Reinforcement-Learning-Practice

Practice code from Reinforcement Learning

Available algorithms

• Policy Iteration from MDP
  • policy_iteration
• Value Iteration from MDP
  • value_iteration
• Monte Carlo Prediction
  • First Visit-MC
  • Every Visit-MC
• Temporal Difference
• N-Step Temporal Difference
• Temporal Difference-Lambda
• SARSA
• Q-Learning

Available Utils

• common_utils:
  • plot_policy
  • plot_state_value_function
  • evaluate_policy
  • improve_policy
  • probability_success
  • mean_return
  • print_policy_success_stats
  • generate_random_policy
  • rmse
  • decay_schedule
  • generate_trajectory
  • generate_trajectory_epsilon_greedy
  • print_action_value_function
  • get_policy_metrics
  • moving_average
  • choose_epsilon_greedy_action