Nearly Optimal Block-Jacobi Preconditioning

 James Demmel 

The goal of Jacobi preconditioning π·β’𝐴⁒𝐷𝐷⁒𝐴⁒𝐷 of a symmetric positive definite matrix π΄π΄ by a diagonal matrix π·π· is to choose π·π· to minimize the condition number πœ…⁑(𝐷⁒𝐴⁒𝐷)πœ…β‘(𝐷⁒𝐴⁒𝐷). In 1969, van der Sluis proved that choosing π·π· so that the diagonal entries of π·β’𝐴⁒𝐷𝐷⁒𝐴⁒𝐷 are all ones reduces πœ…⁑(𝐷⁒𝐴⁒𝐷)πœ…β‘(𝐷⁒𝐴⁒𝐷) to within a factor of the minimum possible, where the factor depends on both the dimension π‘›π‘› and the norms used to define the condition number. We extend this result in two ways to block-Jacobi preconditioning, where π·π· is a block-diagonal matrix with blocks of given sizes, and we consider π·β’𝐴⁒𝐷𝑇𝐷⁒𝐴⁒𝐷𝑇 instead of π·β’𝐴⁒𝐷𝐷⁒𝐴⁒𝐷 to maintain the symmetric positive definite (spd) property. First, we extend van der Sluis’s original bound to include block-Jacobi. Second, we define a new norm in which choosing π·π· so that the corresponding diagonal blocks of π·β’𝐴⁒𝐷𝑇𝐷⁒𝐴⁒𝐷𝑇 are identity matrices minimizes the condition number. We use this to show that the condition number in the 2-norm of this optimally scaled π·β’𝐴⁒𝐷𝑇𝐷⁒𝐴⁒𝐷𝑇 is at least as large as the condition number in the new norm, and at most a factor π‘‘2𝑑2 larger, where π‘‘𝑑 is the number of diagonal blocks. We give an example where the optimal 2-norm condition number nearly attains this new upper bound, which for this example is tighter than van der Sluis’s bound by a factor equal to the matrix dimension. Finally, all these results generalize to the case of one-sided scaling π·β’𝐡𝐷⁒𝐡 of a full row-rank matrix π΅π΅.

URL: https://epubs.siam.org/doi/10.1137/22M1504901