Momentum accelerated power iterations and the restarted Lanczos method
By: Alessandro Barletta, Nicholas Marshall, Sara Pollock
Potential Business Impact:
Speeds up finding important numbers in big problems.
In this paper we compare two methods for finding extremal eigenvalues and eigenvectors: the restarted Lanczos method and momentum accelerated power iterations. The convergence of both methods is based on ratios of Chebyshev polynomials evaluated at subdominant and dominant eigenvalues; however, the convergence is not the same. Here we compare the theoretical convergence properties of both methods, and determine the relative regimes where each is more efficient. We further introduce a preconditioning technique for the restarted Lanczos method using momentum accelerated power iterations, and demonstrate its effectiveness. The theoretical results are backed up by numerical tests on benchmark problems.
Similar Papers
On convergence of greedy block nonlinear Kaczmarz methods with momentum
Numerical Analysis
Solves hard math problems faster with a new trick.
Efficient Computation of Dominant Eigenvalues Using Adaptive Block Lanczos with Chebyshev Filtering
Numerical Analysis
Finds important numbers in big math problems faster.
On finite precision block Lanczos computations
Numerical Analysis
Makes computer math work better for big problems.