[Ipopt] Segmentation fault (LP objFun = 0.0 and mehrotra_algorithm = yes)

ChiaWeng Boon chiaweng at gmail.com
Mon Sep 26 07:04:43 EDT 2011


I am using IPOPT to find the analytic center of a set of linear
inequalities, and I am exploiting the concept of linear programming by
setting the objective function to 0.

min cx
s.t. B-Ax > 0.0

Ipopt gets the analytic center correctly when the "mehrotra_algorithm"
option is set to yes.  However, I noticed that when "tol" is set to a
value smaller 1e-6, Ipopt gives a segmentation fault.   And when I set
"expect_infeasible_problem" to yes while solving an infeasible
problem, I also get a segmentation fault.  Are there any options that
I can set to avoid this.

Another note is that if the objective function cx is not 0, Ipopt
solves the linear program nicely.  I can also find the analytic center
by solving max Sum_log_i (B_i -A_i*x), but then this would invoke the
slower NLP solver.  I would appreciate some advice.


Chia Weng

More information about the Ipopt mailing list