JuMP: Multiplicative parameters for LP

Specific Domains Optimization (Mathematical)
1

I’m trying to rework how “model parameters” work in my model. I know that

using JuMP
import HiGHS

model = direct_model(HiGHS.Optimizer())

p = @variable(model)
fix(p, 1.0; force=true)

x = @variable(model)
@constraint(model, x >= p)

@objective(model, Min, x + 0)

optimize!(model)
fix(p, 10.0; force=true)
optimize!(model)

works for a “somewhat low” overhead parametric RHS.

On the other hand, p * x obviously creates a QuadExpr, which prevents using it for the LHS. Is there any similar trick that comes at less performance loss than going for ParametricOptInterface.jl or ParameterJuMP.jl?

2

You may be interested in this part of the JuMP manual: Modify a variable coefficient.

1 Like
3

Thanks @jd-foster! Unfortunately I know about this, and while it is the “perfect solution” from a performance point of view, it is not really usable for large scale models - without significant “manual” work (that I was hoping to overcome with some other approach).

Consider the case of a variable (with my coefficient) being added to another expression that is then added - together with other expressions - to a constraint. It’s hard to find out - afterwards - what the contribution of my initial coefficient was (given the fact, that the same variable could occur in multiple of those expressions)… :frowning: Also, “knowing the variable” it is not immediately clear in which constraints it ended up.

4

No, there are no multiplicative parameters in JuMP. As you’ve seen, we’ve had two attempts at introducing them (ParameterJuMP and ParametricOptInterface, but they both have issues with performance overhead).

1 Like