src.baseline.metabbo.dedqn

Module Contents

Classes

DEDQN

Introduction

DEDQN is a mixed mutation strategy Differential Evolution (DE) algorithm based on deep Q-network (DQN), in which a deep reinforcement learning approach realizes the adaptive selection of mutation strategy in the evolution process.

API

class src.baseline.metabbo.dedqn.DEDQN(config)[source]

Bases: src.rl.dqn.DQN_Agent

Introduction

DEDQN is a mixed mutation strategy Differential Evolution (DE) algorithm based on deep Q-network (DQN), in which a deep reinforcement learning approach realizes the adaptive selection of mutation strategy in the evolution process.

Original paper

“Differential evolution with mixed mutation strategy based on deep reinforcement learning.” Applied Soft Computing (2021).

Official Implementation

None

Args:

  • config (object): Configuration object containing agent and environment parameters. The constructor modifies and extends this configuration with DEDQN-specific settings, such as state size, action space, learning rate, optimizer, and neural network architecture.

Attributes:

  • config (object): The configuration object with updated DEDQN parameters.

  • model (MLP): The neural network model used for Q-value approximation.

Example:

agent = DEDQN(config)

Notes:

  • The agent uses an MLP with two hidden layers of 10 units each and ReLU activations.

  • The optimizer is set to AdamW, and the loss criterion is MSELoss.

  • Learning rate decay and gradient clipping are not used by default, following the original DEDQN design.

  • The agent’s save directory is automatically constructed based on its string representation and training name.

See Also:

  • DQN_Agent: The base class for DEDQN.

  • MLP: The neural network class used for function approximation.

Initialization

Initializes the DQN agent with the given configuration, networks, and learning rates.Store the initial agent in the checkpoint directory.

Args:

  • config: Configuration object containing all necessary parameters for the experiment.

  • network (dict): A dictionary of neural networks used by the agent.

  • learning_rates (float): Learning rate for the optimizer.

__str__()[source]