Skip to content

Latest commit

 

History

History
33 lines (18 loc) · 1.34 KB

README.md

File metadata and controls

33 lines (18 loc) · 1.34 KB

Sample-Based_RL_Gym

Overview

Environment Animation

Methods Comparison

This project implements and trains sample-based reinforcement learning (RL) algorithms from scratch in Python, using the Taxi-v3 environment from OpenAI's Gym. The focus is on analyzing the performance of different algorithms, including Monte Carlo methods and Temporal Difference (TD) methods such as Q-learning and SARSA.

Methods Implemented

1. Monte Carlo Method

  • Updates occur only at the end of each episode.
  • Evaluates the agent’s performance and convergence speed over multiple episodes.

2. Temporal Difference (TD) Methods

  • Updates occur at every step, leading to faster convergence compared to Monte Carlo methods.
  • Compares Q-learning and SARSA algorithms to evaluate variance in rewards and training speed.

Key Findings

  • TD methods generally exhibited lower variance in rewards and faster training speeds compared to the Monte Carlo method.
  • Among Q-learning and SARSA, Q-learning demonstrated less variance in rewards and lower error percentages.

Results

Training Performance

The performance metrics and analysis of the implemented methods are documented in the provided report. Please refer to the respective files for comprehensive insights into the training outcomes.