Markov Decision Processes using Dynammic programming for the Door-Key problems.
There are results for 7 test scenes
| doorkey-5x5-normal |
|---|
 |
| doorkey-6x6-normal | doorkey-6x6-direct | doorkey-6x6-shortcut |
|---|---|---|
 |
 |
 |
| doorkey-8x8-normal | doorkey-8x8-direct | doorkey-8x8-shortcut |
|---|---|---|
 |
 |
 |
- Install gym-minigrid
- Install dependencies
pip install -r requirements.txtYou will need to modify doorkey.py as the main entrance.
You might find some useful tools in utils.py
- step(): Move your agent
- generate_random_env(): Generate a random environment for debugging
- load_env(): Load the test environments
- save_env(): Save the environment for reproducing results
- plot_env(): For a quick visualization of your current env, including: agent, key, door, and the goal
- draw_gif_from_seq(): Draw and save a gif image from a given action sequence. Please notice that you have to submit the gif!
The example.py shows you how to interact with the utilities in utils.py, and also gives you some examples of interacting with gym-minigrid directly.
-
BFS: gets the shortest cost from all possible states to the given state. returns an array with cooresponding costs and a tree having child nodes for that shortest path.
-
get_shortest_path(): returns the optimal controls which passes throught the best pickup position
-
controls_to_sequence(): takes the optimal con trols and returns the optimal sequence. Note: Control is different from sequence here
-
get_pickup_positions(): returns all the possible pickup positions
-
get_best_pickup_position(): returns best pickup position that minimizes cost from start to key and key to door
-
Start_To_Goal_direct(): computes cost and policy for direct reach if possible
-
Start_To_Goal_viaDoor(): computes cost and policy for path via Door
Just change the variable env_name to the desired environment. It should be located in ./env folder and must have extension .env.