GitHub

OptimizationSystem

System Constructors

ModelingToolkit.OptimizationSystem — Type
struct OptimizationSystem <: ModelingToolkit.AbstractOptimizationSystem

A scalar equation for optimization.

Fields

  • tag: A tag for the system. If two systems have the same tag, then they are structurally identical.

  • op: Objective function of the system.

  • unknowns: Unknown variables.

  • ps: Parameters.

  • var_to_name: Array variables.

  • observed: Observed equations.

  • constraints: List of constraint equations of the system.

  • name: The name of the system.

  • description: A description of the system.

  • systems: The internal systems. These are required to have unique names.

  • defaults: The default values to use when initial guess and/or parameters are not supplied in OptimizationProblem.

  • metadata: Metadata for the system, to be used by downstream packages.
  • gui_metadata: Metadata for MTK GUI.
  • namespacing: If false, then sys.x no longer performs namespacing.
  • complete: If true, denotes the model will not be modified any further.
  • index_cache: Cached data for fast symbolic indexing.
  • parent: The hierarchical parent system before simplification.
  • isscheduled

Examples

@variables x y z
@parameters a b c

obj = a * (y - x) + x * (b - z) - y + x * y - c * z
cons = [x^2 + y^2 ≲ 1]
@named os = OptimizationSystem(obj, [x, y, z], [a, b, c]; constraints = cons)
source

Composition and Accessor Functions

  • get_op(sys): The objective to be minimized.
  • get_unknowns(sys) or unknowns(sys): The set of unknowns for the optimization.
  • get_ps(sys) or parameters(sys): The parameters for the optimization.
  • get_constraints(sys) or constraints(sys): The constraints for the optimization.

Transformations

Analyses

Applicable Calculation and Generation Functions

calculate_gradient
calculate_hessian
generate_gradient
generate_hessian
hessian_sparsity

Problem Constructors

SciMLBase.OptimizationProblem — Method
DiffEqBase.OptimizationProblem{iip}(sys::OptimizationSystem, u0map,
                                    parammap = DiffEqBase.NullParameters();
                                    grad = false,
                                    hess = false, sparse = false,
                                    cons_j = false, cons_h = false,
                                    checkbounds = false,
                                    linenumbers = true, parallel = SerialForm(),
                                    kwargs...) where {iip}

Generates an OptimizationProblem from an OptimizationSystem and allows for automatically symbolically calculating numerical enhancements.

Certain solvers require setting cons_j, cons_h to true for constrained-optimization problems.

source

Expression Constructors

ModelingToolkit.OptimizationProblemExpr — Type
DiffEqBase.OptimizationProblemExpr{iip}(sys::OptimizationSystem,
                                        parammap = DiffEqBase.NullParameters();
                                        u0 = nothing,
                                        grad = false,
                                        hes = false, sparse = false,
                                        checkbounds = false,
                                        linenumbers = true, parallel = SerialForm(),
                                        kwargs...) where {iip}

Generates a Julia expression for an OptimizationProblem from an OptimizationSystem and allows for automatically symbolically calculating numerical enhancements.

source