Kirk M. Soodhalter

Presentations

• Automated Drawing of Weather Fronts
  Invited Talk
  Wednesday, January 427, 2012
  Community College of Philadelphia Math Club
  We present a simplified model for the automatic detection of weather fronts (cold fronts and warm fronts) using data received from weather prediciction simulations. Computed examples are shown and compared with with hand-drawn weather map. This talk is tailored to undergraduates who have completed some calculus courses.
• Block Krylov Subspace Recycling: Theory and Application in a Newton Iteration
  Invited Talk
  Wednesday, January 4, 2012
  2012 Joint Mathematics Meetings
  We derive a version of the GCRODR algorithm (GMRES with subspace recycling) for use in the block Krylov setting. We call this method block GCRODR (block GMRES with recycling). We demonstrate this method's effectiveness as a solver embedded in a Newton iteration arising in fluid density functional theory.
• Block Krylov Subspace Recycling
  Friday, November 4, 2011
  Mid Atlantic Numerical Analysis Day
  We derive a version of the GCRODR algorithm (GMRES with subspace recycling) for use in the block Krylov setting. We call this method block GCRODR (block GMRES with recycling).
• A Schur Complement Method for Solving a Nearly-Hermitian System
  Wednesday, July 20, 2011
  7th International Congress on Industrial and Applied Mathematics
  We discuss our Schur Complement Method for nearly Hermitian linear systems, focusing on instability of a competing method and performance results for larger problems.
• The Schur Complement Method for Nearly-Hermitian Linear Systems: An Effective Solver
  Thursday, June 17, 2011
  Householder Symposium XVIII
  We discuss our Schur Complement Method for nearly Hermitian linear systems, focusing on instability of a competing method and performance results for larger problems.
• A Schur Complement Approach for Solving a Nearly Hermitian System
  Tuesday, March 3, 2011
  SIAM Conference on Computational Science and Engineering
  We discuss an approach for solving an n x n linear system with rank s skew-Hermitian part (s much smaller than n) in a storage efficient manner. This talk focuses on convergence properties of the algorithm.
• Automated Drawing of Weather Fronts
  Wednesday, February 9, 2011
  Temple University Society for Undergraduate Mathematics
  We present a simplified model for the automatic detection of weather fronts (cold fronts and warm fronts) using data received from weather prediciction simulations. Computed examples are shown and compared with with hand-drawn weather map. This talk is tailored to undergraduates who have completed some calculus courses.
• A Schur Complement Approach for Solving a Nearly Hermitian System
  Thursday, July 15, 2010
  SIAM Annual Meeting 2010
  We discuss an approach for solving an n x n linear system with rank s skew-Hermitian part (s much smaller than n) in a storage efficient manner. Numerical results are presented to show competitiveness.
• Automatic Weather Front Detection (with Meredith Hegg)
  Wendsday May 5, 2010
  Temple University Mathematical Modeling Seminar
  We present a simplified model for the automatic detection of weather fronts (cold fronts and warm fronts) using data received from weather prediciction simulations. Computed examples are shown and compared with with hand-drawn weather map.
• Properties of Progressive GMRES and Flexible Conjugate Gradient
  Wendsday October 7, 2009
  Temple University Department of Mathematics Applied Math Seminar
  We study properties and stability of Beckermann and Reichel's Progressive GMRES (ProGMRES) algorithm and also present some numerical experiments for an algorithm called Flexible Conjugate Gradient (FCG) discussed by Notay in a recent paper.
• Properties and Stability of Progressive GMRES
  Wendsday October 26, 2009
  SIAM Conference on Applied Linear Algebra
  We study properties and stability of Beckermann and Reichel's Progressive GMRES (ProGMRES) algorithm. Some possible reasons for instability discussed and backed up with some numerical evidence.
• Progressive GMRES: A Minimum Residual Krylov Method Which is Equivalent to MINRES in the Symmetric Case. Parts I & II
  Wendsday April 8, 2009 and Wendsday April 15, 2009
  Temple University Department of Mathematics Applied Math Seminar
  We discuss Beckermann and Reichel's Progressive GMRES (ProGMRES) algorithm, a minimum residual method which approximates solutions to a nearly symmetric linear system.