iwopy.interfaces.pygmo.optimizer¶

Classes¶

Module Contents¶

class iwopy.interfaces.pygmo.optimizer.Optimizer_pygmo(problem: iwopy.core.Problem, optimizer: str = 'de', *, pop_size: int = 1, seed: int | None = None, vectorize: bool = False, max_iterations: int | None = None, max_evaluations: int | None = None, norandom_first: bool | None = None, c_tol: float = 0.0001, problem_verbosity: int = 0, algorithm_verbosity: int = 1, nlopt_optimizer: str | None = None, name: str = 'Optimizer_pygmo', **kwargs: Any)[source]¶

Bases: iwopy.core.Optimizer

Interface to the pygmo optimizers for serial runs.

Parameters:¶

problem: iwopy.core.Problem¶

The problem to optimize

optimizer: str = 'de'¶

Native driver name, such as de, ipopt, or nlopt.

pop_size: int = 1¶

Absolute population size. Local solvers operate on one selected member.

seed: int | None = None¶

Seed for both the population and native drivers supporting seed.

vectorize: bool = False¶

Enable batch fitness where supported by the selected driver.

max_iterations: int | None = None¶

IPOPT iteration or evolutionary-driver generation limit. Unsupported for NLopt and compass_search. None preserves the native default.

max_evaluations: int | None = None¶

Native driver fitness-call limit for NLopt or compass_search, excluding initial population construction. Unsupported for other drivers. None preserves the native default.

norandom_first: bool | None = None¶

Replace the first individual with problem initial values; None enables this when pop_size is one.

c_tol: float = 0.0001¶

Native population constraint tolerance.

problem_verbosity: int = 0¶

UDP diagnostics and native algorithm verbosity levels.

algorithm_verbosity: int = 1¶

UDP diagnostics and native algorithm verbosity levels.

nlopt_optimizer: str | None = None¶

Native NLopt solver name when optimizer is nlopt. The standalone factory supplies it automatically for family nlopt.

name: str = 'Optimizer_pygmo'¶

Optimizer name.

**kwargs: Any¶

Flat native driver options, such as print_level and ftol_rel.

Raises:¶

• ValueError – If limits are invalid or unsupported by the selected driver, vectorize is not a boolean, or driver selection is inconsistent.

• TypeError – If a superseded or unsupported shared keyword is supplied.

finalize(opt_results: iwopy.core.opt_results.SingleObjOptResults | iwopy.core.opt_results.MultiObjOptResults | int | None = None, verbosity: int = 1) → None¶

This function may be called after finishing the optimization.

Parameters:¶

opt_results: iwopy.core.opt_results.SingleObjOptResults | iwopy.core.opt_results.MultiObjOptResults | int | None = None¶

The optimization results object

verbosity: int = 1¶

The verbosity level, 0 = silent

initialize(verbosity: int = 1) → None[source]¶

Initialize the object.

Parameters:¶

verbosity: int = 1¶

The verbosity level, 0 = silent

print_info() → None[source]¶

Print solver info, called before solving

solve(verbosity: int = 1, callbacks: list[iwopy.core.OptimizerCallback] | None = None) → iwopy.core.SingleObjOptResults[source]¶

Run the optimization solver.

Parameters:¶

verbosity: int = 1¶

The verbosity level, 0 = silent

callbacks: list[iwopy.core.OptimizerCallback] | None = None¶

Ordered callbacks for intermediate optimization states

Returns:¶

results – The optimization results object

algo : Any | None = None¶

c_tol = 0.0001¶

property initialized : bool¶

Flag for finished initialization

Returns:¶

initialized – True if initialization has been done

name = 'optimizer'¶

norandom_first¶

optimizer = 'de'¶

pop : Any | None = None¶

pop_size = 1¶

problem¶

problem_verbosity = 0¶

seed = None¶

udp : iwopy.interfaces.pygmo.problem.UDP | None = None¶

vectorize = False¶

verbosity = 1¶