Convex.jl set some configurations e.g. maximum iteration number, etc

Specific Domains Optimization (Mathematical)
1

I would like to change some configurations of Convex.jl optimiser.
I took a look at the documentation but I didn’t understand how to do it.
Can anybody help me?

Note that I’m mostly using Mosek.

2

I’ve opened a PR to improve things in Convex.jl Use MOI.instantiate to create a new optimizer by odow · Pull Request #431 · jump-dev/Convex.jl · GitHub.

For now, you need to do something like

function new_optimizer()
    model = Mosek.Optimizer()
    MOI.set(model, MOI.RawParameter("QUIET"), false)
    MOI.set(model, MOI.RawParameter("INTPNT_CO_TOL_DFEAS"), 1e-7)
    return model
end
solve!(p, new_optimizer)
1 Like
3

Thanks!
BTW, what will be changed if the PR is merged?

4

You’ll be able to go

solve!(
    p, 
    MOI.OptimizerWithAttributes(
        Mosek.Optimizer, 
        "QUIET" => false,
        "INTPNT_CO_TOL_DFEAS" => 1e-7,
    ),
)

The docs should be updated as well.

1 Like