SymNLPModels.jl

NLPModels.jl adapter for Symbolics.jl

Univariate Concave Example

Minimize the function (x - 3)^2 subject to x^2 ≤ 1 using Ipopt.

using SymNLPModels: SymNLPModel
using NLPModelsIpopt: ipopt
using Symbolics

@variables x

objective = (x - 3)^2
constraints = [x^2 ≲ 1]

model = SymNLPModel(objective, constraints)
stats = ipopt(model; tol=1e-4, print_level=0)

actual = stats.solution

Multivariate Concave Example

Provide variables explicitly in the model, or use parse_solution and value if variable order is important.

using SymNLPModels: SymNLPModel, value, parse_solution
using NLPModelsIpopt: ipopt
using Symbolics

@variables x[1:5]
X = Symbolics.scalarize(x)
center = randn(length(X))

objective = sum((X .- center).^2)
constraints = X.^2 .≲ 1
tol = 1e-4

model = SymNLPModel(objective, constraints;)
stats = ipopt(model; )

expected = clamp.(center, -1, 1)
solution = parse_solution(model, stats.solution)
actual = value(solution, X)   
@show isapprox(expected, actual; atol=tol)

Note

This project makes little sense as you could use ModelingToolkit with OptimizationMOI to do the same.