This repository is a Q-learning implementation of OpenAI Gym Mountain Car game.
The goal is reaching to the flag by using 3 different actions, including 'left', 'nothing', 'right'. However, the force is not enough to reach to the flag just by 'right' action and the agent must use the momentum of car.
Just copy-paste the mountain_car_v0.py to your project or clone this repository and run mountain_car_v0.py using python.
The car reaches to the goal just after a few hundreds of episodes.
| Episodes | Output | Output | Output |
|---|---|---|---|
| Episode 200, 400, 600 |  |
|
 |
| Episode 800, 1000, 1200 |  |
 |
 |
| Episode 1400, 1600, 1800 |  |
 |
 |
| Episode 2000, 2200, 2400 |  |
 |
 |
| Episode 2600, 2800, 3000 |  |
 |
 |
| Episode 3200, 3400, 3600 |  |
 |
 |
| Episode 3800, 4000, 4200 |  |
 |
 |
| Episode 4400, 4600, 4800 |  |
 |
 |
| Episode 5000, 5200, 5400 |  |
 |
 |
| Episode 5600, 5800, 6000 |  |
 |
 |
| Episode 6200, 6400, 6600 |  |
 |
 |
| Episode 6800, 7000, 7200 |  |
 |
 |
| Episode 7400, 7600, 7800 |  |
 |
 |
| Episode 8000, 8200, 8400 |  |
 |
 |
| Episode 8600, 8800, 9000 |  |
 |
 |
| Episode 9200, 9400, 9600 |  |
 |
 |
| Episode 10000, 10200, 10400 |  |
 |
 |
With special thanks to sentdex, An amazing Youtuber!