“…(3.7) tells us that e (ν) p → 0 as ν → ∞. It is well-known from the theory of linear iterative schemes that the convergence can be significantly improved by preconditioning (cf., e.g., BANK et al [1990], BRAMBLE et al [1997], ELMAN [2002], ELMAN and GOLUB [1994], ELMAN and SILVESTER [1996], KLAWONN [1998], RUSTEN and WINTHER [1992]). In terms of the Richardson iteration (3.3), we may use…”