If the the linear program throws a warning (ex. matrix is ill-conditioned), just fail and let MIP deal with the constraint. Note that I could apply LU-factorization (to deal with the not-full-row-rank issue) and MAYBE do regularization on the matrix (to hopefully help with the ill-conditioned matrix issue), but I think that may be over-kill for what penaltymodel-lp was meant to do (that being, it deals with the easy to solve constraints).
Closes #102
If the the linear program throws a warning (ex. matrix is ill-conditioned), just fail and let MIP deal with the constraint. Note that I could apply LU-factorization (to deal with the not-full-row-rank issue) and MAYBE do regularization on the matrix (to hopefully help with the ill-conditioned matrix issue), but I think that may be over-kill for what
penaltymodel-lpwas meant to do (that being, it deals with the easy to solve constraints).