Optimizers¶

class ionworkspipeline.optimizers.Optimizer(**kwargs: Any)¶

Base class for all optimizers.

Optimizers seek a single optimal point in parameter space that minimizes the objective function.

Parameters¶

**kwargs: Arguments passed to the underlying optimizer algorithm.

Extends: ionworkspipeline.data_fits.parameter_estimators.parameter_estimator.ParameterEstimator

run(x0: ndarray) → OptimizerResult¶

Optimize the objective function.

Parameters¶

x0array_like: Initial guess for the independent variables.

Returns¶

resionworkspipeline.OptimizerResult: The result of the optimization.

class ionworkspipeline.optimizers.ScipyDifferentialEvolution(**kwargs: Any)¶

Global stochastic optimizer using differential evolution with parallel evaluation.

Differential evolution is a robust global optimization algorithm that evolves a population of candidate solutions across generations. It excels at handling multi-modal, non-convex objective landscapes and requires no gradient information.

Notes¶

Does not support custom equality or inequality constraints
Parallel workers can speed up optimization; set this optimizer’s num_workers (default 1)
Initial guess x0 is ignored; initial population is generated from bounds
Polish option disabled by default as it conventionally significantly decreases performance
Callback logs only best solution per generation (not individual evaluations)

Parameters¶

workersint, default=1: Number of parallel workers for function evaluations. Use -1 for all CPU cores.
max_iterationsint, default=1000: Maximum number of generations. Passed to scipy as maxiter.
population_sizeint, default=15: Population size multiplier (total population = population_size * dimensionality). Passed to scipy as popsize.
strategystr, default=’best1bin’: Differential evolution strategy. Options include ‘best1bin’, ‘rand1bin’, ‘best2bin’, ‘rand2bin’, ‘currenttobest1bin’.
mutationfloat or tuple, default=(0.5, 1): Mutation constant. Can be float or (min, max) tuple for adaptive mutation.
recombinationfloat, default=0.7: Crossover probability for parameter mixing.
seedint, optional: Random seed for reproducible results.
atol, tolfloat, optional: Absolute and relative tolerance for convergence.
**kwargs: Additional arguments passed to scipy.optimize.differential_evolution. See scipy documentation for complete options.

Examples¶

Basic usage (single worker for doctests):

>>> optimizer = ScipyDifferentialEvolution(max_iterations=50, seed=42)
>>> optimizer.set_objective(sphere)
>>> optimizer.set_bounds((lower, upper))
>>> result = optimizer.run(x0)
>>> result.fun < 1e-3
True

Integration with DataFit:

>>> optimizer = ScipyDifferentialEvolution(max_iterations=500)
>>> isinstance(optimizer, ScipyDifferentialEvolution)
True

Extends: ionworkspipeline.data_fits.parameter_estimators.optimizers.optimizer.Optimizer

run(x0: ndarray) → OptimizeResult¶

Minimize objective using differential evolution.

Parameters¶

x0: Ignored. Initial population is randomly generated within bounds.

Returns¶

OptimizeResult: Optimization result with x (best solution), fun (best cost), success (convergence flag), and generation statistics.

set_evaluation_callback(callback: Callable[[list[ndarray], list[float]], None] | None = None) → None¶

Configure callback for logging best solution after each generation.

Unlike other optimizers, this only logs the best solution per generation, not individual evaluations (which occur in parallel worker processes).

Parameters¶

callback: Function receiving lists of parameter vectors and costs. Called with single-element lists containing the generation’s best solution. Set to None to disable callbacks.

class ionworkspipeline.optimizers.ScipyLeastSquares(**kwargs: Any)¶

Nonlinear least squares optimizer using scipy’s Trust Region Reflective algorithm.

This optimizer is designed for problems where the objective returns a residual vector rather than a scalar cost. It minimizes the sum of squares of the residuals. Best suited for well-behaved, smooth problems with a clear residual structure.

Notes¶

Requires objective functions that return an array (residual vector)
Automatically handles linear algebra errors by returning NaN values
More efficient than general minimization for least-squares structure
Supports bound constraints but not general equality/inequality constraints

Parameters¶

methodstr, optional: Algorithm to use. Options: ‘trf’ (default), ‘dogbox’, ‘lm’.
ftol, xtol, gtolfloat, optional: Tolerance parameters for convergence criteria.
max_nfevint, optional: Maximum number of function evaluations.
**kwargs: Additional arguments passed to scipy.optimize.least_squares. See scipy documentation for complete options.

Examples¶

>>> optimizer = ScipyLeastSquares(method='trf', max_nfev=100)
>>> optimizer.set_objective(sphere_residuals)
>>> optimizer.set_bounds((lower, upper))
>>> result = optimizer.run(x0)
>>> np.allclose(result.x, [0, 0, 0], atol=1e-3)
True

Extends: ionworkspipeline.data_fits.parameter_estimators.optimizers.optimizer.Optimizer

run(x0: ndarray) → OptimizeResult¶

Minimize the sum of squares of the objective residuals.

Parameters¶

x0: Initial parameter values.

Returns¶

OptimizeResult: Optimization result with x (solution), cost (final residual norm), success (convergence flag), and other attributes.

class ionworkspipeline.optimizers.ScipyMinimize(**kwargs: Any)¶

General-purpose scalar minimization with support for constraints.

Wraps scipy’s minimize function, providing access to multiple local optimization algorithms (e.g., L-BFGS-B, SLSQP, trust-constr, COBYQA). Suitable for smooth, scalar-valued objectives with optional equality and inequality constraints.

Notes¶

Requires objective functions that return a scalar value
Supports bound constraints and custom equality/inequality constraints
Choice of method depends on problem structure and constraint types
Some methods (e.g., ‘L-BFGS-B’) support bounds only, not general constraints

Parameters¶

methodstr, optional: Optimization algorithm. Common choices: - ‘L-BFGS-B’: Bound-constrained, gradient-based (default for bounded problems) - ‘SLSQP’: Sequential Least Squares, supports all constraint types - ‘trust-constr’: Modern trust-region method, supports all constraints - ‘COBYQA’: Derivative-free, supports nonlinear constraints
max_iterationsint, optional: Maximum number of iterations. Passed to scipy as maxiter.
tolfloat, optional: Tolerance for termination.
**kwargs: Additional arguments passed to scipy.optimize.minimize. See scipy documentation for complete options.

Examples¶

>>> optimizer = ScipyMinimize(method='L-BFGS-B', max_iterations=100)
>>> optimizer.set_objective(sphere)
>>> optimizer.set_bounds((lower, upper))
>>> result = optimizer.run(x0)
>>> np.allclose(result.x, [0, 0, 0], atol=1e-3)
True

Extends: ionworkspipeline.data_fits.parameter_estimators.optimizers.optimizer.Optimizer

run(x0: ndarray) → OptimizeResult¶

Minimize a scalar objective function.

Parameters¶

x0: Initial parameter values.

Returns¶

OptimizeResult: Optimization result with x (solution), fun (final cost), success (convergence flag), and other method-specific attributes.

class ionworkspipeline.optimizers.ScipyShgo(**kwargs: Any)¶

Global optimizer using simplicial homology techniques.

SHGO (Simplicial Homology Global Optimization) uses topological techniques to identify and sample from all local minima basins. It’s particularly effective for problems with many local minima and supports general nonlinear constraints.

Notes¶

Deterministic algorithm (reproducible results without random seed)
Efficiently handles problems with many local optima
Supports bound, equality, and inequality constraints
May be slower than stochastic methods for high-dimensional problems
Initial guess x0 is ignored; sampling points determined by algorithm

Parameters¶

nint, default=100: Number of sampling points used in the algorithm.
itersint, default=1: Number of iterations for algorithm convergence.
sampling_methodstr, default=’simplicial’: Sampling strategy: ‘simplicial’ (default) or ‘sobol’.
minimizer_kwargsdict, optional: Additional arguments passed to the local minimizer.
**kwargs: Additional arguments passed to scipy.optimize.shgo. See scipy documentation for complete options.

Examples¶

>>> optimizer = ScipyShgo(n=100, iters=1)
>>> optimizer.set_objective(sphere)
>>> optimizer.set_bounds((lower, upper))
>>> result = optimizer.run(x0)  # x0 is ignored
>>> np.allclose(result.x, [0, 0, 0], atol=1e-3)
True

Extends: ionworkspipeline.data_fits.parameter_estimators.optimizers.optimizer.Optimizer

run(x0: ndarray) → OptimizeResult¶

Minimize objective using SHGO algorithm.

Parameters¶

x0: Ignored. Sampling points are determined by the algorithm.

Returns¶

OptimizeResult: Optimization result with x (global minimum), fun (minimum cost), success (convergence flag), and information about local minima found.

class ionworkspipeline.optimizers.ScipyDualAnnealing(**kwargs: Any)¶

Global stochastic optimizer using dual annealing.

Dual annealing combines generalized simulated annealing with fast local search. It’s designed for global optimization with a good balance between exploration and exploitation, particularly effective for rugged objective landscapes.

Notes¶

Does not support custom equality or inequality constraints
Accepts optional initial guess x0 to seed the search
Stochastic algorithm (use seed parameter for reproducibility)
Generally faster convergence than pure simulated annealing
Good choice when gradient information is unavailable

Parameters¶

max_iterationsint, default=1000: Maximum number of global search iterations. Passed to scipy as maxiter.
initial_tempfloat, default=5230: Initial temperature for the annealing schedule.
restart_temp_ratiofloat, default=2e-5: Temperature ratio for restart condition during local search.
visitfloat, default=2.62: Parameter for the visiting distribution (higher = more exploration).
acceptfloat, default=-5.0: Parameter for the acceptance distribution (lower = more exploitation).
seedint, optional: Random seed for reproducible results.
no_local_searchbool, default=False: If True, skip local minimization (pure generalized simulated annealing).
**kwargs: Additional arguments passed to scipy.optimize.dual_annealing. See scipy documentation for complete options.

Examples¶

Basic usage with initial guess:

>>> optimizer = ScipyDualAnnealing(max_iterations=100, seed=42)
>>> optimizer.set_objective(sphere)
>>> optimizer.set_bounds((lower, upper))
>>> result = optimizer.run(x0)
>>> result.fun < 1e-3
True

Extends: ionworkspipeline.data_fits.parameter_estimators.optimizers.optimizer.Optimizer

run(x0: ndarray) → OptimizeResult¶

Minimize objective using dual annealing.

Parameters¶

x0: Initial guess to seed the search. The algorithm may explore beyond this point during the global search phase.

Returns¶

OptimizeResult: Optimization result with x (best solution), fun (best cost), success (convergence flag), and annealing statistics.

class ionworkspipeline.optimizers.AskTellOptimizer(method: str = 'CMAES', log_to_screen: bool = False, sigma0: float | ndarray | None = None, max_iterations: int | None = None, max_unchanged_iterations: int | None = None, max_unchanged_iterations_threshold: float | None = None, min_iterations: int = 1, max_evaluations: int = 1000000, population_size: int | None = None, threshold: float | None = None, absolute_tolerance: float = 1e-05, relative_tolerance: float = 0.01, xtol: float | None = 1e-06, population_convergence_tol: float | None = 0.005, flat_fitness_tol: float | None = None, surrogate_convergence_tol: float | None = 0.001, convergence_patience: int = 3, algorithm_options: dict[str, Any] | None = None, **kwargs: Any)¶

Optimizer using ask/tell algorithms for population-based and simplex optimization.

Supports CMAES, PSO, DifferentialEvolution, XNES, Nelder-Mead, BayesianOptimization, SOBER, and TuRBO.

Parameters¶

methodstr, optional: Optimization method. Default is “CMAES”. Must be one of: “CMAES”, “Nelder-Mead”, “PSO”, “DifferentialEvolution”, “XNES”, “BayesianOptimization”, “SOBER”, or “TuRBO”.
log_to_screenbool, optional: Whether to print optimization progress. Default is False.
sigma0float | np.ndarray, optional: Initial step size for population-based methods. Default is None.
max_iterationsint, optional: Maximum number of iterations. Default is None (auto-computed).
max_unchanged_iterationsint, optional: Stop after this many iterations without improvement. Default is None.
max_unchanged_iterations_thresholdfloat, optional: Threshold for determining improvement. Default is 1e-5.
min_iterationsint, optional: Minimum iterations before checking stopping criteria. Default is 1.
max_evaluationsint, optional: Maximum number of function evaluations. Default is 1e6.
population_sizeint, optional: Population size for population-based methods. Default is method-specific.
thresholdfloat, optional: Target objective value to stop optimization. Default is None.
absolute_tolerancefloat, optional: Absolute tolerance for unchanged iterations. Default is 1e-5.
relative_tolerancefloat, optional: Relative tolerance for unchanged iterations. Default is 1e-2.
xtolfloat, optional: Parameter change tolerance. Stops when the L-infinity norm of the change in x_guessed between generations drops below this value. Default is 1e-6.
population_convergence_tolfloat, optional: Population convergence tolerance. The exact convergence semantics are algorithm-dependent (e.g. DE also checks position-space diversity). Default is 1e-2.
flat_fitness_tolfloat, optional: Flat fitness tolerance. Stops when >= 50% of population members have fitness within this tolerance of the median. Default is None (disabled).
surrogate_convergence_tolfloat, optional: Relative tolerance on the model-native convergence signal for surrogate algorithms (BO, SOBER). Fires when the expected improvement at the algorithm’s best candidate is below tol times the range of observed costs for convergence_patience consecutive iterations. Default 1e-3. Set to None to disable.
convergence_patienceint, optional: Number of consecutive generations the population convergence check must pass before the optimizer actually stops. Higher values guard against transient convergence signals. Default is 3.
algorithm_optionsdict, optional: Algorithm-specific configuration options (for PSO, DE, and CMA-ES). CMA-ES accepts any key from cma.CMAOptions().

Notes¶

"CMAES" is the default general-purpose choice. For expensive low-dimensional problems use "BayesianOptimization"; for expensive parallel or higher-dimensional problems use "TuRBO"; for cheap objectives consider "DifferentialEvolution".

Extends: ionworkspipeline.data_fits.parameter_estimators.optimizers.optimizer.Optimizer

property optimizer: Algorithm | None¶: Access the underlying algorithm instance.

property population_size: int | None¶

The size of the population.

Returns the live algorithm’s population size when available, falls back to the user-specified value, and finally computes the method-specific default from bounds if both are None. This allows callers (e.g. the backend distributed-evaluator setup) to obtain a concrete value after set_bounds() but before run().

property requires_constraint_decomposition: bool¶: True when the configured algorithm consumes constraint decomposition.

property requires_finite_failure_cost: bool¶: True for surrogate algorithms, whose GP fit needs finite costs.

run(x0: list[float]) → OptimizerResult¶

Run the optimization using the configured algorithm.

Drives a single AskTellStart synchronously: each point the start asks for is evaluated in-process and reported back until a stopping criterion fires. This is the same state machine the execution engine pumps for parallel multistart fits, so a direct run and an engine-driven start share one ask/tell loop, bookkeeping, and result-builder. Objective-level parallelism (fanning the objective tasks out to an executor) lives inside each evaluation.

Parameters¶

x0list[float]: Initial parameter values.

Returns¶

iwp.OptimizerResult: Optimization result with optimized parameters and metadata.

set_constraint_names(names: list[str]) → None¶

Record ordered constraint names for constraint-aware surrogates.

Set by DataFit from the objectives’ action constraints. The names order the constraint-violation columns surrogate optimizers consume.

set_parameters_metadata(priors: list | None = None, transforms: list | None = None, scales: list | None = None) → None¶

Record per-parameter metadata forwarded to surrogate algorithms.

Parameters¶

priorslist | None: One prior distribution per fit parameter, or None entries.
transformslist | None: np.log/np.log10/None per fit parameter.
scaleslist | None: Normalization scale per fit parameter.

class ionworkspipeline.optimizers.PointEstimate¶

Point estimate optimizer - returns the initial guess without optimization.

Useful for evaluating a single parameter set or initializing pipelines.

Extends: ionworkspipeline.data_fits.parameter_estimators.optimizers.optimizer.Optimizer

classmethod from_schema(schema: Any) → PointEstimate¶

Construct from a validated ionworks_schema schema instance.

Accepts both ionworks_schema.parameter_estimators.PointEstimateOptimizer and DummyOptimizer (which is an alias for PointEstimateOptimizer). Both schemas have no fields, so the schema instance is consumed only for type symmetry with the rest of the optimizer surface.

run(x0: ndarray) → OptimizerResult¶

Optimize the objective function.

Parameters¶

x0array_like: Initial guess for the independent variables.

Returns¶

resionworkspipeline.OptimizerResult: The result of the optimization.

class ionworkspipeline.optimizers.Dummy(*args, **kwargs)¶

Alias for PointEstimate optimizer.

Extends: ionworkspipeline.data_fits.parameter_estimators.optimizers.point_estimate_optimizer.PointEstimate

Optimizers¶

Parameters¶

Parameters¶

Returns¶

Notes¶

Parameters¶

Examples¶

Parameters¶

Returns¶

Parameters¶

Notes¶

Parameters¶

Examples¶

Parameters¶

Returns¶

Notes¶

Parameters¶

Examples¶

Parameters¶

Returns¶

Notes¶

Parameters¶

Examples¶

Parameters¶

Returns¶

Notes¶

Parameters¶

Examples¶

Parameters¶

Returns¶

Parameters¶

Notes¶

Parameters¶

Returns¶

Parameters¶

Parameters¶

Returns¶

This Page