[Ipopt] How to detect which constraints not satisfying Linear Independence Constraint Qualification?

[James Henderson]

Hi Andreas, 

Thank you for your reply.
But in what situation 
the algorism will be failure like this?
Where or how to find the 
source of error come from?
I really want to identify the source of 
problem, and make some improvement on it.
And one more question, do 
there exist any method to detect which constraints not satisfying LICQ, 
or to reduce the degeneracy situation(because I have so many 
constraints)?
Thank you so much

James


----- 
Original Message ----
From: Andreas Waechter <andreasw at watson.ibm.com>
To: James Henderson <jamesh125125 at yahoo.com>
Cc: ipopt ipopt <ipopt at list.coin-or.org>
Sent: Sat, February 27, 2010 6:07:05 AM
Subject: Re: [Ipopt] How to detect which constraints not satisfying Linear Independence Constraint 
Qualification?

Hi James,

I noticed in your output that 
your restoration phase failure is actually 
for the restoration phase inside the restoration phase.  The restoration 
phase problem by 
construction satisfies LICQ, so this cannot really be the 
issue.  
Maybe your constraint function are just very nasty (I assume you 
verified that the derivative checker does not compain, and that there are 
no "if" statements in the computation of your functions that might make 
the problem non-smooth).  Beside suggesting to try to formulate the 
constraints in a different way, I don't know what to propose.  You might 
want 
to check if other NLP solvers can solve your problem.

Regards,

Andreas

On Tue, 9 Feb 2010, James Henderson wrote:

> Dear all,
>
> I have many constraint functions, but I suppose they are all linear 
independent, but there still have a failed restoration.
> Does 
anyone know how to detect which constraints not satisfying Linear 
Independence Constraint Qualification?
> Thank you.
>
> James
>
> This is Ipopt version 3.8.1, running with linear 
solver ma27.
>
> Number of nonzeros in equality constraint 
Jacobian...:     9672
> Number of nonzeros in inequality 
constraint Jacobian.:    38340
> Number of nonzeros in Lagrangian 
Hessian.............:        0
>
> Total number of 
variables............................:     2136
>                     variables with only lower bounds:        0
>                
variables with lower and upper bounds:        0
>                     variables with only upper bounds:        0
> Total number of 
equality constraints.................:      806
> Total number of 
inequality constraints...............:     4260
>        
inequality constraints with only lower bounds:        0
>   
inequality constraints with lower and upper bounds:        0
>        inequality constraints with only upper bounds:     4260
>
> iter    objective    inf_pr   inf_du lg(mu)  ||d||  lg(rg) alpha_du 
alpha_pr  ls
>   0 1.8076131e+004 8.53e-003 2.21e+001   0.0 
0.00e+000    -  0.00e+000 0.00e+000   0
>   1 6.1451386e+002 
5.88e-001 8.75e+000  -0.2 2.22e+001    -  4.90e-001 4.18e-001f  1
> ...........
> 1380r4.1766689e+002 9.97e-001 1.21e-001  -4.9 
8.64e-001    -  1.00e+000 1.00e+000h  1
> 1381r4.1617759e+002 
9.50e-001 1.40e+000  -4.9 6.68e-001    -  1.00e+000 1.00e+000h  1
> Restoration phase converged to a feasible point that is
> 
unacceptable to the filter for the original problem.
> Restoration phase in the restoration phase failed.
>
> Number of 
Iterations....: 1381
>
>                                   
(scaled)                 (unscaled)
> Objective...............:  
4.1431717107537929e+002   4.1431717107537929e+002
> Dual 
infeasibility......:  4.7797490421788211e+000   4.7797490421788211e+000
> Constraint violation....:  4.0595912702112713e-005   
4.0595912702112713e-005
> Complementarity.........:  
7.7917756206222183e-006   7.7917756206222183e-006
> Overall NLP 
error.......:  4.7797490421788211e+000   4.7797490421788211e+000
>
>
> Number of objective function evaluations             = 2855
> 
Number of objective gradient evaluations             = 196
> 
Number of equality constraint evaluations            = 2856
> 
Number of inequality constraint evaluations          = 2856
> 
Number of equality constraint Jacobian evaluations   = 1386
> 
Number of inequality constraint Jacobian evaluations = 1386
> 
Number of Lagrangian Hessian evaluations             = 0
> Total 
CPU secs in IPOPT (w/o function evaluations)   =    281.311
> 
Total CPU secs in NLP function evaluations           =     11.907
>
> EXIT: Restoration Failed!
>
>
>
>
>
> _______________________________________________
> Ipopt mailing 
list
> Ipopt at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/ipopt
>
>