optimpv.pymooOpti package

Submodules

optimpv.pymooOpti.pymooOptimizer module

PymooOptimizer module. This module contains the PymooOptimizer class, which is used to optimize a given function using pymoo’s genetic algorithms.

class optimpv.pymooOpti.pymooOptimizer.PymooOptimizer(params=None, agents=None, algorithm='NSGA2', pop_size=100, n_gen=100, algorithm_kwargs=None, existing_data=None, suggest_only=False, name='pymoo', **kwargs)

Bases: BaseAgent

Initialize the PymooOptimizer class. This class is used to optimize a given function using pymoo algorithms.

Parameters:

• params (list of Fitparam() objects, optional) – List of Fitparam() objects, by default None

• agents (list of Agent() objects, optional) – List of Agent() objects see optimpv/general/BaseAgent.py for a base class definition, by default None

• algorithm (str, optional) – Optimization algorithm to use, by default ‘NSGA2’ Options: ‘NSGA2’, ‘NSGA3’, ‘MOEAD’, ‘GA’, ‘DE’, ‘PSO’

• pop_size (int, optional) – Population size, by default 100

• n_gen (int, optional) – Number of generations, by default 100

• algorithm_kwargs (dict, optional) – Additional keyword arguments for the algorithm, by default None

• existing_data (DataFrame, optional) – existing data to use for the optimization process, by default None

• suggest_only (bool, optional) – if True, the optimization process will only suggest new points without running the agents, by default False

• name (str, optional) – Name of the optimization process, by default ‘pymoo’

• **kwargs (dict) – Additional keyword arguments including: - parallel_agents: bool, whether to run agents in parallel - max_parallelism: int, maximum number of parallel processes - verbose_logging: bool, whether to log verbose information - seed: int, random seed for reproducibility

create_algorithm()

Create a pymoo algorithm based on the specified algorithm name or return the provided algorithm object.

create_callback()

Create a callback to track optimization progress.

create_problem()

Create a pymoo problem from agents and parameters.

existing_population()

Create an existing population from the provided data.

get_best_solution(balance_objectives=True)

Get the best solution from the optimization results.

Parameters:

balance_objectives (bool, optional) – If True and multiple objectives exist, return the solution with best balance. If False, return the first solution from the Pareto front.

Returns:

Dictionary containing the best parameters

Return type:

dict

get_df_from_pymoo()

Convert the optimization results to a pandas DataFrame.

Returns:

DataFrame containing the optimization results

Return type:

pd.DataFrame

get_pareto_front()

Get the Pareto front solutions for multi-objective optimization.

Returns:

Tuple containing (parameters_list, objectives_array)

Return type:

tuple

optimize()

Run the optimization process using pymoo.

plot_convergence(save_path=None, **kwargs)

Plot the convergence history of the optimization.

Parameters:

• save_path (str, optional) – Path to save the plot, by default None

• **kwargs (dict) –

  Additional keyword arguments for the plot.

  • xscalestr, optional

    Scale for the x-axis, by default ‘linear’

  • yscalestr, optional

    Scale for the y-axis, by default ‘linear’

update_params_with_best_balance(return_best_balance=False)

Update the parameters with the best balance of all metrics.

If multiple objectives exist, find the best balance based on ranks.

Parameters:

return_best_balance (bool, optional) – If True, return the index and parameters of the best balance, by default False

Returns:

If return_best_balance is True, return the index and parameters of the best balance. Otherwise, return None.

Return type:

tuple

Raises:

ValueError – If no evaluations have been performed or no metrics are available.

class optimpv.pymooOpti.pymooOptimizer.PymooProblem(agents, **kwargs)

Bases: ElementwiseProblem

create_search_space(params)

Create a search space suitable for scipy.optimize.minimize based on parameters.

Parameters:

params (list of Fitparam() objects) – List of parameter objects

Returns:

A tuple containing (x0, bounds) where x0 is the initial parameter vector and bounds is a list of (lower, upper) bound tuples

Return type:

tuple

evaluate_single(args)

Evaluate a single agent on a parameter point

get_all_metrics()

Create the objectives for the optimization process.

parse_constraints(constraints)

Parse parameter constraints from string format to numerical format.

Parameters:

constraints (list of str) – List of constraint strings in format like “x1 + 2*x2 <= 5”

Returns:

List of constraint dictionaries with ‘indices’, ‘coefficients’, ‘rhs’, ‘operator’

Return type:

list of dict

reconstruct_params(x)

Reconstruct a parameter dictionary from an optimization vector.

Parameters:

x (array-like) – Parameter vector from the optimizer

Returns:

Dictionary mapping parameter names to values

Return type:

dict

Module contents