Problem setting constraint in simple LP using JuMP

Specific Domains Optimization (Mathematical)
1

Hi there,

Im trying to construct a simple LP using JuMP but I am running into an issue with the constraint in the function. I checked this is the line causing the issue since if I delete it, I can define the function in the REPL but otherwise I can not (error attached below the code)

function quantile_regression_linear()
    # retrieve views of the relevant data 
    y = rand(1000)
    x = rand(1000)
    N = length(y)
    # initialize the optimization model and solve it 
    model = Model(Gurobi.Optimizer)
    set_silent(model)
    @variable(model, u_plus[1:N] ≥ 0.0)
    @variable(model, u_minus[1:N] ≤ 0.0)
    @variable(model, θ[1:2])
    @constraint(model, [i in 1:N], y[i] - θ[1] - θ[2]*x[i] = u_plus[i] + u_minus[i])
end

When defining this function in the REPL I get the following error:

ERROR: LoadError: MethodError: Cannot `convert` an object of type Expr to an object of type Symbol
The function `convert` exists, but no method is defined for this combination of argument types.

Closest candidates are:
  Symbol(::Any...)
   @ Base strings/basic.jl:229
  convert(::Type{T}, ::T) where T
   @ Base Base.jl:126

Stacktrace:
 [1] setindex!(h::Dict{Symbol, Any}, v0::Expr, key0::Expr)
   @ Base ./dict.jl:346
 [2] parse_macro_arguments(error_fn::JuMP.Containers.var"#error_fn#98"{…}, args::Tuple{…}; valid_kwargs::Nothing, num_positional_args::Nothing)
   @ JuMP.Containers ~/.julia/packages/JuMP/i68GU/src/Containers/macro.jl:58
 [3] parse_macro_arguments(error_fn::Function, args::Tuple{Symbol, Expr, Expr})
   @ JuMP.Containers ~/.julia/packages/JuMP/i68GU/src/Containers/macro.jl:41
 [4] var"@constraint"(__source__::LineNumberNode, __module__::Module, input_args::Vararg{Any})
   @ JuMP ~/.julia/packages/JuMP/i68GU/src/macros/@constraint.jl:103
in expression starting at REPL[153]:12
Some type information was truncated. Use `show(err)` to see complete types.

Thanks in advance!

2

Could it be that you just need to replace = with == in your constraint?

3

Absolutely, could not pick up this silly mistake. Thanks!

4

Absolutely, could not pick up this silly mistake

The error message didn’t help! I’ll see if we can improve it :smile:

It looks like we catch some cases but not all

julia> using JuMP

julia> begin
           model = Model()
           @variable(model, x[1:1])
           @constraint(model, x[1] = 1)
       end
ERROR: LoadError: MethodError: Cannot `convert` an object of type Expr to an object of type Symbol

Closest candidates are:
  convert(::Type{T}, ::T) where T
   @ Base Base.jl:84
  Symbol(::Any...)
   @ Base strings/basic.jl:229

Stacktrace:
 [1] setindex!(h::Dict{Symbol, Any}, v0::Int64, key0::Expr)
   @ Base ./dict.jl:367
 [2] parse_macro_arguments(error_fn::JuMP.Containers.var"#error_fn#98"{…}, args::Tuple{…}; valid_kwargs::Nothing, num_positional_args::Nothing)
   @ JuMP.Containers ~/.julia/packages/JuMP/FEKLB/src/Containers/macro.jl:58
 [3] parse_macro_arguments(error_fn::Function, args::Tuple{Symbol, Expr})
   @ JuMP.Containers ~/.julia/packages/JuMP/FEKLB/src/Containers/macro.jl:41
 [4] var"@constraint"(__source__::LineNumberNode, __module__::Module, input_args::Vararg{Any})
   @ JuMP ~/.julia/packages/JuMP/FEKLB/src/macros/@constraint.jl:103
in expression starting at REPL[137]:4
Some type information was truncated. Use `show(err)` to see complete types.

julia> begin
           model = Model()
           @variable(model, x)
           @constraint(model, x = 1)
       end
ERROR: LoadError: At REPL[138]:4: `@constraint(model, x = 1)`: No constraint expression detected. If you are trying to construct an equality constraint, use `==` instead of `=`.
Stacktrace:
5

thats great that the error message is so informative and specific in some cases! I feel like an error like mine should only slip in on a Monday (if at all) haha but better error messages are always welcomed so thanks a lot!

6

This will be fixed by Improve error message when = is used instead of == by odow · Pull Request #3892 · jump-dev/JuMP.jl · GitHub

The next release of JuMP will give:

julia> function quantile_regression_linear()
           # retrieve views of the relevant data 
           y = rand(1000)
           x = rand(1000)
           N = length(y)
           # initialize the optimization model and solve it 
           model = Model()
           set_silent(model)
           @variable(model, u_plus[1:N] ≥ 0.0)
           @variable(model, u_minus[1:N] ≤ 0.0)
           @variable(model, θ[1:2])
           @constraint(model, [i in 1:N], y[i] - θ[1] - θ[2]*x[i] = u_plus[i] + u_minus[i])
       end
ERROR: LoadError: At REPL[11]:12: `@constraint(model, [i in 1:N], (y[i] - θ[1]) - θ[2] * x[i] = begin
        #= REPL[11]:12 =#
        u_plus[i] + u_minus[i]
    end)`: Invalid keyword argument detected. If you are trying to construct an equality constraint, use `==` instead of `=`.
Stacktrace:
 [1] error(::String, ::String)
   @ Base ./error.jl:44
 [2] (::JuMP.Containers.var"#error_fn#98"{String})(str::String)
   @ JuMP.Containers ~/.julia/dev/JuMP/src/Containers/macro.jl:331
 [3] parse_macro_arguments(error_fn::JuMP.Containers.var"#error_fn#98"{…}, args::Tuple{…}; valid_kwargs::Nothing, num_positional_args::Nothing)
   @ JuMP.Containers ~/.julia/dev/JuMP/src/Containers/macro.jl:58
 [4] parse_macro_arguments(error_fn::Function, args::Tuple{Symbol, Expr, Expr})
   @ JuMP.Containers ~/.julia/dev/JuMP/src/Containers/macro.jl:41
 [5] var"@constraint"(__source__::LineNumberNode, __module__::Module, input_args::Vararg{Any})
   @ JuMP ~/.julia/dev/JuMP/src/macros/@constraint.jl:103
in expression starting at REPL[11]:12
Some type information was truncated. Use `show(err)` to see complete types.