[Help-glpk] Re: Using glpk for a sparse LU

[Andrew Makhorin]

>Is there a way to use the library, or lower-level routines, to implement a
>sparse LU solver?

See the module glpluf.c which implements sparse LU-factorization and
can be used to solve sparse linear systems. The typical call sequence
is the following:

   luf_create      create LU-factorization
   luf_decomp      compute LU-factorization for a given matrix A
   luf_solve       solve system A*x = b or A'*x = b
   luf_delete      delete LU-factorization

Glpluf.c is not documented, however, it is provided with detailed
comments; see also comments in glpluf.h.

Andrew Makhorin