Module mdp 

Markov Decision Process (MDP) solvers

This module provides exact and approximate MDP solvers including:

• Value Iteration
• Policy Iteration
• Modified Policy Iteration
• Linear Programming approach
• Q-Learning (model-free)
• SARSA (on-policy TD)
• RTDP (Real-Time Dynamic Programming)
• Prioritized Sweeping
• Stochastic Shortest Path
• Inverse Reinforcement Learning (MaxEnt IRL)

Re-exports

pub use planning::*;

pub use tabular::*;

Modules

planning

Advanced MDP planning algorithms.

tabular

Tabular MDP algorithms: value/policy iteration, Q-learning, SARSA.