[Coin-ipopt] "strange" behavior with Ipopt

[Edson Cordeiro do Valle]

          Hello
  I have been with a "strange" behavior with Ipopt. I have 3 groups of 
inequality constraints, they are grouped because they have similar 
equations. When I change the order which they are added to the 
optimization problem, the solver converges to different optima (I have 
found 2 of them). In one of the permutations, the problem reaches the 
maximum number of iterations and does not converge.
My question is: Is this behavior normal or, based on the algorithm, it 
is possible to happen?
Below are some information of my problem:


List of user-set options:

                                    Name   Value                used
                acceptable_compl_inf_tol = 0.01                  yes
              acceptable_constr_viol_tol = 0.01                  yes
                 acceptable_dual_inf_tol = 1e+10                 yes
                          acceptable_tol = 1e-06                 yes
                      barrier_tol_factor = 10                    yes
                      bound_relax_factor = 1e-08                 yes
                           compl_inf_tol = 0.0001                yes
                         constr_viol_tol = 0.0001                yes
                         derivative_test = first-order           yes
               derivative_test_print_all = no                    yes
                     derivative_test_tol = 0.0001                yes
                            dual_inf_tol = 0.0001                yes
                        file_print_level = 5                     yes
                   hessian_approximation = limited-memory        yes
              limited_memory_max_history = 6                     yes
                           linear_solver = mumps                 yes
                                max_iter = 10000                 yes
                             mu_strategy = monotone              yes
                      obj_scaling_factor = 1                     yes
                             output_file = 
recovropt2_test4_default_options_extra_constr  yes
                             print_level = 5                     yes
                      print_user_options = yes                   yes
                                     tol = 0.0001                yes

******************************************************************************
This program contains Ipopt, a library for large-scale nonlinear 
optimization.
 Ipopt is released as open source code under the Common Public License 
(CPL).
         For more information visit http://projects.coin-or.org/Ipopt
******************************************************************************


Starting derivative checker.

* jac_g [  161,    7] = -1.0000000000000000e+00 v  ~ 
-9.9971657618880272e-01  [ 2.834e-04]
* jac_g [  167,   17] = -1.0000000000000000e+00 v  ~ 
-9.9971657618880272e-01  [ 2.834e-04]
* jac_g [  173,   27] = -1.0000000000000000e+00 v  ~ 
-9.9971657618880272e-01  [ 2.834e-04]
* jac_g [  179,   37] = -1.0000000000000000e+00 v  ~ 
-9.9971657618880272e-01  [ 2.834e-04]
* jac_g [  185,   47] = -1.0000000000000000e+00 v  ~ 
-9.9971657618880272e-01  [ 2.834e-04]
* jac_g [  187,   48] = -1.0000000000000000e+00 v  ~ 
-9.9971657618880272e-01  [ 2.834e-04]
* jac_g [  189,   49] = -1.0000000000000000e+00 v  ~ 
-9.9971657618880272e-01  [ 2.834e-04]
* jac_g [  191,   61] = -1.0000000000000000e+00 v  ~ 
-9.9971657618880272e-01  [ 2.834e-04]
* jac_g [  193,   62] = -1.0000000000000000e+00 v  ~ 
-9.9971657618880272e-01  [ 2.834e-04]
* jac_g [  195,   63] = -1.0000000000000000e+00 v  ~ 
-9.9971657618880272e-01  [ 2.834e-04]
* jac_g [  197,   75] = -1.0000000000000000e+00 v  ~ 
-9.9971657618880272e-01  [ 2.834e-04]
* jac_g [  199,   76] = -1.0000000000000000e+00 v  ~ 
-9.9971657618880272e-01  [ 2.834e-04]
* jac_g [  201,   77] = -1.0000000000000000e+00 v  ~ 
-9.9971657618880272e-01  [ 2.834e-04]
* jac_g [  205,   82] = -1.0000000000000000e+00 v  ~ 
-9.9971657618880272e-01  [ 2.834e-04]
* jac_g [  207,   83] = -1.0000000000000000e+00 v  ~ 
-9.9971657618880272e-01  [ 2.834e-04]

Derivative checker detected 15 error(s).

Number of nonzeros in equality constraint Jacobian...:      846
Number of nonzeros in inequality constraint Jacobian.:      209
Number of nonzeros in Lagrangian Hessian.............:        0

Total number of variables............................:      168
                     variables with only lower bounds:        0
                variables with lower and upper bounds:      168
                     variables with only upper bounds:        0
Total number of equality constraints.................:       96
Total number of inequality constraints...............:      209
        inequality constraints with only lower bounds:      209
   inequality constraints with lower and upper bounds:        0
        inequality constraints with only upper bounds:        0


         
                                                                 Regards
 --

                                                  Edson C. do Valle
                                                edsoncv at enq.ufrgs.br
                                                   Skype: edson.cv