Warning invalid warm start GLPK

Specific Domains Optimization (Mathematical)
,
1

Hi,

I see some topics dated from 2019 talking about the warm start in JuMP with GLPK. I still have the warning :

Warning: MathOptInterface.VariablePrimalStart() is not supported by MathOptInterface.Bridges.LazyBridgeOptimizer{GLPK.Optimizer}.

when using set_start_value
I don’t understand since the merged seems to have been done (see https://github.com/jump-dev/GLPK.jl/pull/101 )

I am using JuMP v.0.21.3, GLPK v.0.13.0 and GLPKMAthProgInterface v.0.5.0

Someone knows if there is another way to fill a start value to the solver ?

Thanks for your help.
Erwan

2

Please provide a reproducible example. GLPKMathProgInterface is deprecated. You should just use GLPK.Optimizer.

3

Here’s an example of a warning when using set_start_value with GLPK.

┌ Warning: MathOptInterface.VariablePrimalStart() is not supported by MathOptInterface.Bridges.LazyBridgeOptimizer{GLPK.Optimizer}. This information will be discarded.
└ @ MathOptInterface.Utilities C:\Users\plowman\.julia\packages\MathOptInterface\5WwpK\src\Utilities\copy.jl:285

Julia 1.5.3, Julia 1.6.0-rc1
JuMP v0.21.6
GLPK v0.14.7

using JuMP, GLPK

model = Model(GLPK.Optimizer)

x = [1,2,3]
@variable(model, w[1:3] >= 1, Int)
set_start_value.(w, [2,2,6])
start_value.(w)

@constraint(model, sum(w) == 10)
@objective(model, Min, sum(x[i] * w[i] for i in 1:3))
optimize!(model)

println("Objective value: ", objective_value(model))
println("Weights: ", value.(model[:w]))
4

Looks like I forgot to update this thread.

The warning is correct. GLPK does not support warm starts.

1 Like