ggga

Run an example optimization benchmark function

ggga [OPTIONS] COMMAND [ARGS]...

Options

--interactive, --no-interactive

Whether to display the generated plots.

--samples <samples>

How many evaluations should be sampled. [default: 50]

--logy

Log-transform the objective function.

--noise <noise>

Standard deviation of test function noise. [default: 0.0]

--model <model>

The surrogate model implementation used for prediction. gpr: Gaussian Process Regression. knn: k-Nearest Neighbor. [default: gpr]

--quiet

Don’t display human-readable output during minimization.

--seed <seed>

Seed for reproducible runs. [default: 7861]

-s, --strategy <strategies>

Which optimization strategy will be used. Can be ‘random’, ‘ggga’, or a YAML document describing the strategy.

--csv <csv>

Write evaluation results to a CSV file. Only use this when running a single strategy.

--style <style>

DualDependenceStyle for the plots.

easom

Easom: Flat function with single sharp minimum.

ggga easom [OPTIONS]

Options

--interactive, --no-interactive

Whether to display the generated plots.

--samples <samples>

How many evaluations should be sampled.

--logy

Log-transform the objective function.

--noise <noise>

Standard deviation of test function noise.

--model <model>

The surrogate model implementation used for prediction. gpr: Gaussian Process Regression. knn: k-Nearest Neighbor.

--quiet

Don’t display human-readable output during minimization.

--seed <seed>

Seed for reproducible runs.

-s, --strategy <strategies>

Which optimization strategy will be used. Can be ‘random’, ‘ggga’, or a YAML document describing the strategy.

--csv <csv>

Write evaluation results to a CSV file. Only use this when running a single strategy.

--style <style>

DualDependenceStyle for the plots.

goldstein-price

Goldstein-Price: Asymetric function with single optimum.

ggga goldstein-price [OPTIONS]

Options

--interactive, --no-interactive

Whether to display the generated plots.

--samples <samples>

How many evaluations should be sampled.

--logy

Log-transform the objective function.

--noise <noise>

Standard deviation of test function noise.

--model <model>

The surrogate model implementation used for prediction. gpr: Gaussian Process Regression. knn: k-Nearest Neighbor.

--quiet

Don’t display human-readable output during minimization.

--seed <seed>

Seed for reproducible runs.

-s, --strategy <strategies>

Which optimization strategy will be used. Can be ‘random’, ‘ggga’, or a YAML document describing the strategy.

--csv <csv>

Write evaluation results to a CSV file. Only use this when running a single strategy.

--style <style>

DualDependenceStyle for the plots.

himmelblau

Himmelblau’s function: Asymetric polynomial with 4 minima.

ggga himmelblau [OPTIONS]

Options

--interactive, --no-interactive

Whether to display the generated plots.

--samples <samples>

How many evaluations should be sampled.

--logy

Log-transform the objective function.

--noise <noise>

Standard deviation of test function noise.

--model <model>

The surrogate model implementation used for prediction. gpr: Gaussian Process Regression. knn: k-Nearest Neighbor.

--quiet

Don’t display human-readable output during minimization.

--seed <seed>

Seed for reproducible runs.

-s, --strategy <strategies>

Which optimization strategy will be used. Can be ‘random’, ‘ggga’, or a YAML document describing the strategy.

--csv <csv>

Write evaluation results to a CSV file. Only use this when running a single strategy.

--style <style>

DualDependenceStyle for the plots.

onemax

One-Max function.

ggga onemax [OPTIONS]

Options

--interactive, --no-interactive

Whether to display the generated plots.

--samples <samples>

How many evaluations should be sampled.

--logy

Log-transform the objective function.

--noise <noise>

Standard deviation of test function noise.

--model <model>

The surrogate model implementation used for prediction. gpr: Gaussian Process Regression. knn: k-Nearest Neighbor.

--quiet

Don’t display human-readable output during minimization.

--seed <seed>

Seed for reproducible runs.

-s, --strategy <strategies>

Which optimization strategy will be used. Can be ‘random’, ‘ggga’, or a YAML document describing the strategy.

--csv <csv>

Write evaluation results to a CSV file. Only use this when running a single strategy.

--style <style>

DualDependenceStyle for the plots.

-D, --dimensions <dimensions>

Number of parameters/dimensions. [default: 4]

onemax-log

One-Max function.

ggga onemax-log [OPTIONS]

Options

--interactive, --no-interactive

Whether to display the generated plots.

--samples <samples>

How many evaluations should be sampled.

--logy

Log-transform the objective function.

--noise <noise>

Standard deviation of test function noise.

--model <model>

The surrogate model implementation used for prediction. gpr: Gaussian Process Regression. knn: k-Nearest Neighbor.

--quiet

Don’t display human-readable output during minimization.

--seed <seed>

Seed for reproducible runs.

-s, --strategy <strategies>

Which optimization strategy will be used. Can be ‘random’, ‘ggga’, or a YAML document describing the strategy.

--csv <csv>

Write evaluation results to a CSV file. Only use this when running a single strategy.

--style <style>

DualDependenceStyle for the plots.

-D, --dimensions <dimensions>

Number of parameters/dimensions. [default: 4]

rastrigin

Rastrigin Function: N-dimensional with many local minima.

ggga rastrigin [OPTIONS]

Options

--interactive, --no-interactive

Whether to display the generated plots.

--samples <samples>

How many evaluations should be sampled.

--logy

Log-transform the objective function.

--noise <noise>

Standard deviation of test function noise.

--model <model>

The surrogate model implementation used for prediction. gpr: Gaussian Process Regression. knn: k-Nearest Neighbor.

--quiet

Don’t display human-readable output during minimization.

--seed <seed>

Seed for reproducible runs.

-s, --strategy <strategies>

Which optimization strategy will be used. Can be ‘random’, ‘ggga’, or a YAML document describing the strategy.

--csv <csv>

Write evaluation results to a CSV file. Only use this when running a single strategy.

--style <style>

DualDependenceStyle for the plots.

-D, --dimensions <dimensions>

Number of parameters/dimensions. [default: 2]

rosenbrock

Rosenbrock function: N-dimensional and asymetric.

ggga rosenbrock [OPTIONS]

Options

--interactive, --no-interactive

Whether to display the generated plots.

--samples <samples>

How many evaluations should be sampled.

--logy

Log-transform the objective function.

--noise <noise>

Standard deviation of test function noise.

--model <model>

The surrogate model implementation used for prediction. gpr: Gaussian Process Regression. knn: k-Nearest Neighbor.

--quiet

Don’t display human-readable output during minimization.

--seed <seed>

Seed for reproducible runs.

-s, --strategy <strategies>

Which optimization strategy will be used. Can be ‘random’, ‘ggga’, or a YAML document describing the strategy.

--csv <csv>

Write evaluation results to a CSV file. Only use this when running a single strategy.

--style <style>

DualDependenceStyle for the plots.

-D, --dimensions <dimensions>

Number of parameters/dimensions. [default: 2]

sphere

Sphere function: N-dimensional, symmetric.

ggga sphere [OPTIONS]

Options

--interactive, --no-interactive

Whether to display the generated plots.

--samples <samples>

How many evaluations should be sampled.

--logy

Log-transform the objective function.

--noise <noise>

Standard deviation of test function noise.

--model <model>

The surrogate model implementation used for prediction. gpr: Gaussian Process Regression. knn: k-Nearest Neighbor.

--quiet

Don’t display human-readable output during minimization.

--seed <seed>

Seed for reproducible runs.

-s, --strategy <strategies>

Which optimization strategy will be used. Can be ‘random’, ‘ggga’, or a YAML document describing the strategy.

--csv <csv>

Write evaluation results to a CSV file. Only use this when running a single strategy.

--style <style>

DualDependenceStyle for the plots.

-D, --dimensions <dimensions>

Number of parameters/dimensions. [default: 2]

trap

Like One-Max, but has misleading gradient.

ggga trap [OPTIONS]

Options

--interactive, --no-interactive

Whether to display the generated plots.

--samples <samples>

How many evaluations should be sampled.

--logy

Log-transform the objective function.

--noise <noise>

Standard deviation of test function noise.

--model <model>

The surrogate model implementation used for prediction. gpr: Gaussian Process Regression. knn: k-Nearest Neighbor.

--quiet

Don’t display human-readable output during minimization.

--seed <seed>

Seed for reproducible runs.

-s, --strategy <strategies>

Which optimization strategy will be used. Can be ‘random’, ‘ggga’, or a YAML document describing the strategy.

--csv <csv>

Write evaluation results to a CSV file. Only use this when running a single strategy.

--style <style>

DualDependenceStyle for the plots.

-D, --dimensions <dimensions>

Number of parameters/dimensions. [default: 2]