src.environment.optimizer.learnable_optimizer

This is a basic class for learnable backbone optimizer. Your own backbone optimizer should inherit from this class and have the following methods: 1. init(self, config) : to initialize the backbone optimizer. 2. init_population(self, problem) : to initialize the population, calculate costs using problem.eval() and record some information such as pbest and gbest if needed. It’s expected to return a state for agent to make decisions. 3. update(self, action, problem) : to update the population or one individual in population as you wish using the action given by agent, calculate new costs using problem.eval() and update some records if needed. It’s expected to return a tuple of [next_state, reward, is_done] for agent to learn.

Module Contents

Classes

Learnable_Optimizer

Introduction

Abstract superclass for learnable backbone optimizers, providing a template for population initialization, update mechanisms, and random seed management for reproducible optimization experiments.

API

class src.environment.optimizer.learnable_optimizer.Learnable_Optimizer(config)[source]

Introduction

Abstract superclass for learnable backbone optimizers, providing a template for population initialization, update mechanisms, and random seed management for reproducible optimization experiments.

Args:

  • config (Any): Configuration object containing optimizer settings, including device information.

Attributes:

  • __config (Any): Stores the configuration object.

  • rng_seed (Optional[int]): The random seed used for reproducibility.

  • rng (np.random.RandomState): Numpy random number generator initialized with the seed.

  • rng_cpu (torch.Generator): PyTorch CPU random number generator initialized with the seed.

  • rng_gpu (Optional[torch.Generator]): PyTorch GPU random number generator, if applicable.

Methods:

  • init_population(problem: Basic_Problem) -> Any: Abstract method to initialize the population for the optimization problem.

  • update(action: Any, problem: Basic_Problem) -> Tuple[Any]: Abstract method to update the optimizer state based on the given action and problem.

  • seed(seed: Optional[int] = None): Sets the random seed for numpy and PyTorch generators for reproducibility.

Raises:

  • NotImplementedError: If init_population or update methods are not implemented in a subclass.

Initialization

abstractmethod init_population(problem: src.environment.problem.basic_problem.Basic_Problem) → Any[source]

Introduction

Initializes the population for the optimization process based on the provided problem definition.

Args:

  • problem (Basic_Problem): An instance representing the optimization problem, containing information such as the search space, constraints, and objective function.

Returns:

  • Any: The initialized population, with the specific type and structure depending on the implementation and problem requirements.

Raises:

  • NotImplementedError: This method must be implemented by subclasses; calling this base method will always raise this exception.

abstractmethod update(action: Any, problem: src.environment.problem.basic_problem.Basic_Problem) → Tuple[Any][source]

Introduction

Updates the optimizer’s internal state based on the provided action and problem instance.

Args:

  • action (Any): The action to be applied for updating the optimizer.

  • problem (Basic_Problem): The problem instance containing relevant information for the update.

Returns:

  • Tuple[Any]: A tuple containing the results of the update operation.

Raises:

  • NotImplementedError: This method must be implemented by subclasses.

seed(seed=None)[source]