Ipopt has trouble with `Vector{AffExpr} in MOI.Zeros`

Specific Domains Optimization (Mathematical)
,
1

I have a JuMP model where solving with HiGHS and Ipopt results in different objective_values.

A JuMP Model
├ solver: Ipopt
├ objective_sense: MIN_SENSE
│ └ objective_function_type: AffExpr
├ num_variables: 54
├ num_constraints: 98
│ ├ AffExpr in MOI.EqualTo{Float64}: 34
│ ├ AffExpr in MOI.GreaterThan{Float64}: 10
│ ├ Vector{AffExpr} in MOI.Zeros: 6
│ ├ VariableRef in MOI.EqualTo{Float64}: 2
│ ├ VariableRef in MOI.GreaterThan{Float64}: 26
│ └ VariableRef in MOI.LessThan{Float64}: 20
└ Names registered in the model

julia> set_optimizer(ldr, HiGHS.Optimizer)
julia> optimize!(ldr)
julia> objective_value(ldr)
119250.0

julia> set_optimizer(ldr, Ipopt.Optimizer)
julia> optimize!(ldr)
julia> objective_value(ldr)
136249.99968899257

Digging a little bit, if I replaced the constraints Vector{AffExpr} in MOI.Zeros: 6 by an explicit JuMP.fix, the issue goes away.

2

Do you have a reproducible example?