16 Pages


Another method is logarithmic scaling, which scales by the logarithm of the greatest magnitude. More sophisticated methods are algorithmic, taking both row and column extremes into account.

scaling argument (homogeneity argument) The idea is to exploit the behavior of norms of functions or vector fields under pull-backs related to scaling transformations x → δx, δ > 0. Very closely related to techniques relying on parametric equivalence.