umf_lusolve — solve a linear sparse system given the LU factors
[x] = umf_lusolve(LU_ptr, b [, st, A])
a pointer to umf lu factors (L,U,p,q,R)
a real or complex column vector or a matrix (multiple rhs)
(optional) a string "Ax=b" (default) or "Ax'=b" (to be written "Ax''=b" in scilab langage: a quote in a string must be doubled !)
(optional) the sparse square matrix corresponding to the LU factors (LU_ptr must be got with LU_ptr = umf_lufact(A))
a column vector or a matrix in case of multiple rhs ( x(:,i) is solution of A x(:,i) = b(:,i) or A'x(:,i) = b(:,i) )
This function must be used in conjonction with umf_lufact which
computes the LU factors of a sparse matrix. The optional
st argument lets us choose between the solving of Ax=b (general case)
or of A'x=b (sometimes useful). If you give the 4th argument then
iterative refinement will be also proceceed (as in umfpack) to
give a better numerical solution.