iOpt Interface Classes

problem

class iOpt.problem.Problem

Базовые классы: abc.ABC

Base class for optimization problems

calculate(point: iOpt.trial.Point, function_value: iOpt.trial.FunctionValue) → iOpt.trial.FunctionValue

Calculate a function at a given point.

For any new problem statement that inherits from Problem, this method should be overloaded

Результат

Calculated value of the function.

calculateAllFunction(point: iOpt.trial.Point, function_values: array([None], dtype=object))

Calculate all functions at a given point.

For any new problem statement that inherits from Problem, this method should be overloaded

Результат

Calculated values of the functions.

get_name()

Get the name of the problem

Результат

self.name.

solution

class iOpt.solution.Solution(problem: iOpt.problem.Problem, best_trials: array([None], dtype=object) = [<iOpt.trial.Trial object>], number_of_global_trials: int = 1, number_of_local_trials: int = 0, solving_time: numpy.float64 = 0.0, solution_accuracy: numpy.float64 = 0.0)

Базовые классы: object

Class of description of the solution to the optimization problem

__init__(problem: iOpt.problem.Problem, best_trials: array([None], dtype=object) = [<iOpt.trial.Trial object>], number_of_global_trials: int = 1, number_of_local_trials: int = 0, solving_time: numpy.float64 = 0.0, solution_accuracy: numpy.float64 = 0.0)

Class constructor

Параметры

• problem – Optimization problem.

• best_trials – Solution the optimization problem.

• number_of_global_trials – Number of global search iterations performed.

• number_of_local_trials – Number of local search iterations performed.

• solving_time – Problem solution time.

• solution_accuracy – Accuracy of the solution found.

solver

solver_parametrs

class iOpt.solver_parametrs.SolverParameters(eps: numpy.float64 = 0.01, r: numpy.float64 = 2.0, iters_limit: int = 20000, evolvent_density: int = 10, eps_r: numpy.float64 = 0.01, refine_solution: bool = False, start_point: iOpt.trial.Point = [], number_of_parallel_points: int = 1, async_scheme: bool = False, timeout: int = - 1, proportion_of_global_iterations: float = 0.95, start_lambdas: list = [], number_of_lambdas: int = 10, is_scaling: bool = False)

Базовые классы: object

The SolverParameters class allows you to define the parameters for searching the optimal solution

__init__(eps: numpy.float64 = 0.01, r: numpy.float64 = 2.0, iters_limit: int = 20000, evolvent_density: int = 10, eps_r: numpy.float64 = 0.01, refine_solution: bool = False, start_point: iOpt.trial.Point = [], number_of_parallel_points: int = 1, async_scheme: bool = False, timeout: int = - 1, proportion_of_global_iterations: float = 0.95, start_lambdas: list = [], number_of_lambdas: int = 10, is_scaling: bool = False)

Constructor of SolverParameters class

Параметры

• eps – The accuracy of the solution to the task at hand. Smaller values – higher search accuracy, less likely to stop prematurely.

• r – Reliability parameter. Higher value of r – slower convergence, higher probability of finding a global minimum.

• iters_limit – maximum number of search trials.

• evolvent_density – density of evolvent construction. The default density is \(2^{-10}\) on the hypercube \([0,1]^N\), which means that the maximum search accuracy is \(2^{-10}\).

• eps_r – parameter affecting the speed of solving the problem with constraints. eps_r = 0 - slow convergence to the exact solution, eps_r>0 - fast convergence to the neighbourhood of the solution.

• refine_solution – if true, the solution will be refined using the local method.

• start_point – point of initial approximation to the solution.

• number_of_parallel_points – number of parallel computed trials.

• timeout – calculation time limit in minutes.

• proportion_of_global_iterations – share of global iterations in the search when using the local method.

to_string() → str

Creates a string containing the values of the main parameters

Результат

string containing the parameters values

trial

class iOpt.trial.FunctionType(value)

Базовые классы: enum.Enum

An enumeration.

OBJECTIV = 1

CONSTRAINT = 2

class iOpt.trial.Point

Базовые классы: object

class iOpt.trial.FunctionValue(type: iOpt.trial.FunctionType = FunctionType.OBJECTIV, functionID: int = 0)

Базовые классы: object

class iOpt.trial.Trial(point: iOpt.trial.Point, function_values: array([None], dtype=object))

Базовые классы: object

iOpt Implementation

• iOpt.method Package
  • listener
  • method
  • optim_task
  • process
  • search_data
  • sol_value
• iOpt.evolvent Package
  • evolvent
• iOpt.output_system.console Package
  • console_output
• iOpt.output_system.painters Package
  • dynamic_painter
  • static_painter