Comparison of Richardson's iteration with Chebyshev acceleration factors to conjugate
, by -
-
Minimization of data variance may not be the best basis to judge iterative matrix inversion
schemes. Instead, one may want to produce a maximum likelihood inversion and one may want
to exercise control over the inversion. On this basis, Richardson's iteration with the
application of Chebyshev acceleration factors has advantages over the conjugate gradient
iteration for use in very large inverse problems that are common in geophysics.
STANFORD