Tabular methods for reinforcement learning

PyTorch implementation of the Q-Learning Algorithm Normalized Advantage Function for continuous control problems + PER and N-step Method

This repository contains the code for automatically generating piano fingerings using a reinforcement learning agent that uses Q-Learning.

Open-zero is a research project aiming to realize the various projects of the company DeepMind

The objective is to teach robot to find and reach the target object in the minimum number of steps and using the shortest path and avoiding any obstacles such as humans, walls, etc usinf reinforcement learning algorithms.

Turn based strategy game with AI

Implementation of Q-learning to solve GridWorld

Q-learning application to find an optimal parking slot

Q-Learning Based Pathfinding in Dynamic Grid Environments

Implementation of Q-Learning, Double Q-Learning, and LSPI for pricing American options under the Black-Scholes model

This repository contains a Jupyter Notebook with an implemenation of a Q-Learning Agent, which learns to solve the n-Chain OpenAI Gym environment

Two intelligent agents (cat and mouse) compete with each other to achieve their goal. Agents are trained through reinforcement learning (Q-learning).

Deep Q Learning blackbox strategies for casino games

The implementation for the paper Continuous Deep Q-Learning in Optimal Control Problems: Normalized Advantage Functions Analysis // NeurIPS 2022

Demonstration of Q-Learning and SARSA algorithms utilizing Python and OpenAI GYM

a Python-based platformer infused with Q-Learning and dynamic level creation from simple JSON files.

A reinforcement learning agent with reflection capabilities for dynamic maze navigation. Implements dual memory system, real-time adaptation, and environment change detection. Open source with research papers and documentation.

Docking robot in a grid environment trained it with Q-learning

🕹️ Welcome to Game-Optimization, a repository dedicated to exploring and implementing various optimization algorithms to solve complex games. This project initially focuses on solving the classic game Sokoban using the Q-learning algorithm, with plans to extend to genetic algorithms and other optimization techniques in the future.

SUTD 50.021 Artificial Intelligence Project - Wordle Solver using Reinforcement Learning

The 3D bin packing problem is a combinatorial optimization problem that involves fitting a given set of items of various sizes into a container of a specific size such that the total volume of the items is as close to the volume of the container as possible.

Markov decision process master thesis

Q-Learning applied to Gymnasium's Toy Text environments: FrozenLake, CliffWalking, BlackJack, and Taxi.

Codes for the AISTATS 2023 paper, A Statistical Analysis of Polyak-Ruppert Averaged Q-learning.

Build an RL (Reinforcement Learning) agent that learns to play Numerical Tic-Tac-Toe. The agent learns the game by Q-Learning.

No description provided.

Dynamic Q-Learning Based Feature Selection approach

This repository contains various networks implementation such as MLP, Hopfield, Kohonen, ART, LVQ1, Genetic algorithms, Adaboost and fuzzy-system, CNN with python.

In this project, we tried two different Learning Algorithms for Hierarchical RL on the Taxi-v3 environment from OpenAI gym. SMDP Q-Learning and Intra Option Q-Learning and contrasted them with two other methods that involve hardcoding based on human understanding. We conclude that the solutions learnt by machine are way superior than humans for this problem. Intra Option Q-Learning outperforms SMDP Q-Learning because of better usage of the SARS samples (similar to experience replay). Our algorithms even outperform the Hardcoded Agent. We also demonstrated and concluded the strong effectiveness of state compression on the model performance.

A collaborative repository for our Bachelor's thesis, focused on optimizing the Cell Outage Compensation (COC) algorithm in Self-Organizing Networks (SONs). Leveraging AI-Hardware Acceleration, the project aims to bolster 5G network reliability, particularly for emerging technologies like autonomous driving.