I'm not certain exactly what characteristic of the problem results in this issue, but I have seen it often. The primal/dual interior point algorithm solves an A x = b linear system of equations to compute the update step, and in some cases the corresponding A matrix becomes singular or nearly singular. It may be that the problem is too flat or something, I'm not sure. If it solves successfully anyway, just printing these warnings, it's fine, but sometimes the optimization will fail completely. In that case, my only solution is to switch to a different solver.
There are probably improvements that can be made to the MIPS algorithm to make it more robust under these circumstances. Any suggestions are appreciated. -- Ray Zimmerman Senior Research Associate 419A Warren Hall, Cornell University, Ithaca, NY 14853 phone: (607) 255-9645 On Jan 31, 2012, at 4:23 AM, Boris Penaloza wrote: > Warning: Matrix is singular to working precision. > > In mips at 422 > In mipsopf_solver at 145 > In opf_execute at 106 > In opf at 225 > In runopf at 96 > In pqinput at 53 > > I am getting this message error as I try to run a simulation. Does anybody > know what it means?? > > thanks...
