rllib

Ray is an AI compute engine. Ray consists of a core distributed runtime and a set of AI Libraries for accelerating ML workloads.

One repository is all that is necessary for Multi-agent Reinforcement Learning (MARL)

VMAS is a vectorized differentiable simulator designed for efficient Multi-Agent Reinforcement Learning benchmarking. It is comprised of a vectorized 2D physics engine written in PyTorch and a set of challenging multi-robot scenarios. Additional scenarios can be implemented through a simple and modular interface.

An open, minimalist Gymnasium environment for autonomous coordination in wireless mobile networks.

A custom MARL (multi-agent reinforcement learning) environment where multiple agents trade against one another (self-play) in a zero-sum continuous double auction. Ray [RLlib] is used for training.

Deep Reinforcement Learning For Trading

Schola is a plugin for enabling Reinforcement Learning (RL) in Unreal Engine. It provides tools to help developers create environments, define agents, and connect to python-based RL frameworks such as OpenAI Gym, RLlib or Stable Baselines 3 for training agents with RL.

An introductory tutorial about leveraging Ray core features for distributed patterns.

Walkthroughs for DSL, AirSim, the Vector Institute, and more

Adaptive real-time traffic light signal control system using Deep Multi-Agent Reinforcement Learning

RLlib tutorials

Dynamic multi-cell selection for cooperative multipoint (CoMP) using (multi-agent) deep reinforcement learning

Reinforcement learning algorithms in RLlib

An open source library for connecting AnyLogic models with Reinforcement Learning frameworks through OpenAI Gymnasium

Tutorial for Ray

An example implementation of an OpenAI Gym environment used for a Ray RLlib tutorial

Autonomous driving agent in Carla simulator leveraging IL and RL techniques.

Legacy case study (2021): Kafka + ksqlDB + Ray/RLlib + Ray Serve. Distributed streaming features and RL training/serving — not a trading product.

Nature and nurture combined adapt faster than either alone. Darwinian evolution (agents selected across generations) plus Multi-Agent Deep RL (agents learning within a lifetime) in one ecosystem, each mechanism amplifying the other — a testbed for the Baldwin effect, coevolution, and emerging cooperation.

Construction of controllers for Shadow-Hand in Mujoco environment, using Deep Learning. 2 Different methods were used to create the controllers: a) Behavioral Cloning b) Deep Reinforcement Learning