Ipopt crashes on this problem on OSX 64bit: <div class="snippet-clipboard-content

I hade a more complicated problem which gave me the error: <div class="snippet-cli

Ipopt crash on a multiplication constraint about ipopt.jl HOT 3 CLOSED

jump-dev commented on August 16, 2024

Ipopt crash on a multiplication constraint

from ipopt.jl.

Comments (3)

ali5h commented on August 16, 2024

I hade a more complicated problem which gave me the error:

ERROR: type: eval_g!: in typeassert, expected Float64, got GenericAffExpr{Float64,Variable}
 in eval_g! at /.../.julia/v0.3/ReverseDiffSparse/src/exprlist.jl:130
 in eval_g at /.../.julia/v0.3/JuMP/src/nlp.jl:274
 in initialize at /.../.julia/v0.3/JuMP/src/nlp.jl:203
 in loadnonlinearproblem! at /.../.julia/v0.3/Ipopt/src/IpoptSolverInterface.jl:269
 in solvenlp at /.../.julia/v0.3/JuMP/src/nlp.jl:474
 in solve at /.../.julia/v0.3/JuMP/src/solvers.jl:9
 in include at /usr/local/Cellar/julia/0.3.5_1/lib/julia/sys.dylib
 in include_from_node1 at /usr/local/Cellar/julia/0.3.5_1/lib/julia/sys.dylib
while loading /.../opt.jl, in expression starting on line 106

In that problem a and b in the above formula where affine expressions.

from ipopt.jl.

mlubin commented on August 16, 2024

It's a known issue that you can't use affine expression objects inside @addNLConstraint. You can easily work around this by defining a new variable and setting it equal to the affine expression, e.g.,

@defVar(m, affvar)
@addConstraint(m, affvar == aff)
@addNLConstraint(m, cos(exp(affvar)) <= 0.5)

The first issue is likely because you're using homebrew's version of ipopt, which isn't supported in Julia. If you don't need it for other reasons, I'd just uninstall it and let Julia set up its own version with Pkg.build("Ipopt").

from ipopt.jl.

ali5h commented on August 16, 2024

Thanks, problem resolved

from ipopt.jl.

Recommend Projects