Gym Environment for Discounted Reward Simulation

This project implements a custom Gym environment that simulates a simple system where an agent moves between states based on probabilistic actions and receives rewards. The goal is to explore various policies (e.g., hard-coded, random) and evaluate the performance of these policies using discounted rewards.