Evaluate a nonlinear expression at a point

Specific Domains Optimization (Mathematical)
1

I have a model with many nonlinear expressions and I would like to evaluate these to test my model.

For example I have something like this

using JuMP
m = Model()

@variable(m,x)

@NLexpression(m,T,x^2)

#What is T(x=2)?

I’m writing an MCP model using the Complementarity package and a really useful model debugging technique is to evaluate the equations (NLexpressions in this case) at a solution point.

Any advice would be greatly appreciated.

1 Like
2

Continue with:

value(v -> Dict(x => 2.0)[v], T)

The first argument takes a JuMP variable name, and returns a value. So essentially the Dict can contain several variables and other ways are legit, as long as they return a value.

1 Like
3

I had the arguments in the value backwards.

Thank you so much.

1 Like
4

Actually, you are right, they are flipped. Had to jump to source to figure this out.

5

We try to follow the Julia style guide: Style Guide · The Julia Language

It suggests that functions should always come as the first argument. The main reason is to allow the do syntax like:

value(T) do v
    return v == x ? 2.0 : 0.0
end