RL-ExpectedSarsa-MountainCar

https://github.com/yuqian5/RL-ExpectedSarsa-MountainCar/blob/master/README.md

Member

Kerry Cao	Junyao Cui

Code Explain

• agent.py
• • Contains the code for the Expected SARSA agent
• MontainCarEnv.py
• • Contains the code for the Mountain Car Environment
• q3.py
• • Produce the chart for Part 3
• q4.py
• • Produce the chart for Part 4
• utility.py
• • Contain utility functions for the agent
• tiles3.py
• • Copy from Rich Sutton for tile coding

Part 3

Agent parameter

ALPHA	EPSILON	NUM_TILE	NUM_TILING
0.1	0.0	8	8

50 Run average

Mean and STD. ERROR

MEAN	STD. ERROR
653.3927	47.1082428602

Part 4

Agent parameter

ALPHA	EPSILON	NUM_TILE	NUM_TILING
0.5	0.01	4	32

50 Run average

Mean and STD. ERROR

MEAN	STD. ERROR
242.4813	11.6725668508

Graph Compare

Legend

ALPHA	EPSILON	NUM_TILE	NUM_TILING	Line Colour
0.1	0.0	8	8	Blue
0.2	0.01	4	32	Green
0.5	0.01	4	32	Red

To Repeat Experiment

• Part 3

• • python3 q3.py
    

• Part 4

• • python3 q4.py