DQN-Acrobot-v1 Open-Source Reinforcement Learning Agent - Free Solution to Acrobot-v1 Control Problem

Dqn Acrobot V1

Developed by sb3

This is a DQN reinforcement learning agent trained using the stable-baselines3 library, specifically designed to solve the Acrobot-v1 control problem.

Physics Model #Double pendulum control #Deep Q-Network #Reinforcement learning training

Downloads 403

Release Time : 6/2/2022

Model Overview

This model uses the Deep Q-Network (DQN) algorithm trained in the Acrobot-v1 environment to learn how to control a double-link pendulum system to reach the target state.

Model Features

Based on stable reinforcement learning framework

Implemented using the stable-baselines3 library, a reliable reinforcement learning framework

Optimized hyperparameter configuration

Tuned hyperparameter settings including learning rate, exploration strategy, etc.

Complete training process support

Supports training, evaluation, and deployment through the RL Zoo framework

Model Capabilities

Reinforcement learning control

Continuous action space processing

Environment state perception

Use Cases

Academic research

Reinforcement learning algorithm comparison

Can serve as a benchmark model to compare performance with other reinforcement learning algorithms in the Acrobot environment

Average reward -72.10 ±6.44

Educational demonstration

Reinforcement learning teaching case

Used to demonstrate the application of DQN algorithm in control problems

Property	Details
Library Name	stable-baselines3
Tags	Acrobot-v1, deep-reinforcement-learning, reinforcement-learning, stable-baselines3
Model Name	DQN
Mean Reward	-72.10 +/- 6.44
Task Type	reinforcement-learning
Dataset	Acrobot-v1