[Ipopt] dead loop problem in Fortran version

[Andreas Waechter]

Hi Yu Yanshan

We have stopped supporting the Fortran version of Ipopt (2.*) a few years 
ago - any use is on your own "risk".  You may want to try the newer 
version.

Regards

Andreas

On Wed, 28 Apr 2010, [GBK] ÓàÑßɽ wrote:

> Hi all,
>
> I am encountering a problem when i use ipot(fortran version). The solver succeeds with some initial guess values,but couldn't stop with some other initial values, even IMAXCPUSEC been set to 300. I think there is a dead loop somewhere, the subroutine CHECK_STOP checking the cpu time missed the bug. Where should I add check_stop to find the dead loop? Any suggestion would be appreciated!
>
> Some basic information about my optimization is as follows:
>
>      parameter  (N = 2041, M = 1)
>      parameter  (NLB = 1, NUB = 1)
>      real*8 DAT(1)
>      integer IDAT(1)
>      external EV_F, EV_C, EV_G, EV_A
>      external EV_H_DUMMY, EV_HLV_DUMMY, EV_HOV_DUMMY, EV_HCV_DUMMY
>
>      data ILB    /   2041  /
>      data BNDS_L /  0.0d0 /
>      data IUB    /   2041 /
>      data BNDS_U / 1.0d0 /
>
>      NARGS = 1
>      ARGS(1) = 1.d-8
>      CARGS(1) = 'dtol'
>      DAT(1) = 1.d0
>      fillinfact = 10.d0
>
> The constraints are: x(1)^2+x(2)^2+.....+x(2040)^2+x(2041)=1, 0<x(2041)<1. Note that x(2041) is a slave variable.
>
>
> The parameters in PARAMS.DAT are set as follows:
> dtoL , 1d-8
> imerit 4
> iquasi 6
> iprint 10
> imaxiter 50
> IMAXCPUSEC 300
> ifile 1
>
>
> Best regards
>
>         Yu Yanshan
>
> _______________________________________________
> Ipopt mailing list
> Ipopt at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/ipopt
>
>