Skip to content

Gostas/Reinforcement-Learning-on-text-based-gridworld

Repository files navigation

Reinforcement Learning on text based gridworld

This is a simulation driven by two separate Reinforcement Learning algorithms called Q-learning and SARSA. The simulation includes an agent which tries to learn the shortest, safest path from a given position to a certain point in a grid. The user is able to compare the algorithms using data generated during the simulation and evaluate their effectiveness and performance. The environment and RL agent are programmed in pure C++ whereas the graphs are generated using Python.

A video demonstration of the program can be found at https://youtu.be/h4inhJ41yXM.

About

Reinforcement Learning agent training and comparison framework

Topics

Resources

Stars

Watchers

Forks