From istefanov87 at abv.bg Thu Jun 2 08:42:56 2022 From: istefanov87 at abv.bg (Ivo Stefanov) Date: Thu, 2 Jun 2022 15:42:56 +0300 (EEST) Subject: [Ipopt] strange behavior by IpOpt solver on a simple problem Message-ID: <1043220795.2650624.1654173779220@nm1.abv.bg> Hello, I am facing some randomness in the results of an optimization with IpOpt. The problem at hand is very simple, basically a linear problem with a single convex quadratic constraint (163 variables, 85 constraints in total). I am running it in 2 environments - using MUMPS as linear solver locally under Windows and using PARDISO as linear solver within a Docker container (Linux) on AWS Lambda. I am consistently getting the same result running with MUMPS under Windows (optimal solution is found), but when I start running with PARDISO under Linux things start to vary a bit. In case the problem is solved, the number of iterations is not consistent between runs (like 56, 54, 40 and similar numbers), which is the first thing that seems strange. What is worse is that some runs also return infeasible. The parameters I am using are: constr_viol_tol=1.0E-8 acceptable_tol=1.0E-8 nlp_scaling_method=none linear_solver=pardiso The reason for the 1e-8 acceptable tolerance is that with 1e-6 it used to terminate with "solved to acceptable level" but the number of iterations and the solution were again a bit variable. The original problem that had this behavior was slightly bigger (some more variables and linear constraints that I removed while preserving the behavior) and for it the log (print_level = 5) under Windows (always consistent) looks like that: Number of nonzeros in equality constraint Jacobian...: 7005 Number of nonzeros in inequality constraint Jacobian.: 1160 Number of nonzeros in Lagrangian Hessian.............: 30381 Total number of variables............................: 495 variables with only lower bounds: 0 variables with lower and upper bounds: 415 variables with only upper bounds: 0 Total number of equality constraints.................: 168 Total number of inequality constraints...............: 338 inequality constraints with only lower bounds: 0 inequality constraints with lower and upper bounds: 3 inequality constraints with only upper bounds: 335 iter objective inf_pr inf_du lg(mu) ||d|| lg(rg) alpha_du alpha_pr ls 0 -9.3890030e-001 7.40e-001 9.53e-001 -1.0 0.00e+000 - 0.00e+000 0.00e+000 0 1 -9.4908233e-001 5.36e-001 2.55e+000 -1.0 5.55e+000 - 7.66e-002 2.75e-001f 1 2 -9.1593284e-001 3.98e-001 2.06e+000 -1.0 5.89e+000 - 3.66e-001 2.59e-001f 1 3 -8.5070790e-001 2.24e-001 1.15e+000 -1.0 8.38e+000 - 4.72e-001 4.36e-001f 1 4 -8.1124641e-001 1.31e-001 1.91e+001 -1.0 1.34e+001 - 6.67e-001 4.16e-001f 1 5 -7.8493779e-001 6.88e-002 6.00e+001 -1.0 2.41e+001 - 6.93e-001 4.75e-001f 1 6 -7.6972130e-001 3.23e-002 1.82e+002 -1.0 2.61e+001 - 1.00e+000 5.30e-001f 1 7 -7.6114851e-001 1.15e-002 3.11e+002 -1.0 1.62e+001 - 1.00e+000 6.45e-001f 1 8 -7.5882188e-001 5.71e-003 1.19e+003 -1.0 5.02e+000 - 1.00e+000 5.02e-001h 1 9 -7.5751120e-001 2.46e-003 2.50e+003 -1.0 2.42e+000 - 1.00e+000 5.70e-001h 1 iter objective inf_pr inf_du lg(mu) ||d|| lg(rg) alpha_du alpha_pr ls 10 -7.5688726e-001 9.08e-004 4.80e+003 -1.0 1.04e+000 - 1.00e+000 6.31e-001h 1 11 -7.5669715e-001 4.35e-004 1.58e+004 -1.0 4.23e-001 - 1.00e+000 5.20e-001h 1 12 -7.5658130e-001 1.47e-004 2.54e+004 -1.0 1.84e-001 - 1.00e+000 6.62e-001h 1 13 -7.5655154e-001 7.33e-005 9.92e+004 -1.0 6.91e-002 - 1.00e+000 5.02e-001h 1 14 -7.5655092e-001 7.17e-005 4.36e+005 -1.0 3.10e-002 - 1.00e+000 2.11e-002f 6 15 -7.5652721e-001 1.28e-005 1.40e+005 -1.0 3.03e-002 - 1.00e+000 8.22e-001h 1 16 -7.5652706e-001 1.24e-005 1.94e+006 -1.0 6.25e-003 - 1.00e+000 2.96e-002f 5 17 -7.5652337e-001 3.22e-006 9.75e+005 -1.0 5.24e-003 - 1.00e+000 7.40e-001h 1 18 -7.5652335e-001 3.19e-006 6.55e+006 -1.0 1.36e-003 - 1.00e+000 1.15e-002f 7 19 -7.5652214e-001 1.81e-007 6.34e+005 -1.0 1.35e-003 - 1.00e+000 9.43e-001h 1 iter objective inf_pr inf_du lg(mu) ||d|| lg(rg) alpha_du alpha_pr ls 20 -7.5652212e-001 1.12e-007 1.03e+007 -1.0 1.86e-004 - 1.00e+000 3.80e-001f 2 21 -7.5652207e-001 2.84e-014 1.00e-006 -1.0 4.73e-005 - 1.00e+000 1.00e+000h 1 22 -7.7092225e-001 8.18e-007 3.10e+006 -5.7 2.79e-003 - 9.10e-001 1.00e+000f 1 23 -1.1535250e+000 1.02e-003 2.65e+006 -5.7 1.17e-001 - 1.45e-001 1.00e+000f 1 24 -1.5162014e+000 2.16e-003 1.53e+006 -5.7 1.98e-001 - 4.22e-001 9.15e-001f 1 25 -1.5313987e+000 1.72e-003 5.66e+005 -5.7 7.83e-002 - 6.31e-001 2.05e-001h 1 26 -1.5756761e+000 8.36e-004 4.15e+005 -5.7 3.37e-001 - 2.66e-001 6.72e-001f 1 27 -1.5873856e+000 4.76e-004 2.53e+005 -5.7 5.45e-001 - 3.90e-001 5.02e-001h 1 28 -1.5960386e+000 2.80e-004 9.44e+004 -5.7 5.00e-001 - 6.27e-001 4.68e-001f 1 29 -1.6042665e+000 1.44e-004 5.64e+004 -5.7 8.50e-001 - 4.02e-001 5.15e-001h 1 iter objective inf_pr inf_du lg(mu) ||d|| lg(rg) alpha_du alpha_pr ls 30 -1.6081474e+000 9.48e-005 3.23e+004 -5.7 8.60e-001 - 4.26e-001 3.61e-001h 1 31 -1.6108033e+000 6.66e-005 1.57e+004 -5.7 1.53e+000 - 5.14e-001 3.24e-001h 1 32 -1.6158989e+000 3.28e-005 1.07e+004 -5.7 2.84e+000 - 3.18e-001 7.55e-001f 1 33 -1.6173806e+000 1.84e-005 6.80e+003 -5.7 3.57e+000 - 3.66e-001 5.26e-001f 1 34 -1.6188377e+000 7.80e-006 2.39e+003 -5.7 3.79e+000 - 6.49e-001 7.46e-001f 1 35 -1.6199353e+000 1.46e-006 7.30e+002 -5.7 4.93e+000 - 6.94e-001 1.00e+000f 1 36 -1.6201399e+000 2.23e-007 1.94e-005 -5.7 1.82e+000 - 1.00e+000 1.00e+000h 1 37 -1.6201630e+000 1.47e-008 1.11e-006 -5.7 4.42e-001 - 1.00e+000 1.00e+000h 1 38 -1.6206358e+000 2.10e-007 4.27e+001 -8.6 3.08e+000 - 7.05e-001 7.71e-001f 1 39 -1.6207397e+000 1.06e-007 3.38e+001 -8.6 3.86e+000 - 5.28e-001 6.20e-001f 1 iter objective inf_pr inf_du lg(mu) ||d|| lg(rg) alpha_du alpha_pr ls 40 -1.6207786e+000 5.09e-008 1.76e+001 -8.6 6.83e+000 - 5.38e-001 5.73e-001f 1 MUMPS returned INFO(1) = -9 and requires more memory, reallocating. Attempt 1 Increasing icntl[13] from 5 to 10. MUMPS returned INFO(1) = -9 and requires more memory, reallocating. Attempt 2 Increasing icntl[13] from 10 to 20. 41 -1.6208005e+000 2.06e-008 1.00e+001 -8.6 8.34e+000 - 5.49e-001 6.37e-001f 1 MUMPS returned INFO(1) = -9 and requires more memory, reallocating. Attempt 1 Increasing icntl[13] from 20 to 40. 42 -1.6208099e+000 7.30e-009 5.20e+000 -8.6 6.53e+000 - 5.70e-001 6.73e-001f 1 43 -1.6208143e+000 1.14e-009 2.03e+000 -8.6 3.34e+000 - 7.01e-001 8.67e-001f 1 44 -1.6208150e+000 9.50e-012 4.90e-001 -8.6 8.60e-001 - 7.96e-001 1.00e+000f 1 45 -1.6208150e+000 1.29e-013 4.30e-012 -8.6 2.68e-002 - 1.00e+000 1.00e+000h 1 Number of Iterations....: 45 (scaled) (unscaled) Objective...............: -1.6208149683584612e+000 -1.6208149683584612e+000 Dual infeasibility......: 4.3001435745537719e-012 4.3001435745537719e-012 Constraint violation....: 5.5511151231257827e-016 5.5511151231257827e-016 Complementarity.........: 2.5826863273493910e-009 2.5826863273493910e-009 Overall NLP error.......: 2.5826863273493910e-009 2.5826863273493910e-009 Number of objective function evaluations = 62 Number of objective gradient evaluations = 46 Number of equality constraint evaluations = 62 Number of inequality constraint evaluations = 62 Number of equality constraint Jacobian evaluations = 46 Number of inequality constraint Jacobian evaluations = 46 Number of Lagrangian Hessian evaluations = 45 Total CPU secs in IPOPT (w/o function evaluations) = 0.314 Total CPU secs in NLP function evaluations = 0.099 EXIT: Optimal Solution Found. ============================================= The same log for a case when it returned infeasible looks like this: Number of nonzeros in equality constraint Jacobian...: 7005 Number of nonzeros in inequality constraint Jacobian.: 1160 Number of nonzeros in Lagrangian Hessian.............: 30381 Total number of variables............................: 495 variables with only lower bounds: 0 variables with lower and upper bounds: 415 variables with only upper bounds: 0 Total number of equality constraints.................: 168 Total number of inequality constraints...............: 338 inequality constraints with only lower bounds: 0 inequality constraints with lower and upper bounds: 3 inequality constraints with only upper bounds: 335 iter objective inf_pr inf_du lg(mu) ||d|| lg(rg) alpha_du alpha_pr ls 0 -9.3890030e-01 7.40e-01 9.53e-01 -1.0 0.00e+00 - 0.00e+00 0.00e+00 0 1 -9.4908233e-01 5.36e-01 2.55e+00 -1.0 5.55e+00 - 7.66e-02 2.75e-01f 1 2 -9.1593284e-01 3.98e-01 2.06e+00 -1.0 5.89e+00 - 3.66e-01 2.59e-01f 1 3 -8.5070790e-01 2.24e-01 1.15e+00 -1.0 8.38e+00 - 4.72e-01 4.36e-01f 1 4 -8.1124641e-01 1.31e-01 1.91e+01 -1.0 1.34e+01 - 6.67e-01 4.16e-01f 1 5 -7.8493779e-01 6.88e-02 6.00e+01 -1.0 2.41e+01 - 6.93e-01 4.75e-01f 1 6 -7.6972130e-01 3.23e-02 1.82e+02 -1.0 2.61e+01 - 1.00e+00 5.30e-01f 1 7 -7.6114851e-01 1.15e-02 3.11e+02 -1.0 1.62e+01 - 1.00e+00 6.45e-01f 1 8 -7.5882188e-01 5.71e-03 1.19e+03 -1.0 5.02e+00 - 1.00e+00 5.02e-01h 1 9 -7.5751120e-01 2.46e-03 2.50e+03 -1.0 2.42e+00 - 1.00e+00 5.70e-01h 1 iter objective inf_pr inf_du lg(mu) ||d|| lg(rg) alpha_du alpha_pr ls 10 -7.5688726e-01 9.08e-04 4.80e+03 -1.0 1.04e+00 - 1.00e+00 6.31e-01h 1 11 -7.5669715e-01 4.35e-04 1.58e+04 -1.0 4.23e-01 - 1.00e+00 5.20e-01h 1 12 -7.5658130e-01 1.47e-04 2.54e+04 -1.0 1.84e-01 - 1.00e+00 6.62e-01h 1 13 -7.5655154e-01 7.33e-05 9.92e+04 -1.0 6.91e-02 - 1.00e+00 5.02e-01h 1 14 -7.5655092e-01 7.17e-05 4.36e+05 -1.0 3.10e-02 - 1.00e+00 2.11e-02f 6 15 -7.5652721e-01 1.28e-05 1.40e+05 -1.0 3.03e-02 - 1.00e+00 8.22e-01h 1 16 -7.5652619e-01 1.04e-05 9.64e+05 -1.0 6.81e-03 - 1.00e+00 1.89e-01f 3 17 -7.5652200e-01 3.50e-07 7.27e+04 -1.0 4.40e-03 - 1.00e+00 9.66e-01h 1 18 -7.5652198e-01 2.82e-07 9.45e+06 -1.0 3.90e-04 - 1.00e+00 1.93e-01f 2 19 -7.5652189e-01 1.36e-07 1.00e-06 -1.0 1.38e-04 - 1.00e+00 1.00e+00h 1 iter objective inf_pr inf_du lg(mu) ||d|| lg(rg) alpha_du alpha_pr ls 20 -7.7059115e-01 9.26e-09 1.24e+06 -5.7 2.79e-03 - 9.10e-01 9.77e-01f 1 21 -1.1444804e+00 1.86e-08 1.42e+06 -5.7 1.16e-01 - 1.46e-01 9.82e-01f 1 22 -1.3894430e+00 1.20e-08 8.32e+05 -5.7 2.00e-01 - 4.14e-01 6.03e-01f 1 23 -1.4727326e+00 9.22e-04 4.55e+05 -5.7 1.22e-01 - 7.73e-01 3.81e-01f 1 24 -1.5185627e+00 6.16e-04 2.79e+05 -5.7 4.56e-01 - 2.53e-01 4.08e-01f 1 25 -1.5788303e+00 3.76e-04 3.42e+05 -5.7 7.67e-01 - 2.35e-01 6.53e-01f 1 26 -1.5895539e+00 2.52e-04 2.54e+05 -5.7 5.90e-01 - 2.15e-01 3.76e-01h 1 27 -1.5967084e+00 1.50e-04 2.19e+05 -5.7 3.87e-01 - 7.23e-02 4.41e-01h 1 28 -1.6041354e+00 5.51e-05 7.07e+04 -5.7 3.67e-01 - 6.61e-01 6.77e-01f 1 29 -1.6082423e+00 3.70e-05 4.81e+04 -5.7 9.03e-01 - 4.65e-01 3.66e-01f 1 iter objective inf_pr inf_du lg(mu) ||d|| lg(rg) alpha_du alpha_pr ls 30 -1.6109351e+00 2.72e-05 4.38e+04 -5.7 1.61e+00 - 4.75e-01 3.29e-01f 1 31 -1.6145529e+00 1.92e-05 1.29e+04 -5.7 2.90e+00 - 2.75e-01 5.41e-01f 1 32 -1.6164804e+00 1.27e-05 8.34e+03 -5.7 3.88e+00 - 3.56e-01 4.46e-01f 1 33 -1.6187947e+00 7.13e-06 1.34e+04 -5.7 3.66e+00 - 6.51e-01 8.08e-01f 1 34 -1.6199463e+00 1.40e-06 1.96e+04 -5.7 4.65e+00 - 7.20e-01 1.00e+00f 1 35 -1.6201434e+00 1.58e-07 2.01e-05 -5.7 1.95e+00 - 1.00e+00 1.00e+00h 1 36 -1.6206211e+00 2.79e-07 2.90e+04 -8.6 3.29e+00 - 6.13e-01 7.58e-01h 1 37 -1.6207307e+00 1.41e-07 2.88e+08 -8.6 3.35e+00 - 4.32e-03 6.10e-01f 1 38 -1.6207314e+00 1.39e-07 2.88e+08 -8.6 3.41e+00 - 7.04e-06 1.07e-02f 1 39 -1.6207461e+00 1.10e-07 2.26e+08 -8.6 3.41e+00 - 5.69e-01 2.14e-01f 1 iter objective inf_pr inf_du lg(mu) ||d|| lg(rg) alpha_du alpha_pr ls 40 -1.6207466e+00 1.09e-07 2.24e+08 -8.6 6.94e+00 - 3.39e-01 7.51e-03h 1 41 -1.6207487e+00 1.06e-07 2.17e+08 -8.6 8.85e+00 - 5.13e-03 3.33e-02f 1 42 -1.6207490e+00 1.05e-07 2.16e+08 -8.6 8.71e+00 - 2.08e-01 5.67e-03f 1 43 -1.6207964e+00 3.61e-08 5.19e+07 -8.6 9.62e+00 - 3.13e-02 7.59e-01f 1 44 -1.6208049e+00 1.95e-08 2.46e+07 -8.6 5.15e+00 - 1.31e-01 5.25e-01f 1 45 -1.6208092e+00 1.97e-08 1.21e+07 -8.6 3.96e+00 - 7.87e-03 5.08e-01f 1 46 -1.6208114e+00 1.03e-07 6.85e+06 -8.6 2.77e+00 - 1.03e-02 4.54e-01f 1 47 -1.6208122e+00 2.51e-08 4.75e+06 -8.6 1.96e+00 - 7.83e-01 3.06e-01f 1 48 -1.6208132e+00 3.56e-08 3.03e+06 -8.6 2.54e+00 - 5.64e-03 3.62e-01f 1 49 -1.6208142e+00 1.85e-08 1.34e+06 -8.6 1.71e+00 - 7.83e-01 5.59e-01f 1 iter objective inf_pr inf_du lg(mu) ||d|| lg(rg) alpha_du alpha_pr ls 50 -1.6208146e+00 1.92e-08 5.84e+05 -8.6 1.12e+00 - 3.85e-02 5.63e-01f 1 51 -1.6208148e+00 1.96e-08 2.92e+05 -8.6 4.87e-01 - 2.02e-03 5.01e-01f 1 52 -1.6208149e+00 1.42e-08 2.12e+05 -8.6 2.46e-01 - 1.00e+00 2.75e-01f 1 53 -1.6208149e+00 1.59e-08 1.52e+05 -8.6 1.84e-01 - 1.00e+00 2.84e-01f 2 54 -1.6208150e+00 1.59e-08 5.91e+04 -8.6 1.34e-01 - 1.00e+00 7.45e-01h 1 55 -1.6208150e+00 1.04e-08 5.24e+04 -8.6 3.47e-02 - 1.00e+00 3.45e-01f 2 56 -1.6208150e+00 1.96e-08 2.92e+03 -8.6 2.28e-02 - 1.00e+00 9.63e-01h 1 57 -1.6208150e+00 1.96e-08 1.61e+04 -8.6 8.33e-04 - 2.26e-01 5.25e-02f 2 58 -1.6208150e+00 1.98e-08 4.22e+04 -8.6 7.91e-04 - 1.00e+00 4.79e-01h 1 59 -1.6208150e+00 2.00e-08 7.54e-10 -8.6 4.12e-04 - 1.00e+00 1.00e+00f 1 iter objective inf_pr inf_du lg(mu) ||d|| lg(rg) alpha_du alpha_pr ls 60 -1.6208157e+00 2.00e-08 8.13e-10 -9.0 2.67e-01 - 1.00e+00 1.00e+00h 1 61 -1.6208157e+00 2.00e-08 4.66e-10 -9.0 2.38e-03 - 1.00e+00 1.00e+00h 1 62 -1.6208157e+00 2.00e-08 1.00e+05 -9.0 1.72e-06 - 1.00e+00 5.00e-01h 2 63 -1.6208157e+00 2.00e-08 8.00e+04 -9.0 8.57e-07 - 1.00e+00 1.53e-05h 17 64 -1.6208157e+00 2.00e-08 6.30e-10 -9.0 8.57e-07 - 1.00e+00 1.00e+00h 1 65 -1.6208157e+00 2.00e-08 2.00e+05 -9.0 6.72e-09 - 1.00e+00 3.05e-05h 16 66 -1.6208157e+00 2.00e-08 6.32e-10 -9.0 2.50e-09 - 1.00e+00 1.00e+00s 22 67r-1.6208157e+00 2.00e-08 1.00e+03 -7.7 0.00e+00 - 0.00e+00 0.00e+00R 1 68r-1.6198258e+00 1.19e-04 1.72e+03 -7.7 1.24e+00 - 6.06e-02 4.06e-01f 1 69r-1.6194585e+00 2.46e-05 1.34e+03 -7.7 3.39e-02 - 3.86e-01 1.00e+00h 1 iter objective inf_pr inf_du lg(mu) ||d|| lg(rg) alpha_du alpha_pr ls 70r-1.6186508e+00 1.61e-05 1.00e+03 -7.7 3.94e-02 - 1.33e-01 4.69e-01h 1 71r-1.6186091e+00 1.51e-05 9.96e+02 -7.7 4.23e-02 - 1.58e-03 6.37e-02h 1 72r-1.6186075e+00 1.50e-05 9.94e+02 -7.7 4.21e-02 - 2.29e-01 2.56e-03h 1 73r-1.6180237e+00 1.17e-06 7.16e+02 -7.7 5.32e-02 - 8.39e-02 9.68e-01h 1 74r-1.6180235e+00 7.89e-07 7.09e+02 -7.7 5.75e-02 - 5.83e-02 3.27e-01h 1 75r-1.6180235e+00 7.88e-07 7.06e+02 -7.7 5.50e-02 - 1.44e-01 1.43e-03h 1 76r-1.6180120e+00 4.87e-07 4.95e+02 -7.7 6.17e-02 - 7.75e-01 3.84e-01h 1 77r-1.6179092e+00 2.25e-08 7.48e+01 -7.7 1.73e-01 - 6.18e-01 1.00e+00h 1 78r-1.6177525e+00 9.42e-07 4.18e+01 -7.7 1.49e-01 - 4.62e-01 1.00e+00h 1 79r-1.6171353e+00 4.51e-06 1.51e+01 -7.7 3.03e-02 - 6.51e-01 1.00e+00h 1 iter objective inf_pr inf_du lg(mu) ||d|| lg(rg) alpha_du alpha_pr ls 80r-1.6170959e+00 2.14e-05 6.06e+00 -7.7 5.37e-02 - 5.99e-01 6.30e-01h 1 81r-1.6150629e+00 5.38e-06 2.93e+00 -7.7 5.03e-02 - 5.17e-01 1.00e+00h 1 82r-1.6114684e+00 5.40e-07 1.23e+00 -7.7 2.22e-02 - 5.81e-01 1.00e+00h 1 83r-1.6052244e+00 2.72e-08 4.14e-01 -7.7 1.26e-02 - 6.62e-01 1.00e+00h 1 84r-1.5893649e+00 4.10e-06 2.02e-01 -7.7 1.89e-02 - 5.12e-01 1.00e+00h 1 85r-1.5645076e+00 5.53e-13 9.04e-02 -7.7 1.96e-02 - 5.53e-01 1.00e+00h 1 86r-1.5302150e+00 5.53e-13 1.17e-02 -7.7 5.61e-02 - 8.71e-01 5.14e-01h 1 87r-1.3986112e+00 5.53e-13 4.49e-03 -7.7 3.74e-02 - 6.17e-01 1.00e+00h 1 88r-1.4164161e+00 5.53e-13 4.06e-08 -7.7 6.95e-03 - 1.00e+00 1.00e+00h 1 89r-1.4163084e+00 5.53e-13 6.90e-10 -7.7 9.47e-04 - 1.00e+00 1.00e+00h 1 iter objective inf_pr inf_du lg(mu) ||d|| lg(rg) alpha_du alpha_pr ls 90r-1.5153375e+00 2.58e-12 6.07e+00 -11.0 4.54e-01 - 5.47e-01 9.87e-01h 1 91r-1.5164054e+00 1.68e-12 3.28e-01 -11.0 7.02e-03 - 9.48e-01 1.00e+00h 1 92r-1.5338332e+00 4.22e-15 8.28e-02 -11.0 8.05e-02 - 7.47e-01 1.00e+00h 1 93r-1.5584530e+00 8.88e-16 2.05e-02 -11.0 1.08e-01 - 7.53e-01 1.00e+00h 1 Number of Iterations....: 93 (scaled) (unscaled) Objective...............: -1.5777388885605270e+00 -1.5777388885605270e+00 Dual infeasibility......: 3.1308350614807106e+00 3.1308350614807106e+00 Constraint violation....: 2.2204460492503131e-16 2.2204460492503131e-16 Complementarity.........: 3.1172753702342579e-11 3.1172753702342579e-11 Overall NLP error.......: 3.1308350614807106e+00 3.1308350614807106e+00 Number of objective function evaluations = 164 Number of objective gradient evaluations = 69 Number of equality constraint evaluations = 164 Number of inequality constraint evaluations = 164 Number of equality constraint Jacobian evaluations = 96 Number of inequality constraint Jacobian evaluations = 96 Number of Lagrangian Hessian evaluations = 94 Total CPU secs in IPOPT (w/o function evaluations) = 5.276 Total CPU secs in NLP function evaluations = 0.356 Total CPU secs in NLP function evaluations = 0.356 EXIT: Converged to a point of local infeasibility. Problem may be infeasible. It seems that up to iteration 36 things are moving relatively close (objective value and inf_pr are different up to something I can attribute to the different setup). The inf_du however seem quite different even before that, but at this iteration the two runs start to diverge because the consistent run is actually reducing it while on the variable run it spikes compared to the previous iteration and this is (maybe?) the reason for it to continue and ultimately end in a locally infeasible point. I am pretty confident that I am doing something wrong, but not really able to figure out what. Any clues on what may be going on ? Thanks a lot in advance for any insights ! -------------- next part -------------- An HTML attachment was scrubbed... URL: