

4:00-4:25 Graph Sparsification by Effective Resistances

Nikhil Srivastava and Daniel Spielman, Yale University

4:30-4:55 Graph and Hypergraph Models and Algorithms for Sparse Matrix-Vector Multiplication

Erik G. Boman and Michael M. Wolf, Sandia National Laboratories