GitHunt
TA

tadiwa-aizen/Reinforcement-Learning

Designed a Q-learning agent for navigating a 'Four-Rooms' grid-world, mastering package collection with increasing complexity and handling both deterministic and stochastic action spaces.

Grettings, welcome to The read me for Reinforcement Learning Project
There are 3 python scripts, each corresponding to a part of the project
The scripts are scenario 1 , whihc has the leaner search for 1 package
Scenario 2 has the leanrer search for mukltiple packges and scenario3 searches for multiple packages packages in order(RED,BLUE THE GREEN)
There is a makefile available, it downloads the virtual enviromnement and the necessary packages
from there its all very simple, to operate the porgram just type 'make scene{n}' and this will run the corresponding RL scenario
The results are displayed as a png file in the same directory

The option to make the scenarios stochastic is added an can be specified by the user when they first run the programs
Thank You

Note, if the leaner does not perfom satifactory, please just run it again and i am sure its pefromance will increase

Languages

Python98.0%Makefile2.0%

Contributors

Created November 11, 2023
Updated January 14, 2024
tadiwa-aizen/Reinforcement-Learning | GitHunt