KNITRO + JuMP: linear --> nonlinear constraints?

[Alan Crawford]

This is probably an old issue - but i wondered why linear constraints are being converted in to nonlinear constraints when called through JuMP?

For reference i am using Julia v0.5 with Knitro 10.0.1. Also 

julia> Pkg.status("JuMP")

 - JuMP                          0.14.1

julia> Pkg.status("KNITRO")

 - KNITRO                        0.2.0

--
You received this message because you are subscribed to the Google Groups "julia-opt" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
Visit this group at https://groups.google.com/group/julia-opt.
For more options, visit https://groups.google.com/d/optout.

[Miles Lubin]

Not sure what you mean exactly, but possibly see https://github.com/JuliaOpt/KNITRO.jl/commit/67a6e8379d1b39476282abb56a6e65598e2896c6 which hasn't been tagged yet.

On Wednesday, November 30, 2016 at 10:14:09 AM UTC-5, Alan Crawford wrote:

This is probably an old issue - but i wondered why linear constraints are being converted in to nonlinear constraints when called through JuMP?

For reference i am using Julia v0.5 with Knitro 10.0.1. Also 

julia> Pkg.status("JuMP")

 - JuMP                          0.14.1

julia> Pkg.status("KNITRO")

 - KNITRO                        0.2.0

--
You received this message because you are subscribed to the Google Groups "julia-opt" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
Visit this group at https://groups.google.com/group/julia-opt.
For more options, visit https://groups.google.com/d/optout.

[Miles Lubin]

If you use that you should update to Knitro 10.1.2 otherwise you may silently get incorrect answers on "rare" occasions.

On Wednesday, November 30, 2016 at 10:16:24 AM UTC-5, Miles Lubin wrote:

Not sure what you mean exactly, but possibly see <a href="https://github.com/JuliaOpt/KNITRO.jl/commit/67a6e8379d1b39476282abb56a6e65598e2896c6" target="_blank" rel="nofollow" onmousedown="this.href=&#39;https://www.google.com/url?q\x3dhttps%3A%2F%2Fgithub.com%2FJuliaOpt%2FKNITRO.jl%2Fcommit%2F67a6e8379d1b39476282abb56a6e65598e2896c6\x26sa\x3dD\x26sntz\x3d1\x26usg\x3dAFQjCNGTJozoTyism49lpcxYNRl8bgc74A&#39;;return true;" onclick="this.href=&#39;https://www.google.com/url?q\x3dhttps%3A%2F%2Fgithub.com%2FJuliaOpt%2FKNITRO.jl%2Fcommit%2F67a6e8379d1b39476282abb56a6e65598e2896c6\x26sa\x3dD\x26sntz\x3d1\x26usg\x3dAFQjCNGTJozoTyism49lpcxYNRl8bgc74A&#39;;return true;">https://github.com/JuliaOpt/KNITRO.jl/commit/67a6e8379d1b39476282abb56a6e65598e2896c6 which hasn't been tagged yet.

On Wednesday, November 30, 2016 at 10:14:09 AM UTC-5, Alan Crawford wrote:

This is probably an old issue - but i wondered why linear constraints are being converted in to nonlinear constraints when called through JuMP?

For reference i am using Julia v0.5 with Knitro 10.0.1. Also 

julia> Pkg.status("JuMP")

 - JuMP                          0.14.1

julia> Pkg.status("KNITRO")

 - KNITRO                        0.2.0

--
You received this message because you are subscribed to the Google Groups "julia-opt" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
Visit this group at https://groups.google.com/group/julia-opt.
For more options, visit https://groups.google.com/d/optout.

[Alan Crawford]

Thanks Miles. To clarify:

In my jump model i get the following summary information:

Minimization problem with:

 * 5000 linear constraints

 * 3784 variables

Solver is Knitro

Then when i solve the model I get that knitro output telling me i have 1000 nonlinear constraints:

=======================================

           Academic License

       (NOT FOR COMMERCIAL USE)

         Artelys Knitro 10.0.1

=======================================

Knitro presolve eliminated 0 variables and 0 constraints.

outlev:               1

Knitro changing algorithm from AUTO to 1.

Knitro changing bar_initpt from AUTO to 3.

Knitro changing bar_murule from AUTO to 4.

Knitro changing bar_penaltycons from AUTO to 1.

Knitro changing bar_penaltyrule from AUTO to 2.

Knitro changing bar_switchrule from AUTO to 2.

Knitro changing linsolver from AUTO to 5.

Problem Characteristics                    ( Presolved)

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

Objective goal:  Minimize

Number of variables:                  3784 (      3784)

    bounded below:                       0 (         0)

    bounded above:                       0 (         0)

    bounded below and above:             0 (         0)

    fixed:                               0 (         0)

    free:                             3784 (      3784)

Number of constraints:                5000 (      5000)

    linear equalities:                   0 (         0)

    nonlinear equalities:                0 (         0)

    linear inequalities:                 0 (         0)

    nonlinear inequalities:           5000 (      5000)

    range:                               0 (         0)

Number of nonzeros in Jacobian:      94750 (     94750)

Number of nonzeros in Hessian:     7161220 (   7161220)

EXIT: Locally optimal solution found.

Final Statistics

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

Final objective value               =  -3.50504042566882e-01

Final feasibility error (abs / rel) =   9.60e-11 / 9.60e-11

Final optimality error  (abs / rel) =   2.64e-10 / 2.64e-10

# of iterations                     =         19 

# of CG iterations                  =          0 

# of function evaluations           =         21

# of gradient evaluations           =         21

# of Hessian evaluations            =         19

Total program time (secs)           =     170.44518 (   170.338 CPU time)

Time spent in evaluations (secs)    =     106.45541

===============================================================================

 

On Wednesday, 30 November 2016 15:18:11 UTC, Miles Lubin wrote:

If you use that you should update to Knitro 10.1.2 otherwise you may silently get incorrect answers on "rare" occasions.

On Wednesday, November 30, 2016 at 10:16:24 AM UTC-5, Miles Lubin wrote:

Not sure what you mean exactly, but possibly see <a href="https://github.com/JuliaOpt/KNITRO.jl/commit/67a6e8379d1b39476282abb56a6e65598e2896c6" rel="nofollow" target="_blank" onmousedown="this.href=&#39;https://www.google.com/url?q\x3dhttps%3A%2F%2Fgithub.com%2FJuliaOpt%2FKNITRO.jl%2Fcommit%2F67a6e8379d1b39476282abb56a6e65598e2896c6\x26sa\x3dD\x26sntz\x3d1\x26usg\x3dAFQjCNGTJozoTyism49lpcxYNRl8bgc74A&#39;;return true;" onclick="this.href=&#39;https://www.google.com/url?q\x3dhttps%3A%2F%2Fgithub.com%2FJuliaOpt%2FKNITRO.jl%2Fcommit%2F67a6e8379d1b39476282abb56a6e65598e2896c6\x26sa\x3dD\x26sntz\x3d1\x26usg\x3dAFQjCNGTJozoTyism49lpcxYNRl8bgc74A&#39;;return true;">https://github.com/JuliaOpt/KNITRO.jl/commit/67a6e8379d1b39476282abb56a6e65598e2896c6 which hasn't been tagged yet.

On Wednesday, November 30, 2016 at 10:14:09 AM UTC-5, Alan Crawford wrote:

This is probably an old issue - but i wondered why linear constraints are being converted in to nonlinear constraints when called through JuMP?

For reference i am using Julia v0.5 with Knitro 10.0.1. Also 

julia> Pkg.status("JuMP")

 - JuMP                          0.14.1

julia> Pkg.status("KNITRO")

 - KNITRO                        0.2.0

--
You received this message because you are subscribed to the Google Groups "julia-opt" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
Visit this group at https://groups.google.com/group/julia-opt.
For more options, visit https://groups.google.com/d/optout.

[Alan Crawford]

I checked out the master of KNITRO.jl and downloaded the KNITRO v 10.1.2, I ran the example from KNITRO.jl 

using KNITRO, JuMP
m = Model(solver=KnitroSolver())
@variable(m, x[1:3]>=0)
@objective(m, Min, 9.0 - 8.0*x[1] - 6.0*x[2] - 4.0*x[3]
                        + 2.0*x[1]^2 + 2.0*x[2]^2 + x[3]^2
                        + 2.0*x[1]*x[2] + 2.0*x[1]*x[3])
@constraint(m, x[1] + x[2] + 2.0*x[3] <= 3)
solve(m)

KNITRO now correctly picks up problem structure. 

 

On Wednesday, 30 November 2016 15:14:09 UTC, Alan Crawford wrote:

This is probably an old issue - but i wondered why linear constraints are being converted in to nonlinear constraints when called through JuMP?

For reference i am using Julia v0.5 with Knitro 10.0.1. Also 

julia> Pkg.status("JuMP")

 - JuMP                          0.14.1

julia> Pkg.status("KNITRO")

 - KNITRO                        0.2.0

--
You received this message because you are subscribed to the Google Groups "julia-opt" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
Visit this group at https://groups.google.com/group/julia-opt.
For more options, visit https://groups.google.com/d/optout.