Nonconvex.jl

Nonconvex.jl copied to clipboard

julia> using JuMP

julia> import Nonconvex

julia> Nonconvex.@load Ipopt
[ Info: Attempting to load the package NonconvexIpopt.
[ Info: Loading succesful.

julia> model = JuMP.Model();

julia> @variable(model, x >= 0)
x

julia> @objective(model, Min, x^2)
x²

julia> ncvx_model = DictModel(model)
ERROR: AssertionError: obj isa AffExpr
Stacktrace:
 [1] get_objective_info(model::JuMP.Model, nvars::Int64)
   @ NonconvexCore ~/.julia/packages/NonconvexCore/YjDSM/src/models/jump.jl:184
 [2] DictModel(model::JuMP.Model)
   @ NonconvexCore ~/.julia/packages/NonconvexCore/YjDSM/src/models/jump.jl:204
 [3] top-level scope
   @ REPL[7]:1

cc @ccoffrin

Adding an epigraph variable also fails:

julia> using JuMP

julia> import Nonconvex

julia> Nonconvex.@load Ipopt
[ Info: Attempting to load the package NonconvexIpopt.
[ Info: Loading succesful.

julia> model = JuMP.Model();

julia> @variable(model, x >= 0)
x

julia> @variable(model, t)
t

julia> @constraint(model, t >= x^2)
-x² + t ≥ 0.0

julia> @objective(model, Min, t)
t

julia> ncvx_model = DictModel(model)
ERROR: AssertionError: obj isa AffExpr
Stacktrace:
 [1] get_objective_info(model::JuMP.Model, nvars::Int64)
   @ NonconvexCore ~/.julia/packages/NonconvexCore/YjDSM/src/models/jump.jl:184
 [2] DictModel(model::JuMP.Model)
   @ NonconvexCore ~/.julia/packages/NonconvexCore/YjDSM/src/models/jump.jl:204
 [3] top-level scope
   @ REPL[17]:1

Yes currently I only support linear expressions because this is the main utility of using JuMP over Nonconvex directly to define objectives or constraints. Extending this is possible but it will need to add support for more expression types in https://github.com/JuliaNonconvex/NonconvexCore.jl/blob/master/src/models/jump.jl.