The so-called grade of a vector with respect to a nonsingular matrix is the dimension of the (largest) Krylov (sub)space generated by from . It determines in particular, how many iterations a Krylov space method with linearly independent residuals requires for finding in exact arithmetic the solution of (if the initial approximation is the zero vector). In this talk we generalize the grade notion to block Krylov spaces and show that this and other fundamental properties carry over to block Krylov space methods for solving linear systems with multiple right-hand sides.
We consider linear systems with the same nonsingular coefficient matrix , but different right-hand sides , which we gather in a block vector . The systems are then written as
The block grade of with respect to or, the block grade of with respect to is the positive integer defined by
Among the results we have established for the block grade are the following ones.
LEMMA 1 For ,
LEMMA 2 The block grade of the block Krylov space and the grades of the individual Krylov spaces contained in it are related by
LEMMA 3 The block grade is characterized by
THEOREM Let be the block solution of and let be any initial block approximation of it and the corresponding block residual. Then
We also discuss the effects of the size of the block grade on the efficiency of a block Krylov space method.