[Ipopt] EXIT: Restoration Failed!

[Stefano Avallone]

Hello all,

I am trying to solve a convex optimization problem (nonlinear convex objective 
function, linear constraints), but I get a "restoration failed" on exit. 
Please find below an excerpt of the output with the highest print level. I 
would really appreciate it if you could give me any insight how to solve the 
problem.

Best regards and many thanks in advance,
Stefano

-----------------------------------------------------------------------------------

This is Ipopt version 3.9trunk, running with linear solver ma27.

Number of nonzeros in equality constraint Jacobian...:       40
Number of nonzeros in inequality constraint Jacobian.:       20
Number of nonzeros in Lagrangian Hessian.............:      210

Total number of variables............................:       20
                     variables with only lower bounds:        0
                variables with lower and upper bounds:       20
                     variables with only upper bounds:        0
Total number of equality constraints.................:       14
Total number of inequality constraints...............:       20
        inequality constraints with only lower bounds:        0
   inequality constraints with lower and upper bounds:       20
        inequality constraints with only upper bounds:        0

[...]

**************************************************
*** Finding Acceptable Trial Point for Iteration 48:
**************************************************

--> Starting line search in iteration 48 <--
Acceptable Check:
  overall_error =  6.2432661230067832e-01   acceptable_tol_             =  
9.9999999999999995e-07
  dual_inf      =  6.2432661230067832e-01   acceptable_dual_inf_tol_    =  
1.0000000000000000e+10
  constr_viol   =  4.4295999887965323e-11   acceptable_constr_viol_tol_ =  
1.0000000000000000e-02
  compl_inf     =  9.2240643020082419e-07   acceptable_compl_inf_tol_   =  
1.0000000000000000e-02
  curr_obj_val_ =  9.1052327487239893e+02   last_obj_val                =  
9.1052327481317286e+02
  fabs(curr_obj_val_-last_obj_val_)/Max(1., fabs(curr_obj_val_)) =  
6.5046187439939525e-11 acceptable_obj_change_tol_ =  1.0000000000000000e+20
test iter = 48
The current filter has 8 entries.
                phi                    theta            iter
    1  9.1052354071339846e+02  3.0670202228995563e-10    29
    2  9.1052354072776302e+02  2.1724936776828671e-15    39
    3  9.1052354084525871e+02  8.8596425344253528e-16    41
    4  9.1052354084525996e+02  8.8596398874738628e-16    42
    5  9.1052354108485770e+02  6.5761296596426808e-16    43
    6  9.1052354083602870e+02  2.0631604761870023e-15    44
    7  9.1052354084260230e+02  1.9546522963837980e-15    45
    8  9.1052354084260378e+02  1.9546519952930661e-15    46
Relative step size for delta_x = 9.894576e-09
Trying soft restoration phase step with step length  1.000000e+00
Checking acceptability for trial step size alpha_primal_test= 0.000000e+00:
  New values of barrier function     =  9.1052354091528207e+02  (reference  
9.1052354091522284e+02):
  New values of constraint violation =  2.0036811308125425e-10  (reference  
1.8173182417423936e-10):
Checking sufficient reduction...
Failed...
  Primal-dual error at current point:   5.0020802340580574e-02
  Primal-dual error at trial point  :   2.7778642326579883e-01
  Trial step rejected.
Acceptable Check:
  overall_error =  6.2432661230067832e-01   acceptable_tol_             =  
9.9999999999999995e-07
  dual_inf      =  6.2432661230067832e-01   acceptable_dual_inf_tol_    =  
1.0000000000000000e+10
  constr_viol   =  4.4295999887965323e-11   acceptable_constr_viol_tol_ =  
1.0000000000000000e-02
  compl_inf     =  9.2240643020082419e-07   acceptable_compl_inf_tol_   =  
1.0000000000000000e-02
  curr_obj_val_ =  9.1052327487239893e+02   last_obj_val                =  
9.1052327481317286e+02
  fabs(curr_obj_val_-last_obj_val_)/Max(1., fabs(curr_obj_val_)) =  
6.5046187439939525e-11 acceptable_obj_change_tol_ =  1.0000000000000000e+20
test iter = 48
Restoration phase is called at point that is almost feasible,
  with constraint violation 1.817318e-10. Abort.
Exception of type: RESTORATION_FAILED in file 
"../../../../Ipopt/src/Algorithm/IpBacktrackingLineSearch.cpp" at line 562:
 Exception message: Restoration phase called, but point is almost feasible.

Number of Iterations....: 48

                                   (scaled)                 (unscaled)
Objective...............:   9.1052327487239893e+02    9.1052327487239893e+02
Dual infeasibility......:   6.2432661230067832e-01    6.2432661230067832e-01
Constraint violation....:   4.4295999887965323e-11    4.4295999887965323e-11
Complementarity.........:   9.2240643020082419e-07    9.2240643020082419e-07
Overall NLP error.......:   6.2432661230067832e-01    6.2432661230067832e-01