James Brannick, CU-Boulder; Marian Brezina, CU-Boulder; Rob Falgout, CASC, LLNL; Tom Manteuffel, CU-Boulder; Steve McCormick, CU-Boulder; John Ruge, CU-Boulder
Multigrid methods provide optimal-order solution techniques for the linear systems arising from simulation of many physical problems. Efficient black-box solution of many linear systems may be attained using algebraic multigrid methods that automatically construct the multigrid hierarchy based only on a given matrix. We present recent research into generalizing the limiting assumptions of typical algebraic multigrid methods, particularly emphasizing the use of multiple prototypes of algebraically smooth error in determining interpolation operators.