Skip to content

A Taxi driver agent that learns using Q learning and Sarsa

Notifications You must be signed in to change notification settings

nazaninsbr/Intellignet-Taxi

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Intelligent Taxi

I solved the taxi-v2 problem (it's a really famouse one, you could just google it), using Q-Learning, Sarsa and using the OpenAI Gym enviroment.

How do I run it?

you could use one of the following commands to run the corresponding code:

python main.py --qlearning 

or 

python main.py --sarsa

or 

python main.py --gym

The test is hard coded at the moment, but you could change it by changing the testTheModel function.

The output of the OpenAI Gym version is a graphical one:

but Q-Learning and Sarsa output a text based result which includes, action number, state and choosen action.

they also draw a plot showing the total reward in each step:

Q-Learning and Sarsa

The States are modeled as follows:

(POSITION[0], POSITION[1], PASSENGER, GOAL)

where POSITION is the taxi location in the map (row and column, top left is [0, 0]), and PASSENGER is one of the following characters ['R', 'G', 'Y', 'B', 'I'], where 'R', 'G', 'Y' & 'B' are points on the map and 'I' stands for Inside-the-taxi. GOAL is also of the following strings: ['R', 'G', 'Y', 'B']