piglot.objectives package

Submodules

piglot.objectives.analytical module

Provide analytical functions for optimisation.

class AnalyticalMultiObjective(parameters: ParameterSet, objectives: Dict[str, AnalyticalSingleObjective], stochastic: bool = False, scalarisation: Scalarisation | None = None, composite: bool = False, output_dir: str | None = None)[source]

Bases: GenericObjective

Multi-objective problem derived from a set of analytical expressions.

plot_case(case_hash: str, options: Dict[str, Any] | None = None) → List[Figure][source]

Plot a given function call given the parameter hash.

Parameters

case_hashstr, optional: Parameter hash for the case to plot.
optionsDict[str, Any], optional: Options to pass to the plotting function, by default None.

Returns

List[Figure]: List of figures with the plot.

classmethod read(config: Dict[str, Any], parameters: ParameterSet, output_dir: str) → AnalyticalMultiObjective[source]

Read the objective from a configuration dictionary.

Parameters

configDict[str, Any]: Terms from the configuration dictionary.
parametersParameterSet: Set of parameters for this problem.
output_dirstr: Path to the output directory.

Returns

SyntheticObjective: Objective function to optimise.

class AnalyticalObjective(parameters: ParameterSet, expression: str, variance: str | None = None, stochastic: bool = False, random_evals: int = 0, output_dir: str | None = None, maximise: bool = False, weight: float = 1.0, bounds: Tuple[float, float] | None = None)[source]

Bases: GenericObjective

Objective function derived from an analytical expression.

plot_case(case_hash: str, options: Dict[str, Any] | None = None) → List[Figure][source]

Plot a given function call given the parameter hash.

Parameters

case_hashstr, optional: Parameter hash for the case to plot.
optionsDict[str, Any], optional: Options to pass to the plotting function, by default None.

Returns

List[Figure]: List of figures with the plot.

classmethod read(config: Dict[str, Any], parameters: ParameterSet, output_dir: str) → AnalyticalObjective[source]

Read the objective from a configuration dictionary.

Parameters

configDict[str, Any]: Terms from the configuration dictionary.
parametersParameterSet: Set of parameters for this problem.
output_dirstr: Path to the output directory.

Returns

SyntheticObjective: Objective function to optimise.

class AnalyticalSingleObjective(parameters: ParameterSet, expression: str, variance: str | None = None, random_evals: int = 0, maximise: bool = False, weight: float = 1.0, bounds: Tuple[float, float] | None = None)[source]

Bases: IndividualObjective

Objective function derived from an analytical expression.

evaluate(params: ndarray, use_random: bool = True) → Tuple[float, float][source]

Evaluate objective value for the given results.

Parameters

paramsnp.ndarray: Parameter values for this evaluation.
use_randombool: Whether to use random evaluations (true by default).

Returns

Tuple[float, float]: Mean and variance of the objective.

plot_1d(values: ndarray, append_title: str) → Figure[source]

Plot the objective in 1D.

Parameters

valuesnp.ndarray: Parameter values to plot for.
append_titlestr: String to append to the title.

Returns

Figure: Figure with the plot.

plot_2d(values: ndarray, append_title: str) → Figure[source]

Plot the objective in 2D.

Parameters

valuesnp.ndarray: Parameter values to plot for.
append_titlestr: String to append to the title.

Returns

Figure: Figure with the plot

classmethod read(config: Dict[str, Any], parameters: ParameterSet) → AnalyticalSingleObjective[source]

Read the objective from a configuration dictionary.

Parameters

configDict[str, Any]: Terms from the configuration dictionary.
parametersParameterSet: Set of parameters for this problem.

Returns

AnalyticalSingleObjective: Objective function to optimise.

class ScalarisationComposition(scalarisation: Scalarisation)[source]

Bases: Composition

Composition for scalarisation of multiple objectives.

composition_torch(inner: Tensor, params: Tensor) → Tensor[source]

Compute the composition for all objectives.

Parameters

innertorch.Tensor: Return value from the inner function.
paramstorch.Tensor: Paratemers for the given responses.

Returns

torch.Tensor: Composition results.

piglot.objectives.compositions module

piglot.objectives.design module

Module for curve fitting objectives

class DesignSingleObjective(name: str, prediction: List[str], quantity: Reduction, maximise: bool = False, weight: float = 1.0, bounds: Tuple[float, float] | None = None, flatten_utility: FlattenUtility | None = None, prediction_transform: ResponseTransformer | None = None)[source]

Bases: ResponseSingleObjective

Single objective for design optimisation objectives.

plot(axis: Axes, raw_results: Dict[str, OutputResult]) → Dict[Line2D, str][source]

Plot the response for this objective.

Parameters

axisplt.Axes: Axis to plot the response on.
raw_resultsDict[str, OutputResult]: Raw responses from the solver.

Returns

Dict[Line2D, str]: Mapping of lines to response names (for dynamically updating plots).

classmethod read(name: str, config: Dict[str, Any], output_dir: str) → DesignSingleObjective[source]

Read the objective spec from the configuration dictionary.

Parameters

namestr: Name of the objective.
configDict[str, Any]: Configuration dictionary.
output_dir: str: Output directory.

Returns

ResponseSingleObjective: Single objective to use.

class ResponseDesignObjective(parameters: ParameterSet, solver: Solver, objectives: List[ResponseSingleObjective], output_dir: str, scalarisation: Scalarisation | None = None, stochastic: bool = False, composite: bool = False, full_composite: bool = True, transformers: Dict[str, ResponseTransformer] | None = None)[source]

Bases: ResponseObjective

Class for design of response-based objectives.

classmethod read(config: Dict[str, Any], parameters: ParameterSet, output_dir: str) → ResponseDesignObjective[source]

Read the objective from a configuration dictionary.

Parameters

configDict[str, Any]: Terms from the configuration dictionary.
parametersParameterSet: Set of parameters for this problem.
output_dirstr: Path to the output directory.

Returns

ResponseDesignObjective: Objective function to optimise.

piglot.objectives.fitting module

Module for curve fitting objectives

class FittingSingleObjective(name: str, reference: Reference, prediction: List[str], reduction: Reduction, weight: float = 1.0, bounds: Tuple[float, float] | None = None)[source]

Bases: ResponseSingleObjective

Single objective for curve fitting optimisation objectives.

plot(axis: Axes, raw_results: Dict[str, OutputResult]) → Dict[Line2D, str][source]

Plot the response for this objective.

Parameters

axisplt.Axes: Axis to plot the response on.
raw_resultsDict[str, OutputResult]: Raw responses from the solver.

Returns

Dict[Line2D, str]: Mapping of lines to response names (for dynamically updating plots).

classmethod read(name: str, config: Dict[str, Any], output_dir: str) → FittingSingleObjective[source]

Read the objective spec from the configuration dictionary.

Parameters

namestr: Name of the objective.
configDict[str, Any]: Configuration dictionary.
output_dir: str: Output directory.

Returns

ResponseSingleObjective: Single objective to use.

class Reference(filename: str, output_dir: str, x_col: int = 1, y_col: int = 2, skip_header: int = 0, transformer: ResponseTransformer | None = None, filter_tol: float = 0.0, show: bool = False)[source]

Bases: object

Container for reference solutions.

get_data() → ndarray[source]

Get the data column of the reference.

Returns

np.ndarray: Data column.

get_time() → ndarray[source]

Get the time column of the reference.

Returns

np.ndarray: Time column.

has_filtering() → bool[source]

Check if the reference has filtering.

Returns

bool: Whether the reference has filtering.

prepare() → None[source]: Prepare the reference data.

static read(filename: str, config: Dict[str, Any], output_dir: str) → Reference[source]

Read the reference from the configuration dictionary.

Parameters

filenamestr: Path to the reference file.
configDict[str, Any]: Configuration dictionary.
output_dir: str: Output directory.

Returns

Reference: Reference to use for this problem.

class ResponseFittingObjective(parameters: ParameterSet, solver: Solver, objectives: List[ResponseSingleObjective], output_dir: str, scalarisation: Scalarisation | None = None, stochastic: bool = False, composite: bool = False, full_composite: bool = True, transformers: Dict[str, ResponseTransformer] | None = None)[source]

Bases: ResponseObjective

Class for fitting of response-based objectives.

prepare()[source]

Prepare the objective for optimisation.

For curve fitting, this involves preparing the reference data and updating both the flatten utility and the transformer.

classmethod read(config: Dict[str, Any], parameters: ParameterSet, output_dir: str) → ResponseFittingObjective[source]

Read the objective from a configuration dictionary.

Parameters

configDict[str, Any]: Terms from the configuration dictionary.
parametersParameterSet: Set of parameters for this problem.
output_dirstr: Path to the output directory.

Returns

ResponseFittingObjective: Objective function to optimise.

piglot.objectives.synthetic module

Provide synthetic test functions

class SyntheticObjective(parameters: ParameterSet, name: str, output_dir: str, transform: str | None = None, composition: Reduction | None = None, **kwargs)[source]

Bases: GenericObjective

Objective function derived from a synthetic test function.

static get_test_functions() → Dict[str, Type[SyntheticTestFunction]][source]

Return available test functions.

Returns

Dict[str, Type[SyntheticTestFunction]]: Available test functions.

static read(config: Dict[str, Any], parameters: ParameterSet, output_dir: str) → SyntheticObjective[source]

Read the objective from a configuration dictionary.

Parameters

configDict[str, Any]: Terms from the configuration dictionary.
parametersParameterSet: Set of parameters for this problem.
output_dirstr: Path to the output directory.

Returns

SyntheticObjective: Objective function to optimise.

Module contents

Interface for objectives.

read_objective(config: Dict[str, Any], parameters: ParameterSet, output_dir: str) → Objective[source]

Read the objective from the configuration dictionary.

Parameters

configDict[str, Any]: Configuration dictionary.
parametersParameterSet: Parameter set for this problem.
output_dirstr: Path to the output directory.

Returns

Objective: Objective to optimise for.