This is a repository of examples and tutorials for learning how to use the POMDPs.jl package for working with Markov Decision Processes (MDPs) and Partially Observable Markov Decision Processes (POMDPs). See also QuickPOMDPs.jl for a concise way to define problems.
Each tutorial is a notebook in the notebooks directory. Here is a list:
- Defining a POMDP with the Explicit Interface
- Defining a POMDP with the Generative Interface
- Defining a POMDP with Probability Tables
- Running Simulations
- Using an Offline Solver
- Using an Online Solver
- Defining a Heuristic Policy
- Defining MDP of gridWorld and using VI and MCTS to find solutions
There are also several tutorials contained in the legacy folder which are not tested and may not be up to date, but may nevertheless be useful.
Some additional notebooks created for Stanford's AA-228 course may be found at https://github.com/sisl/aa228-notebook