Compute the INVERSE of B = [ INVERSE(R) + cI ] with the SHERMAN-MORRIS
formula modified to take into account the sparseness of the matrix R.
INPUT : R - square matrix stored columnwise as a vector of lengt
c - scalar.
I - identity matrix of order n, which need not to be sto
OLDINV - working square matrix stored columnwise as a vector
length n*n, containing the inverse of B associated w
successive change of the k-th diagonal element (k=1,
n - order of matrices R, OLDINV and BINV.
OUTPUT : BINV - square matrix stored columnwise as a vector of lengt
containing the inverse of B following a change of th
diagonal element (k=n). |