pynol.online_learning

online_learning

pynol.online_learning.online_learning(T, env: Environment, learner: Base | Model)[source]

Combine the environment and learner, start the online learning process.

Parameters:

  • T (int) – Number of total round.

  • env (Environment) – Environment.

  • learner (Base, Model) – Learner.

Returns:

tuple contains:

x (numpy.ndarray): Decisions over \(T\) rounds.

loss (numpy.ndarray): Losses over \(T\) rounds.

surrogate_loss (numpy.ndarray): Surrogate losses over \(T\) rounds.

tm (numpy.ndarray): Time cost over \(T\) rounds.

Return type:

tuple

multiple_online_learning

pynol.online_learning.multiple_online_learning(T, env: Environment, learners: list, processes=4)[source]

Combine the environment and multiple learners, start the online learning process with multiprocessing to speed up.

Parameters:

  • T (int) – Number of total round.

  • env (Environment) – Environment.

  • learners (list) – Learners.

Returns:

tuple contains:

x (numpy.ndarray): Decisions of all learners over \(T\) rounds.

loss (numpy.ndarray): Losses of all learners over \(T\) rounds.

surrogate_loss (numpy.ndarray): Surrogate losses of all learners over \(T\) rounds.

tm (numpy.ndarray): Time cost of all learners over \(T\) rounds.

Return type:

tuple